Professional Documents
Culture Documents
Oom Mash Manual
Oom Mash Manual
TABLE OF CONTENTS
Due to large file size, the MASH 2016 Manual was split into 4 sections. To find which section you want, review this Table
of Contents.
✓ Part I (pages 1-101)
✓ Part II (pages 102-138)
✓ Part III (pages 139-226)
✓ Part IV (pages 227-264)
MASH 2016 Manual - Part I - includes through pages 1-101 (back to top)
Chapter 1 – Introduction
1.1 Purpose and Scope 1
1.2 Underlying Philosophy 2
1.3 Performance Limitations 3
1.4 Safety Features 4
1.5 Test Levels 5
1.6 International Harmonization 6
1.7 Analytical and Experimental Tools 6
1.8 Organization of Manual 7
MASH 2016 Manual - Part II - includes through page 102-138 (back to top)
5.2.1 Structural Adequacy 104
5.2.2 Occupant Risk 105
5.2.3 Post-Impact Vehicular Response 108
5.3 GEOMETRIC FEATURES 110
MASH 2016 Manual - Part III - includes through pages 139-226 (back to top)
A2.3.2.2 Tests with 10000S, 36000V, and 36000T Vehicles 141
CHAPTER THREE 142
A3.2 TESTING SITE 142
A3.3 SOIL 142
A3.3.1 Standard Soil 143
A3.3.2 Soil Strength 143
A3.3.3 Special Soils 144
A3.3.4 Embedment of Test Article 144
A3.4 TEST ARTICLE 144
A3.4.2.1 Longitudinal Barriers 144
A3.4.2.4 Truck-Mounted Attenuators (TMA) 145
CHAPTER FOUR 145
A4.2 TEST VEHICLE DESCRIPTIONS 145
A4.2.1 Production Vehicles 146
A4.2.1.1 Test Vehicle Mass 147
A4.2.1.2 Ballast 148
A4.2.1.4 Vehicle Damage 148
A4.2.1.5 Surrogate Occupants 149
A4.2.2 Surrogate Test Vehicles 149
A4.2.3 TMA Support Truck 151
A4.3.1 Instrumentation Specifications 152
A4.3.2 Acceleration and Rate Gyro Placement and Data Reduction 152
CHAPTER FIVE 156
A5.1 GENERAL 156
A5.2 EVALUATION FACTORS AND CRITERIA 156
A5.2.1 Structural Adequacy 156
A5.2.2 Occupant Risk 157
A5.2.3 Post-Impact Vehicular Response 162
A5.3 GEOMETRIC FEATURES 163
CHAPTER SIX 164
A6.1 GENERAL REPORTING RECOMMENDATIONS 164
CHAPTER SEVEN 164
A7.1 PURPOSE 164
MASH 2016 Manual - Part IV - includes through pages 227-264 (back to top)
Appendix F—Determination of THIV, PHD, and ASI 228
F1 Introduction 228
F2 A Guide to the Measurement of Theoretical Head Impact Velocity (THIV) and the Post-Impact
Head Deceleration (PHD) 228
F2.1 General 228
F2.2 Theoretical Head Impact Velocity (THIV) 228
F2.3 Post-Impact Head Deceleration (PHD) 232
F2.4 Summary of Procedure to Compute THIV and PHD 232
F3 A Guide to the Measurement of the Acceleration Severity Index (ASI) 234
F3.1 Procedure 234
F3.2 Summary 235
AASHIO
Manual for
o
A ssessrng
H ardware
Second Edition
201,6
PROPERTY OF
TRAFFIC ENGINEERING DESIGN DIVISION
DO NOT REMOVE FROM THE OFFICE
@ 2016, by American Associalion of State Highway and Transportation Officials. All rights reserved,
This book, or parts thereol may not be reproduced in any form without written permission of the publisher.
Printed in the United States of America,
ISBN: 978-1-56051-665-1
Regional Representatives
REGION I
REGION II
Paul Mattox, West Vîrginia, One-Year Term
Charles Kilpatrick, Virginia, Two-Year Term
REGION I¡!
Charles A. Zelle, Minnesota, One-Year Term
Randall S. Blankenhorn, lllinois, Two-Year Term
REG]ON IV
Brian Ness, ldaho, One-YearTerm
Carlos Braceras, Utah, Two-Year Term
Non-Voting Members
lmmediate Past President: (Vacant)
Executive Director: Bud Wright, Washington, DC
ALABAMA, \Mlliam Colquett, Eric Christie, Blake Mullins OKLAHOMA, Steven Jacobi, Walter Peters
ALASKA, Richard A. Pratt OREGON, Tanarat Potisuk, Hormoz Seradj
ARIZONA, David Eberhart, David Benton, Pe-Shen Yang PENNSYLVANIA, Thomas P. Macioce, James Long,
ARKANSAS, Charles Richard Ellis Lou Ruzzi
CALIFORNIA, Thomas Ostrom, Susan Hida, Dolores Valls PUERTO RICO, Vacant
COLORADO, Behrooz Far, Stephen Harelson, Jessica Marti- RHODE ISLAND, ceorgette Chahine
nez SOUTH CAROLINA, Terry Koon, Barry W. Bowers, Jeff
CONNECTICUT, ïmothy Fields Sizemore
DELAWARE, BarryA. Benton, Jason Hastings SOUTH DAKOTA, Steve Johnson
DISTRICT OF COLUMBIA, Konjit "Connie' Eskender, TENNESSEE, \Â/byne Seger, John Hastings
Donald Cooney, Richard Kenney TEXAS, Gregg Freeby, Bemie Carrasco, Jamie Fanis
FLORIDA, Sam Fallaha, \Mlliam Potter, Jeffrey Pouliotte UTAH, Carmen Swanwick, Cheryl Hersh Simmons
NEW MEXICO, Ray Trujillo, Ted Barber, Jeff Mgil Christopher Westbrook
NEWYORK, Richard Marchione, Wahid Albert U.S. COAST GUARD, Kamal Elnahal
Representatives f¡om the Subcommittee on Deslgn Members lrom the U.S. Department of Transportat¡on
Kevin Herritt, Califomia Richard Albin, FHWA
C. Andy Casey, Georgia Eduardo Arispe, FHWA
Scott King, Kansas Frank Julian, FHWA
Jason Siwula, Kentucky \Mll Longstreet, FHWA
Paul Fossier, Jr., Louisiana
Teri Soos, Maryland Assoclate Member-lnternatlonal
Alexander BardoW Massachusetts MarkAyton, Ontario
Michael Elle, Minnesota
Arielle Ehrlich, Minnesota Associate Membel-Other
David Bizuga, New Jersey Mark Bush, TRB
Bradley Bortn¡ck, New York
Bucky Galloway, North Carolina AASHTO
Bernie Clocksin, South Dakota Kevin Sablan, LÍaison
Christopher Lindsey, Texas
Alex Price, Mrginia
Erik Emerson, Wsconsin
William Wilson, Vgoming
PREFACE
The use ofeffective roadside safety features provides a reasonable level ofsafety for the traveling pub-
lic. New systems ale continually emerging to address safety problems. Devices and practices improve in
response to an increased understanding ofsafety performance, a changing vehicle fleet, the emergence
of new materials, and other factors.
This second edition of the Manual for Assessing Safety Hardware (MASH) is the latest in a long series
of crash testing guidance documents dating back to 1 962. This update incorporates many changes from
the previous edition. Some of the more significant changes include:
This document's purpose is to encourage consistency in crash testing and evaluation. Full-scale crash
will continue to be the most common method of evaluating the impact performance
testing has been and
of safety hardware. Consistency in crash testing and evaluation benefits states, testing facilities, manu-
facturers, and the public.
Note that MASH addresses only the crash testing of roadside safety features. It does not contain in-
stallation or maintenance guidance, nor does it supersede any of the guidelines found in the AASHTO
Roadside Design Guide.
TABLE OF CONTENTS
Chapter 1-lntroduction....,....,,
I.I PURPOSEAND SCOPE
3.4. I General 73
F2 A Guide to the Measwement of the Theoretical Head Impact Velocity (THIV) and
LIST OF FIGURES
Figure 2- I . Impact Conditions fol Longitudinal Barrier Tests .20
Figure 2-24. Critical Cable Barrier Placement f'or' 4H:l V V-Ditch .21
Figure 2-28. Critical Cable Barrier Placement for 6H: I V V-Ditch .22
Figure 2-34. Impact Conditions for Terminal and Redirective Crash Cushion Tests ............... .30
Figure 2-3B. Impact Conditions for Non-Redirective Crash Cushion Tests .32
.38
Figure 2-5. Impact Conditions for Support Structures, Work-Zone Traffic Control Devices,
and Breakaway Utility Poles.... .....................43
Figure 2-6. Critical Impact Point for Test I 0, Test Level I 5l
Figure 2-7. Critical Impact Point for Test 10, Test Level 2 .52
Figure 2-15. Critical Impact Point for Test 21, Test Level I 60
Figure 2-16. Critical Impact Poínt for iest 21, Test Level 2 61
Figure 2-17. Critical Impact Point fol'Test 21, Test Levels 3, 4, 5, ætd 6 62
Figure 2-18. Critical Irnpact Point forTest 34 on Non-Gating Crash Cushions......... 64
Figure 3-1. Recommended Summary Sheet lbr Shong Soil Test Resu|ts.................. ...............................70
Figure 3-2. Example of Test Day Static Soil Strength Documentation .............................71
Figure B-3. Dynamic and Static Test Results flor Standard Post Test......... .....................170
Figure B-5. Test Day Static Load Test Compared to Standard Test.................. ..............173
Figure E- | . Ple-Intpact Measurement.. ,.220
Figure E-2. Placement of Swivel Laser Bracket. ..221
Figule E-3. Measurement of Vertical Positions ..222
Fi gure E-4. Veh icle Del'ormation Spreadsheet. ..223
..224
Figule E-6. Field Length Measulement ..225
Figure E-7. Crush Depth Measurements. ..226
..227
Figure F-l. Vehicle and Ground Reference Frames............ ..229
Figule F-2. Impact of the Theoretical Head on the Left Side .231
LIST OF TABLES
Table 2-L Vehicle Gross Static Mass Upper and Lovvel Limits.. .............13
Table2-2A. Recomrnended Test Matrices fol Longitudinal Ban'ie1s......... .......................15
Table 2-28. Recommended TL-3 Test Matrix for Single Median Barlier Designed for Placenrent Anywhele in
Table 2-2C. Recommended TL-3 Test Matrix for Single or Double Median Barrier Designed for Placement
Between 0- to 4-ft Offset from Slope Break.... ..................17
Table2-2D. Recommended TL-3 Test Matrix for Single Median Barrier Designed for PlacementAnywhere in
Table2-28. MASH TL-3 Test Matrix for Single or Double Median Barrier Placed at 0- to 4-ft Offset from
SBP of 6H:lVV-Ditch............... ......:.........19
Table 2-3. Recommended Test Matrices for Terminals and Crash Cushions ....................27
Table 2-4. Recommended Test Matrices for Tnrck- and Trailer-Mounted Attenuators..................... ....,...37
Table 2-5. Recommended Test Matrices I'or Support Structures, Work-Zone Tralfic Control
Devices, and Breakaway Utility Poles.... ...........................42
Table2-6. Recommended Post Spacing forEvaluating Cable Barriers Placed wittrin Median Ditches....,..,..,................47
Table 2-7 . Critical Impact Point for Rigid Barrier Tests witlì I l00C and 2270P Vehicles................ ........50
Table 2-8. Critical lmpact Point for Heavy Vehicle TesS ..........................63
Table 4-1. Recommended Properties of ll00C, 15004, and 2270PTestVehicles......,.......... ...................85
Table 4-2. Recommended Properties of 100005, 36000V and 36000T Test Vehicles...... ........................86
Table 5-lA. Safety Evaluation Guidelines for Structural Adequacy .......102
Table 5-lB. Safety Evaluation Guidelines for Occupant Risk................. ........................103
Table 5-l C. Safety Evaluation Guidelines for Post-Impact Vehicular Response ........ ............................1 04
Table 6-l. Recommended Table of Contents for Crash Test Reports .... .......................... I l4
Table 6-2. Recommended Fonnat for Reporting of Findirìgs................... ....................... I l8
Table 6-3. Example of Recommended Assessment Summary Page for Individual Clash Tests ll9
Table 6-4. Example of Recommended Assessment Summary Page for Multiple Crash Tests........................................1l9
Table A-1. Ploperties of Common Barrier Rail Elements 139
Table A-3. Dynamic Yield Forces of Posts Embedded in Strong Soi|.................. ...........141
1t
lntroduction
h. purpose of this manual is to present uniform guidelines for the crash testing of both permanent
T
I and temporary highway safety features and recommended evaluation criteria to assess test results.
Guidelines are also presented for the in-service evaluation ofsafety features. These guidelines and
criteria, which have evolved over the past 40 years, incorporate current technology and the collective
judgment and experlise of professionals in the field of roadside safety design. They provide: (l) a basis
on which researchers and user agencies can compare the impact performance merits of candidate safety
features, (2) guidance for developers ofnew safety features, and (3) a basis on which user agencies can
formul ate perform ance spec ifi cati ons for safety features.
A goal of a highway safety feature is to provide a forgiving roadway and roadside that reduces the risk
of a serious crash when a motorist leaves the roadway. The safety goal is met when the feature either
contains and redirects the vehicle away from a roadside obstacle, decelerates the vehicle to a safe stop,
readily breaks away or fractures or yields, allows a controlled penetration, or is traversable, without
causing serious injuries to the vehicle's occupants or to other motorists, pedestrians, or work zone
personnel.
Ideally, the roadside would be clear of all obstructions, including unnecessary roadside hardware, and
be traversable so that an en'ant motorist could recover control of the vehicle and stop or retum to the
travelway. However, thele are numerous roadside areas that cannot practically be cleared of all fixed
objects or made traversable. At these sites, the use of an appropriate safety feature or safety treatment is
intended to reduce the consequences ofa departure from the roadway.
The crash testing guidelines presented herein cover vehicular tests to evaluate the impact performance
of permanent and temporary highway safety features. Performance is evaluated in terms of the risk of
injury to occupants ofthe irnpacting vehicle, the structural adequacy ofthe safety feature, the exposure
to workers and pedestrians that may be behind a barrier or in the patlr of debris resulting from impact
with a safety feature, and the post-impact behaviol of the test vehicle. Other factoi's that should be eval-
uated in the design ofa safety feature, such as aesthetics, costs (initial and maintenance), and durability
(ability to withstand environmental conditions such as freezing and thawing, wind-induced fatigue
loading, effects of moisture, ultraviolet radiation. etc.) are not addressed in this document.
The procedures described herein include guidelines for direct impact performance evaluation through
full-scale crash testing as well as general procedures for evaluating the field performance ofa safe-
ty feature. New safety features or significant revisions to existing designs should first be evaluated
through full-scale crash testing. After a system has been proven to meet the recommended impact
performance guidelines, the evaluation should switch to an in-service evaluation of the feature's field
performance. It is recommended that in-service performance evaluations be conducted when new
safety features are placed in service.
The crash testing guidelines provide a minimum set of requirements that a roadside safety feature
has to meet in order to demonstrate its satisfactory impact performance. However, it should be noted
that, while these guidelines are representative and applicable to an affay of highway features and
traffic conditions, they are by no means all-inclusive. Experience has shown that as new designs are
developed, current test procedures may not properly evaluate critical conditions for these designs.
Experience has also shown that evaluation and testing offeatures not addressed by the current guide-
lines will be made. Thereforé, specific features and site conditions may arise that require special tests
and evaluation criteria. Deviations from the guidelines are warranted when other tests or evaluation
criteria are more appropriate and representative of site or design conditions. However, it should also
be understood that it is impractical to test a particular feature for all conditions that may be encoun-
tered in the field and engineering judgment should be exercised when developing policies for the use
ofthese features under differing conditions.
These crash testing and impact performance guidelines supersede those contain ed in NCHfuP Report
350: Recotnmended Procedures for the Safety Perþrmance Eyaluation of Highway Features (129).
Major revisions incorporated herein relative to Report 350 include (a) changes to the test vehicles,
(b) changes to the number and impact conditions of the test matrices, (c) changes to the evaluation
criteria, and (d) addition ofnew features to the test guidelines.
The underlying philosophy in the development of the guidelines is that of 'ïorst practical condi-
tions." When selecting test parameters, such as the test vehicle, impact speed and angle combination,
point of impact, test matrix, etc., every effort is made to speciff the worst, or most critical, conditions.
For example, the weight of the small passenger car test vehicle was selected to represent approxi-
mately the 2nd percentile of passenger type vehicles, i.e., only two percent of vehicles weigh less than
the specified test weight. The impact speed and angle combination represents approximately the
93rd percentile of real-world crashes, i.e., only seven percent of crashes are more severe. V/hen the
combined effects of all testing parameters are considered, the testing represents the extremes of im-
pact conditions to be expected in real-world situations.It is also implicitly assumed that, if a roadside
safety feature performs satisfactorily at the two extremes, then the fealure would also work well for
all impact conditions in between. This assumption has shown to be reasonable for most roadsíde
safety features.
O 2016 by the Anrerican Association ol'State Highway antl Transportat ion Olïcials.
All rights reserved Duplication is a violation olapplicable larv
Chapter 1-lntroduction | 3
On the other hand, the selection of the test parameters must be practical so that the roadside safety
features developed in accordance with the guidelines are cost-effective and provide increased levels of
safety without placing an unrealistic financial burden on user agencies. Considerations need to be given
to available technology and associated costs. The relevancy ofthe test parameters should also be taken
into account, such as increases in the level of safety and potential effects on existing and newly devel-
oped feafures. In many respects, the selection of test parameters is a policy decision as to what level of
safety should be provided and at what cost to the user agencies.
Another underlying philosophy used in developing the guidelines for selected roadside safety features
is that of the "state-of-the-possible." Examples for such features include breakaway sign and luminaire
supports and Category II temporary work-zone traffic control devices. For these roadside safety fea-
tures, technology is already available for designing and manufacturing devices that can meet evaluation
criteria more stringent than those specified for other roadside safety features, thus the term "state-of-the-
possible." For example, the limit of the occupant impact velocity (OlV) for breakaway pole structures is
set at 16 fl/s (5.0 m/s) instead of 39 ff/s (12.0 m/s) and more stringent guidelines are recommended for
evaluating windshield damage for Type I and Type II temporary work-zone traffic control devices. The
rationale for this underlying philosophy is that, since technology is readily available to meet these more
stringent criteria without undue financial burden, it is to the benefit of motorists to provide a higher
level ofsafety.
It should be recognized that the impact performance of a highway feature cannot be measured by a
series of crash tests alone. Crash testing must be viewed as a necessary, but not sufficient, condition to
indicate that a feature would peform satisfactorily under real-world conditions. First, crash testing is
conducted under idealized conditions so that impact performance can be evaluated and compared under
controlled conditions. Second, even the most carefully researched device has performance limits dictat-
ed by physical laws, vehicle stability, and vehicle crashworthiness.
For example, at some sites, sufficient space is lacking to safely decelerate a vehicle, regardless of the
crash cushion design. lrrespective of the breakaway feature, ceftain structural supports may be so
massive that the impacting vehicle is abruptly decelerated, thus limiting achievable impact performance
without a change in support technology. There is no assurance that a feature meeting the test recom-
mendations herein for a tracking vehicle will perform satisfactorily if impacted by a vehicle sliding
sideways or rotating. Some vehicle types may lack sufficient size or mass or necessary crashworthi-
ness features such as interface strength, stiffness, controlled crush properlies, and stability to provide
occupants with an acceptable level of protection, e.g., no provisions are made herein for the design
and testing of safety features for two-wheeled vehicles. Seemingly insignificant site conditions such as
curbs, slopes, and soft soil conditions can contribute to the unsuccessful performance ofa safety feature
for some inrpact conditions.
For these reasons, safety features are generally developed and tested for selected idealized situations
that are intended to encompass a large majority, but not all, of the possible in-seryice collisions. Even
so, it is essential that test results be evaluated and interpreted by competent researchers and that the
evaluation be guided by sound engineering judgment. It is to be expected that certain features, while
meeting all test and evaluation criteria recommended herein, may encounter in-service conditions that
are not covered by the testing. Variations in material characteristics, such as increases and decreases
in steel yield strength from one batch to the next or the thermal sensitivity of the modulus of elasticity
of polymer materials, have been shown to significantly alter the strength and/or stiffness of roadside
safety features. Further, variations in field installation details can materially affect the performance of
some roadside safety features. Thus, the user agency may, at its discretion, require additional testing
and evaluation requirements beyond those set forth herein. The corollary of this is also to be expect-
ed, i.e., certain features not meeting all test and evaluation criteria recommended herein may still be
cost-effective altematives for selected in-service applications. In this case, highway agencies could
continue to utilize safety features that have demonstrated good impact performance through an in-
service performance evaluation.
Finally, it should be emphasized that these guidelines are intended for crash testing and evaluation
ofroadside safety features and not as use warrants. In other words, these guidelines do not address
when, where, and how roadside safety features are to be employed in the field. User agencies should
follow the guidelines set forth in the AASHTO Roadside Design Guide (4) and formulate intemal
policies for directions regarding use \ryarrants.
The impact performance evaluation guidelines cover both permanent and temporary highway safety
features, including:
Terminals
- Guardrails
- Median barriers
It should be noted that this list of roadside safety features is not all-inclusive and new features may be
developed that are not covered by this list. Current testing and evaluation procedures may not properly
address the critical conditions and impact performance for these new designs. Special tests and evalu-
ation criteria may, therefore, be needed for proper evaluation. Also, the list does not include baniers or
devices intended for other purposes, such as security barriers designed to stop impacting vehicles with
little regard for the occupant risk. Testing and evaluation requirements for such devices should refer to the
appropriate agencies, such as the U.S. Department of State oTASTM specifications.
Longitudinal barriers may be tested to six test levels and other roadside features may be tested to three
test levels. A test level is defined by impact conditions (speed and angle of approach) and the type of
test vehicle (ranging in size fi'om a small car to a fully loaded tractor-trailer truck), as summarized in
Table l- l. The first three test levels are limited to passenger vehicles while the last three incorporate
some form of heavy truck. Note that longitudinal barriers are the only safety features for which all
six test levels are defined at this time. All other safety features are designed exclusively for passenger
vehicles, such as automobiles and light trucks. A feature designed and tested for a low test level would
generally be used on a low-speed, low-volume, or both roadway such as a rural collector, local road, or
urban street. A feature designed and tested for a high test level ûould typically be used on a high-speed,
high-volume, or both, roadway, such as a freeway. It must also be noted that features that meet a given
test level will generally have different performance characteristics. Although a rigid barrier and a flex-
ible barrier can be designed to satis! a given test level, they will have different applications. The rigid
barrier will produce higher vehicle decelerations and prevent any lateral deflection while the flexible
barrier will ploduce lower accelerations, allow large lateral deflections, and be less likely to redirect the
impacting vehicle back toward the travelway. Further, there are different performance classifications for
some safety features, such as crash cushions. For example, a crash cushion can be designed to redirect
a vehicle impacting the side of the cushion (termed a redirective crash cushion), or it can be designed
to decelerate the vehicle to a stop when impacted on the side (tenned a non-redirective crash cushion).
Both designs can be made to satisfy a given test level.
While the guidelines were formulated purposely to offer the user considerable latitude in the design
and testing of a feature, it is not the pulpose nor is it within the purview of this document to detennine
where a feature, satisfying a given test level and having specific performance characteristics, would find
appropriate applications along the nation's roadways. That determination rests with the transporlation
agency responsible for the design, operation, and maintenance of the roadway.
Test Conditions
Test Test Vehicle
Level Designation'and Type Speed
mph (km/h)
1100C (Passenger Car) 31 (50) 25
1
2270P (Pickup Truck) 31 (50) 25
Concun'ent with the preparation of the first edition of this document, the European Committee for
Standardization (CEN) was preparing a sirnilar document for the European Union (EU). Developments
in both the United States and CEN were monitored, and every effort was made to harmonize the im-
pact perfornrance standards, e.g., using the sarne or similar testing conditions and evaluation criteria.
However, given the inherent diffelences in highway and traffic conditions between the United States
and the EU, differences between the U. S. guidelines and CEN standards are to be expected.
Design and development of a new safety feature is a complicated process in which full-scale crash test-
ing is used to demonstrate the satisfactory impact performance of the feature. During the early stages
of design and development, analytical and experimental tools are typically used to aid in the process,
including:
The initial design is typically developed using structural loading and design procedures based on the
principles of mechanics. Static tests are often conducted on certain critical components and connec-
tions to develop such data as ultimate capacity of the materials, strength of connections,load/deflection
characteristics, etc. Dynamic tests using a pendulum or bogie vehicle are used to test subsystems or
prototypes of the feature, e.g., to determine the energy absorption characteristics of a material under
dynamic impact conditions. Results from the static and dynamic testing are then incorporated into
computer models to simulate and evaluate the impact performance of the feature under varying condi-
tions, including parametric studies. The initial design is then modified based on results of the static and
dynamic tests and the computer simulation.
Note that designers/developers may differ in their approaches and may or may not use one or more of
these analytical and experimental tools, depending on the feature. Some features are relatively simple
to design or their characteristics are well-known from previous work such that the initial design can
be crash tested without any of these intermediate steps. Other features are very complicated and may
require the use of every tool available. These analytical and experimental too.ls can be invaluable to
the design and development process and should be used to the fullest extent possible. A more detailed
discussion on these analytical and experimental tools is presented in Appendix D.
Computer modeling using a finite element analysis code, such as LS-DYNA, has made significant
advances over recent years and is now a major tool in the development and testing of roadside safety
features. Computer modeling provides a means for developers of safety features to assess the impact
performance of safety features without actual crash testing. The evaluation can range from individual
components to subsystems, or to the entire system. V/hile computer modeling is gaining mol'e accep-
tance and reliance in the development and testing process, its effectiveness still depends heavily on the
expertise ofthe individualthat builds the models. Further, without extensive examination of a simu-
lation program's input parameters. it is impossible for another parry to critically evaluate the model's
accuracy. Therefore, it is prenrature at this time to consider replacing the crash testing recommended
herein with computer modeling to evaluate the impact performance of roadside safety features.
Chapter 2 outlines the test matrices and conditions recommended for testing and evaluating various
roadside safety features. It also presents recommended tolerances on impact conditions and proce-
dures to identiff the critical impact point for ceftain features. Chapter 3 describes the requirements for
construction of the test installations, including soil type and conditions. Chapter 4 describes the test
vehicles, specifications, and recommended instrumentation. It also identifies parameters that should be
measured before, during, and after the test. Chapter 5 presents the evaluation criteria used for assessing
test results. Chapter 6 recomnrends the manner in which a given test and its results are to be document-
ed. Chapter 7 contains guidelines on how in-service performance evaluation of a feature should be
conducted. Appendix A is a commentary on Chapters I through 7 and presents further elaboration
and discussion. Appendix B presents procedures for validating and conducting in-situ soil testing
procedures. Appendix C contains electronic and photographic instrumentation specifications, repro-
duced with permission from the Society of Automotive Engineers. Appendix D presents a number of
analytical and experimental tools. Appendix E presents techniques for measuring occupant compart-
ment and vehicle deformation. Appendix F presents procedures for calculating Theoretical Head Impact
Velocity (THIV), Post-Impact Head Deceleration (PHD), and Acceleration Severity Index (ASI) as
measures of occupant risk. Appendix G contains a proposed methodology for analyzing staged atten-
uation systems for mid-sized vehicle impacts. Appendix H outlines a procedure for re-evaluating and
selecting new test vehicles in response to changes in the vehicle fleet. A glossary of terms and a bibliog-
raphy complete the document.
Test Matrices
and Conditions 2
2.l GENERAL
limited series of crash tests are recommended to evaluate each class of roadside safety feature.
I
.{ \The purpose of this testing is to provide an established minimum level of performance for safety
features as well as a basis for comparing different designs within each class. Each of the recommended
tests is designed to assess one or more of the three principal evaluation criteria: occupant risk, vehicle
trajectory and structural adequacy. Further, as mentioned previously, the crash testing guidelines cannot
include all possible impact conditions that may be experienced in the real world and it is not practical
or feasible to test for all possible impact conditions. However, if a significant "window of vulnerability"
is identified for any given design, the test matrices should be supplemented to explore the additional
impact conditions.
The primary parameters that define a full-scale crash test include impact speed; impact angle, test
vehicle mass, and impact location. Each of these parameters is selected to represent a "worst practical
condition" for a roadside feature crash. For impact speed and angle, the "worst practical condition" has
been traditionally set at the 85th percentile level. Test vehicles are normally selected based upon vehicle
body style and weight. V/eights have generally been selected to approximate the 2nd and 90th percen-
tile levels for passenger vehicles. Impact locatjons on a safety feature are often selected to represent a
critical impact point (CIP) that creates the greatest probability of test failure. Hence, the combination of
85th percentile impact speed, 85th percentile impact angle, 5th and 95th percentile vehicle weights, and
critical impact point is believed to represent a worst practical condition.
Note that many safefy features, including guardrail terminals, crash cushions, and truck-mounted atten-
uators, are designed to accommodate end-on impacts. Crash testing has shown that high-angle, end-on
impacts on these features are much less likely to result in serious injury than impacts with lower angles.
In keeping with the spirit of the "worst practical condition" philosophy, impact angles for end-on crash
tests have been set well below the 85th percentile level.
'O 2016 by the American AssociaLion olSLate Highway antl Transportation Ollìcials
All rights reserved Duplication is a violation ofapplicable law
10 | Manual for Assessing Safety Hardware
these values were selected to represent the limiting impact conditions for high-speed, high-volume
roadways. For lower volume, lower speed facilities, or both, the impact speed was reduced from
62 mph (100 km/h) to 44 mph (70 krn/h) or 3l mph (50 km/h), as appropriate. In recognition of the
reduced cornering characteristics of large trucks, the irnpact angles for all large truck tests are reduced
to 15 degrees.Impact speeds for single-unit trucks have also been reduced to 56 mph (90 km/h) while
those for the heavy tractor-trajler type trucks are reduced to 50 mph (80 km/h) to further compensate
for the lower operating speed and reduced comering capabilities associated with combination vehicles.
Selection of the type, size, and weight of the test vehicles could have a significant bearing on the
magnitude of the impact associated with the crash tests. Both small and large passenger vehicles can
pose a significant and unique set of challenges for most types of roadside safety hardware. Evaluation
of 2002 vehicle sales (159) reveals that weights of all classes of passenger vehicles have risen signifi-
cantly since these guidelines were last revised (129).ln order to minimize the weight increase while
maintaining a large enough sales volume to assure ready availability of the test vehicles, the small car
test vehicle was selected to be representative ofthe 2nd percentile heaviest passenger vehicle (instead
of the traditional 5th percentile), and the light truck test vehicle was selected to be representative of
the 90th percentile (instead ofthe traditional 95th percentile) heaviest passenger vehicle. The selected
test vehicles are a sedan weighing approximately 2,420 lb ( l,l 00 kg) for the small car test vehicle
(designation 1100C) and a four-door, two-wheel drive, half-ton pickup truck weighing 5,000 lb
(2,270 kg) for the light truck test vehicle (designation 2270P).An altemative test vehicle, designated
as 15004, was also identified for use in evaluating the impact performance of staged energy absorb-
ing systems. The 15004 test vehicle is a four-door passenger sedan weighing approximately 3,300 lb
(1,500 kg). Three heavy truck test vehicles-a22,000-lb (10,000-kg) single-unit truck, an 80,000-lb
(36,000-kg) tractor-van-trailer combination, and an 80,000-lb (36,000-kg) tractor-tank-trailer
combination-have also been identified for evaluation of higher performance baniers. These vehicles
are designated as 100005, 36000V and 36000T, r'espectively. Detailed specifications for the test vehi-
cles are presented in Chapter 4.
Impact locations for most full-scale crash tests are selected to represent the critical condition that
would be most likely to lead to test failure. For longitudinal barriers, critical impact points (CIPs)
are selected to maximize loading at rail splices and maxirnize the potential for wheel snag and vehi-
cle pocketing. Note that any splice connection expected to be used in the field must be implemented
in the critical region during full-scale crash testing. Critical impact points for post-and-beam type
barriel'terminals and crash cushions are selected to represent the point where the system is believed to
transition from gating to redirective behavior. General guidelines for selecting CIP locations for each
class ofsafety feature are described in Section 2.3. V/here possible, testing agencies are encouraged
to utilize more detailed analysis, such as computer simulation, to estimate the CIP location for each
full-scale crash test. Detailed procedures for estimating CIP locations for roadside safety features are
presented elsewhere (125) and will not be repeated herein.
In some circumstances, longitudinal barriers (e.g., flexible cable baniers) may be positioned within
sloped medians in order to shield motorists from obstacles and/or prevent cross-median crashes. For
these situations, longitudindal barriers should be crash tested and evaluated with the barriers installed in
representative sloped median terrain. When barriers are placed on sloped terrain, such as a flexible cable
barrier placed within a median ditch, the target impact conditions (i.e., speed and angle) should be ref-
erenced to the time when the vehicle reaches the front slope break point (SBP). For most barriers placed
on the front slope of depressed median ditches, vehicles will impact barriers with speeds and angles
similar to those exhibited at the front SBP, except for minor speed changes attributed to free fall while
extended over the front slopes prior to barrier contact and aerodynamic drag. For barriers positioned
on the back slope of depressed median ditches, there is an increased likelihood of deviations in vehicle
impact speed and angle due to vehicle-ground contact as the vehicle traverses through the ditch prior
to barrier contact. For these events, testing laboratories should also document, to the extent possible,
actual impact conditions as the vehicle strikes the barrier.
The severity of an impact is normally measured in terms of impact severity (IS) for crash tests involving
vehicle redirection, and kinetic energy (KE) for crash tests involving end-on impacts or breakaway de-
vices. IS, as defined in Equation 2-1, has been shown to be a good indicator of the magnitude of loading
on a longitudinal barrier.
Where
IS impact severity, kip-ft (kJ)
M vehicle mass, lb (kg)
V impact speed, ft/s (m/s)
0 impact angle, degrees
Total vehicle kinetic energy (KE), as defined in Equation 2-2,is considered a better measure of the se-
verity of all head-on or end-on impacts, including tests of breakaway devices, crash cushions, terminals,
and truck-mounted attenuators. Note that KE is also used as the measure of crash severity for oblique
impacts on the ends of terminals and crash cushions.
'Where
KE kinetic energy, kip-ft (kJ)
Even when test speeds and impact angles are within the acceptable tolerances, the IS or KE values of
a crash test can be unacceptably low. For this reason, an additional limiting condition is applied to the
IS and KE values for full-scale crash tests. IS values for tests involving vehicular redirection and KE
values for high-speed tests involving end-on impacts must be no more than 8 percent below'the target
values. For most full-scale crash tests, excessive impact speeds and angles do not improve the likeli-
hood ofa successful test. Therefore, excessive speed and angles are not considered to be a cause for
failing these tests, provided all impact performance evaluation criteria are met. The exceptions to this
general rule are the low-speed tests, i.e., l9-mph (3O-km/lr) tests for support structures and work-zone
traffic control devices, and 3l-mph (50-km/h) test for breakaway utility poles. Since these low-speed
tests are intended to evaluate the kinetic energy required to activate the breakaway, fracture, or yielding
mechanism of the device, it is recommended that the KE value should be no greater Than2} percent
above the target value. Tests that do not fall into the acceptable ranges for IS or KE are considered to
be invalid and should be repeated. Acceptable ranges for IS and KE values are listed in the crash test
description tables in Section 2.2. Note that limiting IS and KE values are calculated based upon the
vehicle test inertial mass and exclude the weight of loose ballast or dunrmies used in the test.
Although test vehicles are selected to have the appropriate mass, it is sometimes impossible to adjust
vehicle weight to exactly match target values. Excessive vehicle mass can increase barrier loadings
and can sometimes enhance vehicle stability. Similarly, inadequate vehicle mass can reduce barrier
loading and possibly decrease vehicle stability. Therefore, upper and lower limits have been estab-
lished on vehicle gross static mass, as shown in Table 2- I . Note that ballast can be added to increase
test vehicle mass and some vehicle components can be removed to decrease mass. Details of vehicle
mass adjustment procedures are presented in Chapter 4.
Impact locations are normally selected to maximize the risk of test failure. Obviously, large errors
in the impact location can dramatically affect the safety feature performance. Most vehicle guidance
systems used in roadside safety testing have limited levels of accuracy. Based on a suryey of crash
testing laboratories, acceptable variations in target impact location have been established: l2 in.
(300 mm) for side impacts and 6 in. ( I 50 mm) for frontal collisions. In other words, the recommended
maximum acceptable tolerance for the impact point is +12 in. (300 mm), as measured along the face
of the barriers, for tests of longitudinal barriers and sides of terminals or crash cushions. Similarly,
test vehicles involved in frontal impacts, including tests of crash cushions, terminals, and structural
supports, should contact the test article within +6 in. (150 mm) of the target impact point as measured
along the front of the test vehicle. Recommended impact points for all tests are presented in Section
2.3.
TABLE 2-1. Vehicle Test lnertial Mass Upper and Lower Limits
Test Vehlcle Deslgnatlon Target Vehlcle Weight, Acceptable Varlatlon,
and Type lb (kg) lb (ks)
1100C (Passenger Car) 2,420 (1J00) r55 (25)
Roadside baniers are generally classified according to stiffness into one of three categories:flexible,
semi-rigid, and rigid. Although the test matrix is the same for all three classifications of barrier, some
consideration should be given to the type of barrier when constructing the test installation. Height
variations during construction can adversely affect barrier performance. Increased mounting height
is most likely to adversely affect barrier perfonrance for small car impacts while reduced height is
considered most likely to affect barrier performance for light truck impacts. Thus, for barrier systems
with large allowable variations in mounting heights, small car tests should be conducted with barrier
rail elements installed at the maximum acceptable height and light truck tests should be conducted
with rail elements at the minimum acceptable height. Refer to Section 3.4.2 for specific installation
details of various barriers.
When shielding motorists from roadside or median obstacles, it may be desired or necessary to install
longitudinal bariers on roadside or median slopes. For example, it is not uncommon for flexible
cable barrier to be placed within a depressed median. In these circumstances, full-scale crash testing
should also be performed on barriers installed on or near sloped terrain representative ofactual field
conditions. Toward this goal, test matrices have been developed for evaluating cable barriers placed
in depressed medians, more specifically symmetric V-shaped ditches. Successful testing and eval-
uation of cable barriers installed in V-ditches should allow for their placement in other less critical
ditch configurations, such as stepped medians or flat bottom trapezoidal ditches. It should be noted
that these test matrices were developed only for Test Level 3 (TL-3) applications. If a Test Level 4
(TL-4) cable barrier system is desired, it is recommended that any additional crash testing with large
single-unit trucks be performed with the cable barrier installed on level or relatively flat terrain. These
matrices were primarily developed based on research results obtained from a vehicle trajectory study
of simulated small car, mid-size sedan, and light-truck passenger vehicle encroachments into 4H:lV
and 6H: I V V-ditches at a speed of 62 mph and at an angle of 25 degrees (95). Further refinement of
the test matrices occurred using results from prior crash tests of cable barriers placed in V-ditches and
input from roadside safety experts and test laboratories.
For 4H:lV and 6H:lV median ditches, crash test matlices are provided for evaluating: (1) a single
cable baruier system placed anywhere within a median ditch; (2) a single cable barrier system placed
between 0 to 4 ft beyond the front or back slope break point (SBP); and (3) two cable barier systems
(i.e., double cable barrier) placed within a median ditch, each 0 to 4 ft away from a SBP. For single
and double cable barrier systems placed at 0 to 4 ft away from a SBP, the test matrix is a subset of
the matrix developed for placement of a cable barrier system anywhere within the ditch. Guidance
for determining the appropriate subset of required crash tests is provided below. For cable barriers
installed in 4H:lV depressed medians, a 46-ft wide ditch is recommended for the crash testing pro-
gram. A 3O-ft wide ditch is recommended for evaluating cable barriers installed in 6H: lV depressed
medians. These widths were determined based on vehicle trajectory and kinematics through the ditch
derived from vehicle dynamics simulations and crash testing experience. Further details for crash test-
ing cable barriers installed on slopes or in ditches are provided below. Test matrices for the selected
cable barrier-ditch combinations are summarized in Tables 2-28 through2-28. Recommended cable
ban'ier placement for testing and evaluating cable ban'ier systems in 4H:lV and 6H:lV V-ditches is
provided in Figures 2-2A and2-28, respectively.
lmpact
V-Ditch Accept-
Test Gonditions Barrler Barrier Critical
Veh. able Eval.
Desig.
Type Speed, W¡dth Posi- Loca- lmpact
Angle lS Range,b Criteriac
No. mph (ft) tion tiona Point
(deg) k¡p-ft (kJ)
fkm/hl
Level Midspan
3-10 1100c 62 (100) 25 NAd NAd >51 (69.7) A,D,F,U,t
Terrain Location
1ft
Level
3-11 2270P 62 (100) 25 NAd NAd Upstream >106 (144) A,D,F,H,I
Terrain
from Post
12ft
1ft
Front from
3-13 2270P 62 (100) 25 46 Upstream >106 (144) A,D,r,H,l
Slope Front
from Post
SBPE
12ft
Front from Midspan
3-14 1100c 62 (100) 25 46 >51 (6e.7) A,D,F,H,I
Slope Front Location
SBPE
4 ft from
Back Midspan
3-15 1100c 62 (100) 25 46 Ditch >51 (69.7) A,D,F,H,I
Slope Location
Bottom
4 ft from
Back Midspan
3-'16 1100c 62 (100) 25 46 Back >51 (69.7) A,D,F,H,I
Slope Location
SBPE
Front Midspan
3-17 15004 62 (100) 25 46 Variabler >70 (95.1) A,D,ñH,l
Slope Location
Back
I ft from 1ft
3-18 2270P 62 (100) 25 46 Back Upstream >106 (144) A,D,F,H,I
Slope
SBPE from Post
a See Figure 2-2\'1or banier placement.
b See Section 2.1 .2lo¡ lolerances on impact conditions.
c See Table 5-'1.
d Not applicable.
Slope break point.
f Tesüng laboratory should determine critical barrier position on front slope of ditch in order to max¡mize propensity for front end
of 1500Avehicle to penetrate belween vertically adjacent cables. Critical factors may include vertical cable spacing, position
of cables relative to front bumper, location and type of cable release mechanisms, trajectory of vehicle's front bumper, etc.
TABLE 2-2C. Recommended TL-3 Test Matrix for Single or Double Median Barrier Desiglned for
Placement between O to 4-ft Offset from Slope Break Point of 4H:1V V-Ditch
lmpact
V-Ditch Accept-
Test Conditions Barrier Barrler Critical
Veh. able Eval.
Desig. Speed, w¡dth Posi- Loca- lmpact
Type Angle lS Range,b Criteriac
No. mph (ft) tion tiona Point
(des) kip-ft (kJ)
fkm/h)
Level Midspan
3-10 11 00c 62 (100) 25 NAd NA >51 (69.7) A,D,F,H,I
Terrain Location
1ft
Level
3-11 2270P 62 (100) 25 NAd NA Upstream >106 (144) A.D,F,H,I
Terrain
from Post
4 ft from 1ft
Front
3-13 2270P 62 (100) 25 46 Front Upstream >106 (144) A,D,F,H,I
Slope
SBPE from Post
4 ft from
Front Midspan
3-14 1100c 62 (100) 25 46 Front >51 (69.7) A,D,F,H.I
Slope Location
SBPE
4 ft from
Back Midspan
3-15f's 1100c 62 (1oo) 25 46 Ditch >51 (69,7) A,D,F,H,I
Slope Location
Bottom
4 ft from
Back Midspan
3-16s 1100c 62 (100) 25 46 Back >51 (69.7) A,D,F,H,I
Slope Location
SBPE
Back
I ft from 1ft
3-1 8s 2270P 62 (100) 25 46 Back Upstream >106 (144) A,D,F,H,I
Slope
SBPi from Post
TABLE 2-2D.Recommended TL-3 Test Matrix for Single Median Barrier Designed for Placement
Anywhere in 6H:1V V-Ditch
lmpact
V-Ditch A,ccept-
Test Conditions Barrier Barrier Critical
Veh. able Eval.
Desig
Type Speed, width Posi- Loca- lmpact
Angle lS Range,b Griteriac
No, mph (ft) tion tiona Point
(des) kip-ft (kJ)
ftm/hì
Level Midspan
3-10 1100c 62 (100) 25 NAd NAd >51 (69.7) A,D,F,H,I
Terrain Location
1ft
Level
3-11 2270P 62 (100) 25 NAd NAd Upstream >106 (144) A,D,F,H,I
Terrain
from Post
9 ft from 1ft
Front
3-13 2270P 62 (100) 25 30 Front Upstream >106 (144) A,D,r,H,l
Slope
SBPE from Post
9 ft from
Front Midspan
3-14 1100c 62 (100) 25 30 Front >51 (69.7) A,D,ñH,t
Slope Location
SBPE
4 ft from
Back Midspan
3-1 5rs 1100c 62 (100) 25 30 Ditch >51 (6e.7) A,D,F,H,I
Slope Location
Bottom
1 ft from
Back Midspan
3-1 6s 11 00c 62 (100) 25 30 Back >51 (69.7) A,D,F,H,I
Slope Location
SBPE
Front Midspan
3-17 15004 62 (100) 25 30 Variabler >70 (95.1) A,D,F,H,I
Slope Location
1ft
Back Back
3-1 8e 2270P 62 (100) 25 30 Upstream >106 (144) A,D,F,H,I
Slope SBPE
from Post
TABLE 2-2E.Recommended TL-3 Test Matrix for Single or Double Median Barrier Designed for
Placement between O to 4-ft Offset from Slope Break Point of 6H:1V V-Ditch
lmpact
V-Ditch Accept-
Test Conditions Barrier Barrier Critical
Veh. able Eval.
Desig Speed, width Posi- Loca- lmpact
No.
Type Angle lS Range,b Criteriac
mph (ft) tion tiona Point
(des) k¡p-ft (kJ)
lkm/hl
Level Midspan
3-1 0 1100c 62 (100) 25 NAd NAd >51 (69.7) A,D,F,H,I
Terrain Location
1ft
Level
3-1',\ 2270P 62 (100) 25 NAd NAd Upstream >106 (144) A,D,r,H,l
Terrain
from Post
9 ft from 1ft
Front
3-1 3 2270P 62 (100) 25 30 Front Upstream >106 (144) A,D,r,H,l
Slope
SBPE from Post
9 ft from
Front Midspan
3-14 1100c 62 (100) 25 30 Front >51 (69.7) A,D,F,H,I
Slope Location
SBPE
4 ft from
Back Midspan
3-1sf's 1100c 62 (100) 25 30 Ditch >51 (6e.7) A,D,F,H.I
Slope Location
Bottom
1 ft from
Back Midspan
3-1 6s 1100c 62 (100) 25 30 Back >51 (69.7) A,D,F,H,I
Slope Location
SBPE
Front Midspan
3-17 15004 62 (100) 25 30 Variabler >70 (95.1) A,D,r,H,l
Slope Location
1ft
Back Back
3-1 8e 2270P 62 (100) 25 30 Upstream >106 (144) A,D,r,n,l
Slope SBPE
from Post
Lêngth ofTest
Reference
Critical lmpact Poinl (ClP)
End Anchor
TEST O (DEG)
t0 25
11 25
12 15
Transltio n
Sliffer Barrier
End Anchor
TEST O(DEG)
--ro- 25
21 25
22 15
)
TEST20,21 and22
I
B..kSP FrentsBP K SBP
T€sl 18
1227oPl
[See Note 4]
NOTES: 1. For bârrier placement anlMhsrs, use e I 2-fr lâteral ofisel. Othem¡se, use a 4-ft lateral ofset for barr¡er plaæment w¡th¡n 0 to 4 ft on front SBP.
2. For single or doubla median barier placement at 0 to 4 n oÍset from SBe Test No. 15 ¡s unnecessary for V-dilches greater lhan or €qual to
26 ff, as m€asured from front SBP to beck SBP.
3. Testing leboratory should dêtermina criticel berier position from 0 to 'X' on front slope of dltch or on level terrain ln order to mâx¡mize the
propensity lo¡ the front end of 15004 vehlcl€ lo penetrate between vertlcally ad¡acent cables. Critlcal fectors may lnclude verllæl ceble spaclng,
posiùon of cables relatlve to front bumper, locatlon and lype of cable releese mechanisms, têjectory of vehicle's front bumper, etc.
4. A 46-ft wide dltch was selôctod to 6impl¡ñ/ the test matrix, thus resulling ¡n barisr placement beyond the 0 to 4 ft Enge. Howêver, narower ditch
w¡dths prov¡de slmllar risks for ovôBide wilh a barrior placed al the back SBP.
I
-T1
&d SBP Aå4SP Fbnl SAP
MSBP
T€st 18
1227oPl
P6r@ k saÞ
&ck s8Þ
NOTES: 1, For bani€r placemenl. anywhere, use a g-fr laleEl offsel. OtheN¡se, use a 4-ft latefal offset for barier placement w¡lhin 0 to 4 fr on front SBP,
2. For single or double median barrier placement at 0 to 4 n offset from SBB Tesl No. 1 5 ¡s unnecessary for V-ditches greater than or equal to
24 fr, as measured from front SBP to back SBP.
3. Testlng laboratory should detem¡ne cítlcal barier pos¡tion from 0 to "X' on ftont slope ot dilch or on level terain in order to max¡m¡ze lhe
propensity fo¡ lhe front end of 15004 vehicle to penelãte between verlically adjacent cables, Crit¡cal lactoß may include vertiæl cable spac¡ng,
posit¡on of cables relative to fronl bumper, lo€t¡on and type of cable release mechanisms, tra¡ectory of vehicle's front bumper, elc.
TEST 10
Test l0 is designed to investigate a barrier's ability to successfully contain and redirect small pas-
senger vehicles impacting within the length-of-need. For small cars, the primary concems are the
potential for vehicle under-ride, wheel snag, rollover, and head-slap. For flexible cable barriers,
additional concerns include cable interaction with and damage to the A-pillar, windshield, and roof.
In order to evaluate the full range of barrier impact performance, testing agencies should consider
installing the barrier at the maximum allowable height for small car tests. This is especially true for
post-and-beam barrier systems to maximize the risk of under-ride and wheel snag.
TESTS ll
and2l
Tests I I and 2l provide maximum strength tests for Test Levels I through 3 and veriff a barrier's
performance for impacts involving light trucks and SUVs for all test levels. Due to the high rollover
frequencies observed in crash data ancl during historical full-scale crash testing with light truck vehicles,
Tests I I and 21 are required for all banier systems. For flexible cable'ban iers, the primary concems
include vehicle containment, vehicle stability, A-pillar integrity, and working width. These tests are
now required to also meet all occupant risk measures, including both lateral and longitudinal occupant
impact velocity (OIV) and ridedown acceleration (RA) values.
TEST 20 (Optional)
Test 20 for a transition section is an optional test to evaluate the occupant risk and post-impact trajec-
tory criteria for all test levels. It should be conducted ifthere is reasonable uncertainty regarding the
impact performance of the system for impacts with small passenger vehicles.
TESTS 12 and22
Tests l2
and22 are conducted for Test Levels 4, 5, and 6. These tests are intended to evaluate the
strength of the barrier in containing and redirecting heavy trucks.
TEST 13
Test l3 is designed to assess a cable barrier's ability to contain and redirect light trucks and SUVs as
well as prevent barrier override within the length-of-need when placed on a roadside slope or front
slope of a median ditch. The 2270P vehicle provides a critical test due to its large mass moment of
inefiia, high center of mass, and high bumper trajectory relative to the ditch ground surface. For cable
barriers designed to be installed anywhere in a median ditch, the cable ban ier shall be placed 12 and 9
ft away from the fi'ont SBP of a 46-ft. wide 4H: lV V-ditch and a 30-ft wide 6H: lV V-ditch, respectively
For cable barriers intended for use between 0 to 4 ft away from the slope break point, the barrier shall
be placed 4 ft away from the front SBP of the 4H: lV or 6H:lV V-ditch.
TEST 14
The primary objective for Test l4 is to assess a cable barrier's ability to safely contain and redirect
small passenger vehicles without resulting in excessive vehicular instabilities and/or rollover when
installed on a roadside slope or front slope of a median ditch. The risk for rollover may be aggravated
by several factors, including: (l) roll and pitch rotations induced into the airborne vehicle prior to con-
tacting the cable barrier; and (2) instabilities resulting from vehicle contact with the cable barier while
airborne. For cable barriers installed anywhere in a median ditch, the barrier shall be placed 12 and 9
ft away from the front SBP of a 4H: I V and a 6H: I V V-ditch, respectively. For cable barriers intended
for use between 0 to 4 ft away from the slope break point, thE barrier shall be placed 4 fr. away from the
front SBP of the 4H:lV or 6H:lV V-ditch.
TBST 15
Test l5 is designed to assess a cable barrier's ability to safely contain and redirect small passenger ve-
hicles as well as prevent barrier underride, component penetration into the occupant compaftnrent, and
excessive deformations of the A-pillar, roof, or windshield. For cable barriers installed anywhere in a
median ditch, the barrier shall be placed 4 ft up the back slope from the bottom of the 4H:lV or 6H:lV
ditch. For cable bariers intended for use between 0 to 4 ft away from the slope break point, Test l5 is
not required for 4H:lV and 6H: lV V ditches greater than or equal to 26 and24 fr.wide, respectively,
as measured from front SBP to back SBP. If Test l5 is required for cable barriels intended for use
between 0 to 4 ft away from the slope break point due to the likelihood of installation within narrow
V-ditches, it is recommended that testing be performed using the placement guidelines provided for
barriers installed anywhere within a median ditch. Further, Test l5 is not required for evaluating a
double cable barrier system that has one barrier placed on each side ofa median ditch, each from 0 to
4 ft away from a SBP.
TEST 16
Test l6 is designed to barrier's ability to safely contain and redirect small passenger
assess a cable
vehicles after traveling across the center ofa ditch and up the back slope. In previous full-scale crash
tests of cable barrier systems placed on the backslope of a V-ditch, small passenger vehicles have
shown tendencies to either achieve an increased impact angle, acquire a yaw velocity, or rebound off
of the ditch surface and become airborne. These vehicle behaviors may lead to an increased propen-
sity for occupant compartment deformation and penetration, vehicular instability, rollover, and/or
barrier overide. For 4H:lV V-ditches, the cable banier shall be placed 4 ft away from the back SBP
for both a barrier designed for placement anywhere in the ditch or intended for use between 0 to 4 ft
from the SBP. For 6H: lV V-ditches, the cable barrier shall be placed I ft away from the back SBP
for both a barrier designed for placement anywhere in the ditch or intended for use between 0 to 4 ft
from the SBP. Test 16 is not required for evaluating a double cable barrier system designed to have a
separate cable barier installation on each side of the ditch placed from 0 to 4 ft away from each SBP.
TEST 17
Test l7 is designed to evaluate a cable balrier's ability to contain and redirect mid-size passenger
sedans by preventing vehicle penetration through veftically adjacent cables. Although vehicle penetra-
tion through the barrier is the primary concern, this test also assesses risk associated with component
penetration into the occupant compartment and excessive deformations to the A-pilla¡ windshield,
or roof. For Test 17 and as determined by the testing Iaboratory, the critical barrier placement will
occur on the front slope and range between 0 ft (i.e., front SBP) and some lateral offset away from
the front SBP. For cable barriers installed on mostly level terrain or adjacent to steep roadside slopes
(i.e., steeper than 3H:lV), Test l7 is recornmended for evaluating the risk for passenger vehicles to
penetrate between cables depending on banier configuration (i.e., cable spacing. cable heights, etc.).
Critical barrier placement will consider vehicle and ban'ier geometries and maximize tlie propensity
for the front end ofthe I 5004 vehicle to penetrate between adjacent cables. The objective is to po-
sition the front bumper between the adjacent cables with the widest veftical cable spacing at time of
impact. If multiple cable spacings have a similar maximum vertical spacing, the lowest of this group
should be targeted to maximize the number of cables potentially in contact with the A-pillar, wind-
shield, and roof during redirection.
TEST 18
Test l8 is designed to assess a cable banier's ability to safely contain and redirect light trucks and
SUVs after traveling across the center of a ditch and up the back slope. ln previous full-scale crash
tests of cable barrier systems placed on the backslope of a V-ditch, passenger vehicles have shown
tendencies to either achieve an increased impact angle, acquire a yaw velociry or rebound off of the
ditch surface and become airbome. These vehicle behaviors can lead to an increased propensity for
vehicular instability, rolloveç and/or barrier override. For 4H: I V V-ditches, the cable barrier shall be
placed 8 ft away from the back SBP for both a barrier designed for placement anywhere in the ditch or
intended for use between 0 to 4 ft from the SBP. For 6H: lV V-ditches, the cable barrier shall be placed
at the back SBP for both a barrier designed for placement anywhere in the ditch or intended for use
between 0 to 4 ft from the SBP. Test l8 is not required for evaluating a double cable barrier system
designed to have a separate cable barrier installation on each side ofthe ditch placed from 0 to 4 ft away
from each SBP.
Figure 2-l displays impact conditions for Tests 10, 11, 12,20,21, and22. As shown in this figure,
critical impact points are measured from a potential snag or fracture location as well as to maximize the
propensity for underride, override, or penetration offlexible bariers such as cable barriers installed on
a roadside slope or in a median ditch. In some cases, such as transition systems discussed previously, a
barrier may have two or more critical impact locations. In this situation, testing agencies should con-
sider testing each of the locations to veriff acceptable barrier performance. Critical impact locations for
each class ofsafety feature are defined in Section 2.3.
These differences arise primarily due to the classification of crash cushions as either gating or non-gat-
ing. Gating crash cushions are designed to allow vehicles impacting near the beginning or nose of the
system to safely pass through the unit and travel behind the cushion. Non-gating crash cushions are
designed to capture almost all vehicles striking the end of the device and safely decelerate them to a
stop. A non-gating crash cushion must capflrre impacting vehicles during angular tests on the end of the
system.
Non-redirective crash cushions are designed to safely accommodate most vehicles striking the front of
the cushion, but have no capability to redirect vehicles impacting near the rear of the device. As a result,
most non-redirective cushions are designed to be wider than the hazard to be shielded and are typically
used farther from traffic where the risk of high-energy impacts near the rear of the cushion is lower. It is
the responsibility of user agencies to detennine where safety features addressed in this document have
application, including redirective and non-redirective crash cushions.
Impact attenuating traffic gates, referred to as resistance gates, are beginning to be placed in front of
drawbridges , af-grade railroad crossings, or other locations whele penetration of the gate could lead
to a high risk of a severe crash. Because they are designed to attenuate head-on impacts and have
no redirective capacity, resistance gates are considered a subset ofnon-redirective crash cushions.
Resistance gate systems are designed to block the wide areas at locations where they are deployed. As
a result, most impacts involve vehicles striking the gate at angles near 90 degrees.Minor reductions
in the impact angle from 90 to 75 degrees are not believed to increase the likelihood of test failure.
Hence, impact attenuating gates are not subjected to low-angle, head-on tests included in the non-re-
directive crash cushion test matrix. Further, impact attenuating gates cannot be struck along the side
of the system and, thus, the crash test along the side of non-redirective crash cushions is also waived.
Therefore, impact attenuating gates need only be subjected to Tests 40 and 4l . Existing designs are
believed to have only one attenuation stage and therefore Test 45 should not normally be necessary.
Further, the impact point for all of tests is recommended to be at the quafter point along the face of
the system. This impact point should explore asymmetric loading of the attenuation system and be
representative of typical impact conditions for a two-lane gate design.
Some longitudinal barriers, such as flexible cable baniers or concrete median barriers, may be
installed on roadsidefill slopes or within median ditches. However, termination of these systems typi-
cally occurs on generally flat terrain. Consequently, the crash testing and evaluation of end treatments
for these systems, such as crash cushions and guardrail end terminals, shall continue to be performed
with those systems installed on flat, level terrain unless new information suggests procedural changes
are in order to be more representative of actual field conditions.
Recommended tests for evaluating the impact performance of terminals and crash cushions are
presented in Table 2-3.lmpact conditions are presented in Figure 2-3Afor teminals and redirective
crash cushions and Figure 2-38 for non-redirective crash cushions. These guidelines are applicable to
both permanent features and temporary features used in work or construction zones. Reference should
be made to the Glossary for definitions of these features.
@ 2016 by the American Association ol" State l-lighway antl Transportation Ofl'icials.
All rights reserved. Duplication is a violation ofapplicable law
Chapter 2-Test Matrices and Conditions | 27
TABLE 2-3. Recommended Test Matrices for Terminals and Crash Cushions
> 148
G/NG 1-31 2270P 31(50.0) 0 KE (e) C,D,F,H,I,N A,D,F,H,I
(202.0)
>148
G/NG 1-33 2270P 31(50.0) 5-'15 KE (e) C,D,F,H,I,N A,D,F,H,I
(202.0)
Termi-
nals and
G/NG 1-34 1100c 31 (50,0) 15 IS >s (6.5) (e,g) C,D,F,H,I,N A,0,nH,l
Redirec-
tive Crash
Cushions G/NG 1-35 2270P 31 (50.0) 25 IS >27 (36.0) (e) A,D,F,H,I A,D,F,H,I
Non+e- >116
(j 1-42 1100c 31 (50.0) 5-1 5 KE (f) C,D,F,H,I,N
directive (158.0)
Crash
G 1-43 2270P 31 (50.0) 5-1 5 KE >26 (35,6) (0 C,D,F,H,I,N
Cushions
a G/NG-Test applicable to gating and non-gating devices. G-Test applicable to gating devices.
b See Section 2.1.2 fortolerances on impact conditions.
c See Table 5-1.
d See Equations 2-1 and2-2.
e See Figure 2-34 for impact point.
f See Figure 2-38 for impact point.
g See Section 2.3.3 for impact point.
TABLE 2-3. Recommended Test Matrices for Terminals and Crash Cushions (continued)
>141
G/NG 2-30 1100c 44 (70.0) 0 KE (e) C,D,F,H,I,N A,D,F,H,I
(1e1 .0)
>291
G/NG 2-31 2270P 44 (70.0) 0 KE (e) C,D,F,H,I.N A,D,F,H,I
(3e5.0)
>141
G/NG 2-32 1100c 44 (70.0) 5-1 5 KE (e) C,D,F,H,I,N A,D,F,H,I
(1 e1.0)
Termi- >291
G/NG 2-33 2270P 44 (70.0) 5-1 5 KE (e) C,D,F,H,I,N A,D,F,H,I
nals and (3e5.0)
Redi-
G/NG 2-34 11 00c 44 (70.0) 15 IS >9 (12.8) (e,s) C,D,F,H,I,N A,D,F,H,I
rective
Crash
Cush- G/NG 2-35 2270P 44 (70.0) 25 ls >52 (70.5) (e) A,D,F,H,I A,D,F,H,I
ions
>192
G/NG 2-38 15004 44 (70.0) 0 KE (e) C,D,F,H,I,N A,D,F,H,I
(261.0)
>,l06
G 241 2270P 44 (70.0) 0 KE (f) C,D,F,H,I,N
(144.0)
Non+e- >228
G 242 '1100c 44 (70,0) 5-1 5 KE (f) C,D,F,H,I.N
directive (30e.0)
Crash
Cush- G 243 2270P 44 (70.0) 5-1 5 KE >51 (69.7) (0 C,D,F,H,I,N
t0ns
¿106
Lt 2-44 2270P 44 (70.0) 20 KE (f) C D,F,N
('r44.0)
>192
G 2-45 1 5004 44 (70,0) 0 KE (f) C,D,F,H,I,N
(261.0)
a G/NG-Test applicable to gating and non-galing devices. G-Test applicable to gat¡ng devices.
b See Section 2.1 .2 for tolerances on impact conditions.
c See Table 5-1.
d See Equations 2-'l and 2-2.
e See Figure 2-34 for impact point.
f See Figure 2-38 for impact point.
g See Section 2.3.3 for impact po¡nt.
) TABLE 2-3. Recommended Test Matrices for Terminals and Crash Cushions (continued)
62
G/NG 3-31 2270P 0 KE >594 (806) (e) C,D,F,H,I,N A,D,F,H,I
(1 0o.o)
o¿
G/NG 3-32 1100c 5-1 5 KE >288 (390) (e) c,D,lH,l,N A,D,F,H,I
(100.0)
62
G/NG 3-33 2270P 5-1 5 KE >594 (806) (e) c,D,F,H,l,N A,D,F,H,I
Terminals (100.0)
and Re- 62
directive
G/NG 3-34 1100c 15 IS >19 (26) (e,s) C,D,F,H,I,N A,D,F,H,I
(100,0)
Crash
62
Cushions G/NG 3-35 2270P 25 IS >106 (144) (e) A,D,F,H,I A,D,r,X,t
(100.0)
62
G/NG 3-36 2270P 25 IS >106 (144) (e,s) A,D,r,H,l A,D,F,H,I
(100.0)
62
G/NG 3-38 1500A 0 KE >392 (532) (e) C,D,F.H,I,N A,D,F,H,I
(100.0)
62
G 3-40 1100c 0 KE >105 (142) (f) C,D,F,H,I,N
(100.0)
62
G 341 2270P KE >216 (294) (f) C,D,F,H,I,N
(100.0)
Non-re- 62
s 342 1100c 5-1 5 KE >465 (631) (f) C,D,F,H,I,N
directive (100.0)
Crash o¿
Cushions
G 343 2270P 5-1 5 KE >105 (142) (0 C,D,F,H,I,N
(100.0)
62
\t 344 2270P 20 KE >216 (294) (f) C,D,F,N
(100.0)
62
\) 3-45 15004 0 KE >392 (532) (f) C,D,F,H,I,N
(100.0)
a G/NG-Test applicable to gating and non-gating devices. G-Test applicable to gating devices.
b See Section 2.1.2 for tolerances on impact conditions.
c SeeTable 5-1.
d See Equations 2-1 and2-2.
e See Figure 2-34 for impact point.
f See Figure 2-38 for impact point.
g See Section 2.3.3 for impact point.
I
lm p act Angle Normal Direction
e (deq.) TerminalTvoe of Travel
5-1 5 G ating
15 N on-G ating
rESTS 32 AND 33
Figure 2-3A. lmpact Conditions for Terminal and Redirective Crash Cushion Tests
TEST g (DEG)
34 15
3s 25 TEST 35 TEST 34
sEE SECTTON 2.3.3.1 e sEE SECTION 2.3.3.2
FOR DETERMINATION OF CIP FOR DETERMINATION oF ctP J
Normal Direction
Normal Direction of Travel
of Travel
TEST 34 AND 35 TEST 36
ut.
..--.............._Normal Direction
Normal Direction of Travel
of Travel t¡)
o
=.
TEST 37 (FOR MED¡AN DEVTCE) TEST 37 (FOR ROADSTDE DEVTCE) o
(t
0)
=
o-
C)
o
o
=
NOTE: Recommended Tolerance on lmpact -.
o
Point in All Side lmpacts = t300 mm f,
Ø
at)
Figure 2-34. lmpact Conditions for Terminal and Redirective Crash Cushion Tests (continued) Þ
L L Ì
o
J
c
s¿.
e.
<
E
õ _L
T o
Ø
o
o
- al,,
J
f¡q
U)
0)
(D
0=0DEG.
Y=OFFSET=W4
Nomal Dlrestlon
of TEfüc
0=0DEG.
OFFSET = 0
Normal Direction
ofTeffic
-
f¡)
o.
É
0)
TEST 40 TESTS 41 AND 45 o
L L
Y
e.
ì
e
lmpact
Ì
e = 15 DEG. Normal D¡recl¡on
Point (ClP)
OFFSET = 0 of Tlafi¡c
0 = 20 DEG. . \ Nomal D¡rect¡on
of Trafic
TESTS 42 AND 43
TEST 44
These tests are conducted with the vehicle traveling parallel to the roadway and the center of the vehicle
offset one quarter of its width to the left or right from the center of the safety feature. For purposes of
locating the impact point for these tests, the center of a safety feature, such as a guardrail terminal,
should be defined as the center of resistance for head-on impacts rather than the geometric center of the
system. These tests are designed to primarily evaluate occupant risk and vehicle trajectory criteria. The
vehicle offset should be chosen to maximize the risk of exceeding occupant risk or inducing vehicle in-
stability. Consideration should be given to the direction of vehicle rotation induced by the vehicle offset
as well as the potential for tire interaction with rails or other anchor components that do not move when
the system is struck. For a W-beam guardrail terminal, the critical vehicle offset is typically toward the
traffic side since vehicle contact with the backside support posts will maximize yaw rotation and the
risk of instability. Vehicle offset to the field side might be considered when evaluating vehicle stability
on sloped terrain behind the guardrail terminal. Dummy location, left or right, should also be selected to
maximize vehicle yaw movement. The risk of vehicle instability can sometimes be increased by reduc-
ing the offset somewhat to allow the vehicle's tires to contact anchor components under the guardrail
terminal or crash cushion.
TESTS 31 and 41
For devices intended to decelerate vehicles to a stop, these tests are designed to evaluate the capacity
of the feature to absorb sufficient energy to stop the 2270P vehicle in a safe and controlled manner. For
gating systems, these tests are intended to evaluate occlìpant risk and vehicle trajectory criteria during
high-enelgy, head-on impacts. These tests are conducted with the vehicle approaching parallel to the
roadway with the center of the vehicle aligned with the centerline of the terminal or cushion. Again, the
centerline of the device is defined as the center of resistance during end-on impacts.
impact angle to be used in tests of non-redirective crash cushions is less clear. The crash cushion
should be analyzed and impact angles selected to maximize occupant impact velocities and ridedown
accelerations.
TEST 34
Test 34 is intended to evaluate the impact performance of terminals and crash cushions at the critical
impact point (CIP) where the behavior of these devices changes from gating or capturing to redirec-
tion. Vehicle trajectory and occupant risk are the primary concerls for this test. Criteria for selecting
the CIP for post-and-beam terminal or crash cushion systems are presented in Section 2.3.3.1.
Whenever practical, finite element analysis should be conducted to identify critical impact points for
post-and-beam systems as well as other terminals and redirective crash cushions.
TEST 35
Test 35 examines the capacity of a terminal or crash cushion for containing and redirecting heavy
passenger vehicles. For this test, a2270P vehicle is directed into the system at the beginning of the
length-of-need at an impact angle of 25 degrees. Note that, for non-gating crash cushions, the be-
ginning ofthe length-oÊneed should be very near the nose ofthe crash cushion. In this case, Test 35
should involve a vehicle impacting on the very end of the system where cushion behavior changes
from capturing to redirective. Hence, for non-gating systems, this test is essentially a CIP impact with
a light truck test vehicle.
TEST 36
This test is designed to examine the behavior of terminals and redirective crash cushions when at-
tached to rigid barriers or other very stiff features. For this test, tbe 2270P test vehicle is directed ìnto
the terminal or crash cushion at its CIP with respect to the transition to the backup structure. Note that
some terminals, including most W-beam guardrail terminals, are not attached directly to a stiff barrier
or backup structure. Test 36 is only recommended for terminals or cushions directly attached to very
stiff barriers or backup stluctures. General guidelines for determining CIP locations for this test are
included in Section 2.3.3.2. Whenever possible, finite element analyses should be used to determine
the CIP for Test 36.
TEST 37
Test 37 examines the behavior of crash cushions and terminals during reverse-direction impacts.
This test is recommended for any safety feature that will be placed within the clear zone of opposing
traffic. This test involves a2270P or I l00C vehicle striking the critical impact point (CIP) for re-
verse-direction impacts. CIP locations for reverse direction impacts vary greatly from one system to
another and a generalized system for identiffing these locations has yet to be developed. Note that the
configurations shown in Figure 2-3Afor Test 37 are intended for illustration purposes only and do not
necessarily reflect the actual test configuration.
For most crash cushions with fender panels lapped against opposing traffic, the CIP should be selected
to maximize the risk of snagging on the end of the last fender panel lapped in thjs manner.
Many crash cushions attached to concrete barriers incorporate a tapered section between the wider
cushion and the nanower barrier face. In this situation, Test 37 should normally be configured to first
strike the barrier or the tapered section in order to maximize the potential for snagging .The 2270P will
generally be the critical vehicle for this test when a crash cushion is being evaluated.
For post-and-beam terminals utilizing a breakaway cable system, the I l00C will generally be the
critical vehicle for this test, and the impact point should be selected to maximize the risk of the vehicle
snagging on the anchor cable.
TESTS 38 and 45
Tests 38 and 45 are intended to examine the performance of crash cushions and end terminals during
impacts by mid-size vehicles. The concem is that attenuator staging can be tuned to meet the testing
requirements for the small car and heavy pickup truck without adequately accommodating mid-sized
vehicles. For these tests, the centerline of the test vehicle is aligned with the centerline of the test article.
Note that Tests 3l and 4l involve heavier vehicles impacting the systems under the same impact condi-
tions. Thus, accelerometer data from Tests 3l and 4l can be used to identify if Tests 38 and 45 are needed.
As described in Appendix G, accelerometer data from Tests 3 I and 4l can be integrated to obtain the force
versus deflection characteristics ofthe terminal or crash cushion. The force vs. deflection data can then be
used to estimate the occupant impact velocity and ride-down acceleration for a 3,300-lb (1,500-kg) vehi-
cle. Note that this analysis will be conservative because impact forces experienced during Tests 3l and 4l
with the 2270P vehicle will be higher than those generated by a mid-sized vehicle due to the heavier mass
and higher crush stiffness of the 2270P vehicle.If the force versus deflection analysis predicts that the
terminal or crash cushion will meet evaluation guidelines for occupant impact velocity (OIV) and RA for a
3,300-lb (1,500-kg) vehicle, Test 38 or 45 is not recommended.
TEST 44
Test 44 is designed to evaluate the ability ofa non-redirective crash cushion to safely stop a large pas-
senger vehicle in a side impact. For this test, the 2270P test vehicle is directed into the crash cushion at
its CIP with respect to the transition to the backup structure. Note that non-redirective crash cushions
are not designed to safely attenuate this impact and therefore occupant risk parameters, evaluation
criteria H and I, are not among the recommended evaluation criteria. Howevår, these values should
be reported as a means for user agencies to estimate the potential risk of using non-redirective crash
cushions. Impacting vehicles should remain stable and uplight during this test. If a system truly has no
redirective capacity, such as a sand inefiial cushìon, the centerline ofthe test vehicle should be direct-
ed at the corner of the shieldedhazard as shown in Figure 2-38. However, if a non-redirective crash
cushion can be expected to provide some redirective capacity, general guidelines for determining CIP
locations presented in Section 2.3.3.2 for Test 36 should be followed.
the basic intent and framework of the recommended tests and evaluation criteria. Recommended test
vehìcles, impact speeds, and impact angles should be applied with the most critical impact points for
the safety feature under consideration.
During the tests, the support truck should be placed in second gear and the parking brake set. Note
that the support truck's parking brake should be in good condition at the time of the test. TMAs
designed for an unlimited support vehicle mass should be tested with the support truck blocked to pre-
vent forward or lateral motion. Table 2-4 presents recommended crash tests and Figure 2-4 displays
impact conditions for truck- and trailer-mounted attenuators (TMAs). Reference is made to Section
3.4.2.4 for supporl-truck parameters.
Portable work-zone traffic control trailers are common roadside fixed objects that are used in both
temporary and permaneñt applications. These devices are often placed near the traveled way for
maximum visibility. Left unprotected, these devices can pose a significant hazard, especially to the
occupants of small cars. Therefore, it may be necessary in some instances to shield these trailers or
to otherwise make them crashwofthy. Shielding may be accomplished by placing the trailer behind
a barrier or crash cushion. Safety treatments designed to be integral or attached to work-zone traffic
control trailers should be subjected to Tests 50 and 5 I from the truck- and trailer-mounted attenuator
category.
TABLE 2-4. Recommended Test Matrices for Truck- and Trailer-Mounted Attenuators
lmpact
lmpact Kinetic Energy
Test Test Speed,a lmpact Evaluation
Vehicle Angle,a Tolerance,
Level No. mph Point Criterlab
(km/h)
q, deg. kip-ft (kJ)
ç.
-ç,
TMA Truck
0= 0 DEG.
OFFSET = 0
TESTS 50,51,AND 54
TMA Truck
0= 0 DEG.
Y= OFFSET = W/3
TEST 52
0
s
q
TMA Truck
0= 10 DEG.
OFFSET = Wi4
TEST 53
TESTS 52 and 53
Tests 52 and 53 examine the capability of a TMA to safely attenuate off-center and angular impacts
from heavy passenger vehicles. Structural adequacy of the TMA and occupant risk are the two prima-
ry concems for these tests. Test 52 should be conducted with the heaviest allowable suppoft truck or a
rigidly blocked support truck for no upper support-truck weight limit while Test 53 should be conducted
with the lightest allowable support truck.
TEST 54 (Optional)
Test 54 is designed to evaluate the staging of energy absorbers in a TMA for impacts involving mid-size
automobiles. It is desirable that TMAs provide acceptable levels of protection for all passenger vehi-
cles. There is some concem that existing designs are finely tuned to minimize the TMA length while
meeting the requirements of the small passenger car and heavy pickup truck tests, and designers do
not consider occupant risk parameters for mid-sized car impacts. On the other hand, if existing designs
must be lengthened to meet the requirements of this new test, there is concem that costs and operational
problems may increase greatly and that durability will be diminished. Therefore, Test 54 is considered
optional. However, manufacturers and user agencies are encouraged to develop and implement TMAs
that can safely accommodate mid-sized vehicles. As presented previously in the description of Tests 38
and 45, this test can be waived with an analysis of the accelerometer data from Test 5l that indicates
proper attenuator staging.
The recommended full-scale crash tests only evaluate the impact performance of a TMA during pas-
senger vehicle collisions. To date, no truck-mounted attenuators have been developed that are capable
of safely accommodating heavy truck impacts. Further, the full-scale crash testing does not evaluate
operational considerations, such as the potential for fatigue failure of structural elements, moisture
absorption that increases unit weight, mobility of the system, or the influences of temperature variations
or other factors. Perþrmance and Operational Experience of Truck-Mounted Attentntors (93) presents
a synthesis of practices related to the selection of truck-mounted attenuators.
the potential flor occupant compartment intrusion, excessive deceleration, and vehicle instability during
high-speed impacts with all features in this category. Note that occupant risk criteria for breakaway
structures and work zone devices are much more stringent than those used for longitudinal barriers
and crash cushions. In recognition of the capability of breakaway systems to provide a higher level of
safety than is possible for longitudinal barriers and other safety features, these more stringent occupant
risk criteria are implemented. Table 2-5 presents recommended crash tests for evaluation of supporl
structures, work-zone traffic control devices, breakaway utility poles, and lorrgitudinal channelizers.
Reference is made to the Glossary for definition of these features.
Support structures include sign supports, mailbox supports, luminaire supports, emergency call box
supports, and road closure gates. Fire hydrants are also commonly located within the clear zone of
urban and suburban roadways. Although not specifically included in these guidelines, whenever the
impact performance of fire hydrants is evaluated, impact criteria from other support structures should
be utilized. Work zone traffic control devices include plastic drums, barricades, cones, vertical panels
and their supports, and delineator posts. The guidelines are applicable to both permanent and temporary
work zone devices.
Lights, batteries, solar panels, flags, and other ancillary features are often attached to support structures
and work-zone traffic control devices. These ancillary features can sometimes become separated and
penetrate through a vehicle's windshield. This behavior is most often observed with base bending sign
supports and work zone devices. These features should be tested with any common ancillary feature
in place. Any other attachments that are normally incorporated in field applications with work zone
traffic control devices, such as sandbags, flags, sign panels, etc., should also be utilized during crash
testing to ensure proper impact performance. Note that when lights and solar panels are mounted on tall
permanent structures and a breakaway device is employed, the risk of windshield penetration is greatly
reduced.
Water-filled plastic barier systems pose a unique situation since they are designed for a wide range of
structural capability and applications. The initial designs were intended as positive barriers with the
ability to contain and redirect errant vehicles. However, some of the more recent designs are intended
as channelizers, either used individually form a continu-
as stand-alone barricades or linked together to
ous unit. These devices, used singly or interlocked, are labeled as longitudinal channelizing devices in
the Manual on Unifonn Trffic Conlrol Devices (53). Fol systems intended to be used as stand-alone
baricades, the device should be tested as a bamicade. For systems designed as positive barriers, it
should be crash tested as a permanent or temporary barrier as presented in Section 2.2.1 -For longitu-
dinal channelizers, or any channelizing device incorporating individual elements that are connected to
form a continuous unit, these systems are considered a separate class of hardware with different test-
ing and evaluation guidelines, as described in Section 2.2.4.2 under tests 90 and 91. A device tested
as a longitudinal channelizer that has the appearance of a positive barrier must be clearly identified as
a channelizing device. The testing laboratory should prominently indicate that these systems are not
barriers on all test documentation and the manufacturers must permanently mark each individual unit
produced with a warning indicating that the system is evaluated only as a channelizing device and not
recommended for use as a positive barrier.
A critical impact angle (CIA) should be determined for each test recommended in Table2-5. The critical
impact angle should represent the worst case impact condition that is consistent with the manner in
which the test article will be deployed along the roadway. For example, safety features that can be used
near an intersection could be struck frorn virtually any direction. In this case, testing should be conduct-
ed at both 90 degrees from the notmal direction and at any orientation between 0 and 25 degrees that is
deemed to represent the highest risk for the system to
fail any of the recommended evaluation criteria.
Features that are designed to be used along the outsideofdivided highways need only be evaluated for
impact angles of 0 to 25 degrees. However, if this same feature can be used in locations where it could
be subjected to reverse-side impacts, 0-to-25-degree and 155-to-180-degree impact envelopes should be
considered. Some features will have more than one CIA. In this case, testing should be conducted with
the CIA that is judged to have the greatest potential for test failure. Whenever no CIA clearly poses the
greatest risk for test failure, multiple tests should be conducted.
Single support structures should normally be tested with the center of the support aligned with the
left or right quarter point on the impacting vehicle. Previous testing has shown that the offset can lead
to vehicle instability during high-speed testing. Further, if low-speed and high-speed tests are to be
conducted with an I l00C vehicle, the same automobile can often be used for both tests, provided there
is only minor vehicle damage during the low-speed testing. Devices with multiple supports, such as a
dual-leg sign system, should be tested such that the impacting vehicle engages the maximum number of
supports possible.
As shown in Table 2-5, three full-scale crash tests are recommended for three of the classes of hardware
in this category. Two of these tests involve high-speed impacts with I l00C and 2270P vehicles. Both
tests are recommended to accurately identiff the potential for test article intrusion into the windshield
or roof of the two classes of test vehicle. Although two high-speed crash tests are recommended, it may
not be necessary to run both tests. High-speed bogie tests have been successfully employed to identi-
fy test article trajectory and occupant risk parameters. High-speed bogie tests can be used to identif,
which of the two high-speed tests represents the greatest potential for failure. Thus, the number of full-
scale crash tests may be reduced. Although it is not recommended as a general practice, some agencies
have accepted work-zone features based solely upon results ofhigh-speed bogie testing, provided there
is clear evidence that the test article trajectory would not pose any hazard to the vehicle occupants and
that occupant risk measures are well within recommended limits. Impact conditions for this class of
safety fearures are shown in Figure 2-5.
TABLE 2-5. Recommended Test Matrices for Support Structures, Work-Zone Traffic Control
Devices, and Breakaway Utility Poles
lmpact Accept-
Test lmpact lm-
Test Speed,a able KE Evaluation
Lev- Feature Vehicle Angle,a pact
No. mPh Range, Criteriab
el
(kmth)
0, deg. Polnt
klp-ft (kJ)
1-60 1100c 19 (30) clA s34 (41) (c) B,D,F,H,I,N
Support
1-61 1100c 31 (50) ctA >72 (e7) (c) B,E,F,H,I,N
Structures
1-62 2270P 31 (50) ctA >148 (202\ (c) B,D,F,H,I,N
00<0>25"
N orm al D irection
of Travel
Figure 2-5. lmpact Conditions for Support Structures, Work-Zone Traffic Control Devices,
and Breakaway Ut¡lity Poles
Test 70 is designed to evaluate the ability of small vehicles to activate any breakaway, fracture, or
yielding mechanism associated with the work zone feature during low-speed impacts. For free-
standing, lightweight features, velocity changes during low-speed impacts will be within acceptable
limits, even when a breakawa¡ fracture, or yielding feature is not incorporated. Therefore, Test 70 is
considered optional for work-zone traffic control devices weighing less than 220 lb ( I 00 kg).
Tests7l and 72 are intended to evaluate the behavior of features during high-speed impacts. The most
common risks of failure for these tests include intrusion of structural components into the vehicle
windshield, vehicle instability, and occupant risk criteria. Note, however, that lightweight free-
standing features cannot cause sufficient velocity change to result in faìlure of the test under occupant
risk criteria. Therefore, Tests 7l and72 can be conducted without the instrumentation necessaty for
deternining occupant risk whenever the test article has a total weight of 220 lb ( I 00 kg) or less. In this
case, vehicle intrusion, windshield damage, and vehicle stability are the primary performance evalua-
tion factors.
effect on the impact performance of a driveway slope. Further, the critical test conditions for some
features change significantly depending upon the design parameters. For example, the rollover risk
during V-ditch traversals is maximized when an encroaching vehicle leaves the ground on the foreslope
and then re-contacts the ground just past the middle of the ditch. For a shallow ditch with steep slopes,
a l5-degree encroachment would produce this behavior whereas, for a deeper ditch or flatter slopes,
higher encroachment angles are required to attain this "worst practical condition." As a result, it is im-
possible to identiff a single set of test conditions that will adequately explore the impact performance of
roadside geometric features.
Roadside geometric features should generally be designed and evaluated within the general framework
of one of the six test levels for longitudinal barriers. However, critical impact angles should be select-
ed for the particular geometric features under evaluation. Critical impact angles for geometric features
should be identified based upon the risk of rollover predicted by computer simulation. Further, it is
recommended that computer simulation be used to identifl critical roadside geometries and that full-
scale crash testing be used to evaluate only these geometric conditions as a way of veriffing simulation
predictions.
Some roadside geometric features, such as driveways and parallel drainage structures, have proven diÊ
ficult to design to meet TL-3 impact conditions. In many cases, agencies have found that designing such
features to meet the 62-mph (100.0-m/h) encroachment condition is impractical and not cost-
effective, even on highways where almost all safety features are designed to meet TL-3 requirements.
Howeve¡ practical and economical designs may be possible for intermediate speeds between TL-2 and
TL-3. In this situation, it is recomrnended that additional encroachment speeds between 44 and 62 mph
(70 and 100 km/h) be considered for use in roadside geometric feature design and testing. Additional
research is needed to develop cost-effective safety treatments for these geometric features.
Pavement discontinuities include any irregularity in the paved surface, such as a pot hole or a drop-off
at the edge of pavement. Full-scale crash testing has been utilized to evaluate vehicle behavior during
impacts with these discontinuities. Unlike roadside features, pavement discontinuities are normally
encountered at low encroachment angles. Further, as is the case with pavement edge drops, the risk to
motorists is also a function of driver behavior. In these situations, it is prudent to identifo the critical
approach condition or driver behavior under which the testing should be conducted. For example, edge
drops are normally tested in an "edge scrubbing condition" whereby the driver drops two wheels off of
the pavement and then attempts to re-enter the travelway at a Very shallow approach angle. If executed
properly with a deep and steeply sloped pavement edge, the front tires will be restrained from regain-
ing the pavement by the exposed edge. Increasing steering at this point can cause loss of control when
the tires suddenly mount the pavement edge. The primary safety concern for pavement discontinuities
include the potential for loss of control that can lead to vehicles intruding into opposing traffic lanes or
run-off-the-road crashes and the possibility for high decelerations associated with major vertical pave-
ment anomalies, such as long, deep pot holes. Wherever possible, impact conditions should be selected
to be within the general framework for longitudinal barriers. Note, however, that low inrpact angles are
much more common and, in many cases, more critical for pavement discontinuities than are high impact
angles. Section 5.3 presents additional discussions on the evaluation ofcrash test results from both
pavement discontinuities and roadside geometric features.
2.3.1GENERAL
The impact point on a safety feature is the point at which the test vehicle first contacts the test article.
The impact point for a redirective barrier can affect its impact performance. The potential for wheel
snag, pocketitrg, aud sil'uctural failure is in solne ways related to the irnpact poilrt for ltatry barrier
systems. V/ithin practical limits, impact points should be selected to represent the critical location along
a barrier system that will maximize the risk of test failure. The impact point that maximizes the risk of
test failure is labeled the critical impact point (CP). The following sections present recommendations
for locating CIPs for redirective bariers, crash cushions, and terminals.
The BARNER VII computer program has been widely used as the primary tool for identiSing CIP
locations for longitudinal barrier tests (125). This program has been shown to be capable ofaccurate-
ly predicting the extent of snagging, pocketing, and barrier loadings during full-scale crash testing.
Therefore, whenever possible, the BARRIER VII program should be used to estimate critical impact
locations for redirective features. Procedures described in Appendix A may be used for this purpose.
More recently, the LS-DYNA computer program has been utilized for this same purpose. When cor-
rectly implemented, the LS-DYNA program offers greater analysis capability and can include the
evaluation of vehicle rollover into the analysis of CIP locations. Howeveç the level of effort required
to conduct an LS-DYNA simulation rnake its use solely for the purposes of determining CIP locations
impractical. However, in cases where an LS-DYNA system model is developed during the process of
designing a safety feature, it should also be used to identiff critical impact locations.
When BARRIER VII or LS-DYNA analyses are impractical for determining critical impact points, the
following guidelines, derived from BARRIER VII, should be utilized.
For flexible cable barrier systems, the critical impact point within the length-of-need should take into
consideration the vehicle type as well as conditions that increase the propensity for override, underride,
and penetration between the cables. For small ( I I 00C) and mid-size ( 15004) passenger vehicles, the
target impact point shall be at mid-span between support posts to evaluate the potential for underride or
penetration between cables. For light-truck passenger vehicles (2270P) as well as other heavy trucks,
the target impact point shall be limited to 12 in. (300 mm) upstream from a post for all test conditions.
Critical impact points for flexible cable barrier systems are provided in Tables 2-28 through2-28. AT
this time, guidelines for selecting the CIP to evaluate approach guardrail transitions between flexible
cable barriers and stiffer longitudinal barrier systems are unavailable. As such, computer simulation
with LS-DYNA or BARRIER VII is recommended for identifling the CIP for the required transition
crash tests.
Cable barrier systems are often implemented with a range of acceptable post spacings. For cable
barriers placed within V-ditches for which a range of post spacing options is desired, critical vehicle be-
haviors andlor barrier system performance cannot be evaluated using only one post-spacing. Howeve¡
it is highly impractical to perform the full matrix of crash tests, as shown in Tables 2-2B through
2-28, on each post-spacing alternative intended for use in median ditches. Recommended guidance for
selecting critical post-spacing for crash testing and evaluating cable barrier systems with multiple post
spacing options intended for use in median ditches are provided in Table 2-6 and further discussed in
Section 42.3 of Appendix A.
TABLE 2-6. Recommended Post Spacing for Evaluating Cable Barriers Placed within Median
Ditches
Test
Vehicle Barrier
Desig. Primary Evaluatlon Factors Post Spacing
Type Position
No.
3-1 0 1100c Level Terrain Stability, Occupant Compartment Deformation, Narrowest
Undenide, and Penetration
3-11 2270P Level Terrain Working \Mdth and Stability Both
3-1 3 2270P Front Slope Stability, Override, and Working lÄ/idth Narrowest
3-'t4 1't 00c Front Slope Stability, Occupant Compartment Narrowest
Deformation, and Penetration
3-1 5 1100c Back Slope Undenide, Occupant Compartment W¡dest
Deformation,and Occupant Risk
3-1 6 1100c Back Slope Stability, Override, and Penetration Narrowest
3-17 1 5004 Fronl Slope Penetration and Occupant Compartment Deformation \Mdest
3-'18 2270P Back Slope Override and Penetration W¡dest
Most crashworthy cable median balrier systems have been constructed with three to four longitudinal
cable elements. Since median barriers can be struck on either side, the side of the support posts to which
the cable elements are attached may altemate from post to post or by cable element. Although several
cables often contribute to successful vehicle containment and redirection, errant vehicles have been
successfully captured with as few as one cable. In crash tests oflight trucks into cable nledian barriers,
there has been a demonstrated propensity for support posts and attached cables (particularly those on
the non-impact side of the support post) to be pushed downward by the impacting vehicle. This behav-
ior can lead to increased propensity for light+ruck passenger vehicles to penetrate through or override
the cable barrier system.
Therefore, it is recommended that a cable median barrier be crash tested and evaluated with the pri-
mary capture cable for the appropriate design vehicle placed in its most critical position (i.e., the back
or non-impact side of the support post). For barrier systems in which the attachment of the primary
capture cable alternates between the impact and non-impact sides of posts, the CIP should be selected
to be upstream from a post with the primary capture cable attached to the back (non-impact) side of the
post. In general, the primary capture cable will correspond to the highest cable element located between
the critical bumper height and mid-height of the front headlight, referred to as the critical captì.ue zone.
Generally speaking, the mid-height of the front headlight is positioned at an elevation below the front
comer of the engine hood. If no cable exists within this critical capture zone, the primary capture cable
should be taken as the lowest cable above this critical region. Considering a vehicle oriented at 25
degrees with respect to a tangent barrier system, the critical bumper height corresponds to a point on the
corner ofthe front bumper that first contacts the cable barrier, or a hypothetical vertical plane parallel
to the cable barrier system. When selecting the primary capture cable, testing laboratories may consider
performing vehicle dynamics simulations or using existing simulated vehicle trajectories for passenger
vehicles traversing 6H:lV and 4H:lV median ditches (61,74,86,87,88, 89, 95) in order to identi$ the
predicted vertical contact location ofthe vehicle in relation to the banier face. Further, testing labora-
tories may also consider the height of critical vehicle components (i.e., top and critical height of front
bumper, mid-height of headlight, engine hood height, etc.) when selecting the critical cable.
F),
Fo= (Eq.2-3)
,s
Where:
F^ post dynamic yield force per unit length of ban ier
Fy dynamic yield force of a single post
,s post spacing
M, is the effective plastic nloment of all barrier rail elements. For a single rail barrier system, Mo is
merely the plastic moment of the rail element. The effective plastic moment of a multiple rail system is
the sum of the plastic moment of the highest beam and the plastic moments of lower beams reduced by
a ratio of the heights of the highest and lower rail elements as given in Equation 2-4.
Hi
Mr=Mn+ZM, @q.2-a)
Hh
Vy'here
MP effective plastic moment of all barrier rail elements
Mn plastic moment of highest rail element above ground or deck
lvI, plastic moment of a lower barrier rail element
H¡ height of a lower rail element
Ht, height of highest rail element
A more detailed discussion of $ and Moaswell as tables of typical values can be found inAppendixA.
Figure 2-6 can be used to locate the critical impact point as defined by the distance x for the length-of-
needportionofpost-and-beamtypelongitudinalbarriers(Tests l0and ll)foragiventestlevel.The
figures show plots of the critical impact distance x for values of Fo and Mrfor a given barrier system.
Distances shown are measured upstream from the reference post/splice as shown in Figure 2-1. A
rail splice should be located at or just upstream of the reference post, provided this is consistent with
in-service practice. Interpolation may be used to find the x values for points between the curves, and ex-
trapolation may be used to find the x values for points above the upper curve or below the lower curve.
Figures 2-12 rhrough 2-17 are used to locate the CIP for transitions between longitudinal barriers
having different lateral stiffness (Test 20 and2l). The figures show plots of the critical impact distance
x for values of Fo and M, for a given barrier system. Distances shown are measured from the upstream
end of the stiffer system as shown in Figure 2-1. Properties of the more flexible barrier should be used
for determining Fo and Mo.lnTerpolation may be used with these figures as described above for Figures
2-6 through 2-l l. Note that these figures were developed with a transition to a rigid barrier. V/hen the
stiffer barrier is not rigid, the distance x will increase slightly. However, crash testing and simulation
have shown that this effect is relatively small and can usually be ignored. A more detailed discussion of
the above procedures can be found in Appendix A.
Rigid banier CIP locations have been estimated from full-scale crash testing and computer simulations.
Table 2-7 shows CIP estimates for rigid baniers. Note that these estimates may be a little imprecise
because of the limited testing and simulation conducted to date with MASH test vehicles. As testing
experience with these vehicles increases, agencies should re-examine the values recommended in Table
2-7 and revise them as needed. AIso, these numbers represent minimum x values for non-rigid bamiers.
Thus, whenever extrapolations of curves shown in Figures 2-6 through 2-17 give CIP values lower than
those shown in Table 2-7,thex distances from Table 2-7 should be used.
Extensive computer modeling with the LS-DYNA program has been conducted with free-standing
concrete barriers. Although most of this effolt has been focused upon barrier design, these models
should be used when practical to estimate critical impact locations. Critical impact locations should
be determined based upon the degree of shear displacement as the test vehicle approaches a joint and
the amount of barrier tipping during the crash. Shear displacements at a joint can lead to snagging and
barier tipping can lead to vaulting and vehicle rollover. When computer nlodels are unavailable, CIP
distances shown in Table 2-7 should be used for temporary barrier testing.
TABLE 2-7. Critlcal lmpact Polnt for Rlgid Barrler Tests with 1100C and 227OP Vehicles
J
@ 2016 by the American Association of State Highway antl Transportation Olïcials.
All rights reserved. Duplicationis aviolation of applicable law.
Chapter 2-Test Matrices and Conditions | 51
3 6
Critical lmpact Point for Test 10, Test Level 1
â
cl
-!.
|l.a
0
l0 r5 20 25
-
Critical lmpact Point for Test 10, Test Level 1
25
)
2.0
è
t 1.5
|la-
1.0
05
30 40 50 70
-
SEE FIGURE 2-1 FOR "x"
Figure 2-6. Critical lmpact Point for Test 10, Test Level 1
J
@ 2016 by the American Assocíalion ol'State Highway and Transportation Ofïicials.
All rights reserved. Duplication is a violation of applicable law.
52 | ManualforAssessingSafetyHardware
4.5
ô
CL
3
=a-
u_
5.5
7.0
.5 8.6
5 10 't5 20 25
4.0
4
4.5
.g
J
q
u_
5.5
30 40 50 60 70
-
SEE FIGURE 2-1 FOR "x"
Figure 2-7. Critical lmpact Point for Test 10, Test Level 2
rË
.g
v
u_
a 3
75
2
10.5
12.O
105
13.5
5 10 15 20 25
CL
=a_ 3
u
9.0
10.5
30 40 50 60 70
-
SEE FIGURE 2-1 FOR "x"
Figure 2-8. Critical lmpact Point for Test 10, Test Levels 3, 4,5, and 6
35
4.0
4.5 40
4
I
ô
=a. 4.5
rL
5.0
5.0
6.0 5.5
ô.5
5 't0 l5 20 25
)
3-5
ct
=a 3
TL
50
55
60
65
30 40 50 60 70
-
SEE FIGURE 2-1 FOR "x"
Figure 2-9. Critical lmpact Point for Test 11, Test Level 1
')
Critical lmpact Point for Test 11, Test Level 2
6
o-
=a_ 3
u-
120 '10.5
1s.5 12î
'15 0
5 't0 l5 20 25
) 40
35
30
o-
=a_
u_ 2_0
7.5
15
10
105
12-i
13.5
05
30 40 50 60 70
-
SEE FIGURE 2-1 FOR'.x,'
Figure 2-1O. Critical lmpact Point for Test 11, Test Level 2
10.s
5 9.0
Ë 4
o
=a-
|l
3 90
'10.5
2
12-O
13.5
5 10 15 20 25
30
25
C
o-
'E- zo
a
L¡-
6.0
15
'10
9.0
5
90
10 20 30 40 50 60 70
-
SEE FIGURE 2-1 FOR "x"
Figure 2-Ll-. Critlcal impact po¡nt for Test 11, Test Levels 3, 4, 5, and 6
12
l0
ã
gô 6
(
.a-
0
0 2D 40 60 80 100 120
-
SEE FIGURE 2-1 FOR "X'
Figure 2-L2. Ctitical lmpact Point for Test 20, Test Level 1
J
@ 2016 by the Amcrican Association of State Híghway and Transporrarion Ol'lìcials.
All rights reserved. Duplication is a violation of applicable law
58 | Manual for Assessing Safety Hardware
10
I
â
.s
J 6
u-q
4,00
4.25
4.50
4.75
5.00
5.25
0 l0 15 20 25
2.5
2.O
o.
! 4.00
a_
u_ 1.5
4.25
4.50
1.0
4.75
5.00
0.5
5.25
0
30 45 60 75 90 't05 120
-
SEE FIGURE 2-1 FOR "x"
Figure 2-13. Critical lmpact Point for Test 20, Test Level 2
5.0
10
o.
J¿ 6
o_
u_
2 7.O
0
0 5 10 15 20 25
)
't0
6
CL
=a-
L
4
60
6.5
7.O
40 60 80 100 '120
-
SEE FIGURE 2-1 FOR'x"
Figure 2-L4. Critical lmpact Point for Test 20, Test Levels 3, 4,5, and 6
ct
! 4
,aa
0
5 l0 l5 20
'12
)
10
CL
6
=
t...o
0
30 45 60 75 90 '105 120
-
SEE FIGURE 2-1 FOR "x'
Figure 2-15. Critlcal lmpact Point for Test 21, Test Level 1
,^\
Critical lmpact Point for Test 21, Test Level 2
)
12
10
CL 5.25
Þ
=a_
u-
6.50
6.75
0
5 l0 l5 20 25
4
o.
=a_
TL
2
5.25
5.50
575
6.00
6.25
6.50
6.75
0
30 45 60 75 90 'i05 120
-
SEE FIGURE 2-1 FOR "x"
Figure 2-16. Critical lmpact Point for Test 21, Test Level 2
,)
@ 2016 by the American Association olState Highway and Transportation Offi c ials
All rights reserved Duplication is a violation ofapplicable law
62 | ManualforAssessingSafety Hardware
10
9.
I '10.50
11.25
o. .50
.Y 6
{a-
4
8,25
2 10.50
11.25
0
t0 't5 20 25
l0
d
! 6
a-
u_
7.50
2
9.00
30 45 60 75 90 105 '120
-
SEE FIGURE 2-1 FOR "x'
Fi$ure 2-L7. Critical lmpact Point for Test 21, Test Levels 3, 4,5, and 6
of full-scale crash testing of vertical instrumented walls and are measured from the point of contact to
the point of maximum loading. Table 2-8 shows CIP estimates for heavy truck impacts with rigid barri-
ers. Slightly larger x values are indicated for non-rigid or safety-shaped concrete barriers.
Note that positive x values indicate that the point of maximum loading is downstream from impact and
negative values indicate that it is upstream of the impact point. Computer simulation should be used to
refine the estimates shown in Table 2-8 whenever practical. Note that little testing has been conducted
with the 100005 vehicle and therefore recommended CIP locations are based upon testing with the
smaller truck recommended by NCHRP Report 350. As testing experience with the new test vehicle
increases, agencies should refine the CIP estimates shown in Table 2-8.
2.3.3.1 Test 34
This test is designed to evaluate the potential for vehicle instability when small cars impact the side of a
terminal or crash cushion near the beginning of the system. For this test, the CIP is defined as the point
at which the behavior of the terminal or crash cushion changes from redirecting the impacting vehicle
to either capturing the vehicle or allowing it to gate through the system. For post-and-beam guardrails,
this definition implies that the CIP is the point farthest from the impact head where an impacting vehicle
will still break the leading post. For these terminals and crash cushions, computer programs, such as
BARRIER VII and LS-DYNA, can be used to estimate this point. The BARRIER VII program has
been employed to identiÐ/ the CIP for tangent W-beam guardrail terminals mounted on 6 in. by 8 in.
(150 mm by 200 mm) wood posts and one full-scale crash test was conducted to examine the findings.
The original simulation indicated that the CIP should be between the leading post and the point where
the cable anchor bracket is attached to the guardrail. After studying the simulation findings and a single
full-scale crash test, it is recommended that the CIP for tangent gualdrail terminals should be approxi-
mately I ro 2 ft. (0.3 to 0.6 m) downstream from the leading post. Steel breakaway posts would be likely
to perform differently. BARRIER VII or LS-DYNA should be used, when practical. to determine spe-
cific CIP locations for the terminal or crash cushion under evaluation. Most non-gating crash cushions
use relatively rigid diaphragms that provide solid supporl for the redirective panels. As a result, vehicles
striking even a few inches downstream from the start of the first full-strength side panel will be redi-
rected. Therefore, the CIP should be considered to be upstream of the beginning of the first full-strength
side panel. Detailed finite element analysis programs, such as LS-DYNA, can be used to help estimate
thc CIP for these systcms. Figure 2-l B displays a recommended impact condition for non-gating crash
cushions that may be used when detailed modeling is not available.
I
D
I e
D12
I
0 = 15 DEG.
Normal Direction
of Traffic
Figure 2-18. Critical lmpact Point for Test 34 on Non-Gating Crash Cushions
2.3.3.2 Test 36
Test 36 is intended to examine the transition between a crash cushion and a rigid or nearly rigid backup
structure or barrier end. The primary concem for this test is that the crash cushion will not provide suÊ
ficient rigidity to prevent severe wheel snag or pocketing on the end of the rigid feature. LS-DYNA, or
other explicit finite element codes, can be used to determine the critical impact point for this test. V/hen
practical, these models should be employed to select the CIP for Test 36. At this time, few studies have
been conducted to identiÕ/ critical impact points for these features. Crash cushions secured against lat-
eral movement by cables or other moderately flexible systems are expected to have a CIP between 9 and
11ft(2.7 and3.4 m) from a rigid support structure. Systems anchored against movement by diaphragms
attached directly to a concrete pad or to steel railings placed on the ground are examples of stiffcrash
cushion designs that are expected to have a CIP betweenT and9 ft(2.1 and2.7 m). Whenever practical,
finite element modeling should be used to identifr the critical impact point for Test 36.
2.3.3.3 Test 37
This test is intended to examine the behavior of crash cushions and terminals during reverse direction
impacts. As discussed previously, CIP locations for reverse direction impacts vary greatly from one
system to another and there is no generalized system for identifying these locations. For most crash
cushions with fender panels lapped against opposing traffic, the CIP should be selected to maximize the
risk of snagging on the end of the last fender panel lapped in this manner. Many crash cushions attached
to concrete bariers incorporate a tapered section between the wider cushion and the narrower barrier
face. In this situation, Test 37 should normally be configured to first strike the barrier or the tapered sec-
tion in order to maximize the potential for snagging. For post-and-bearn teminals utilizing a breakaway
cable system, the impact point should be selected to maximize the risk of the vehicle snagging on the
anchor cable. Finally, for flexible barriers, the critical impact location should be selected to nlaximize
the risk of one or more lqngitudinal elements, such as a cable, riding over the front of an impacting
vehicle. This situation would maximize the risk of the terminal snagging the front of the vehicle.
2.3.3.4 TÞst 44
Test 44 is designed to evaluate the ability ofa non-redirective crash cushion to stop a large passenger
vehicle in a side impact. The CIP is selected to maximize the risk of the vehicle impacting the backup
structure. For non-redirective crash cushions that truly have no redirective capacity, the centerline of
the test vehicle should be directed at the comer of the shielded obstacle. However, if the non-redirective
crash cushion has some redirective capacity, general guidelines for determining CIP locations presented
in Section 2.3.3.2 for Test 36 should be followed.
All impact performance guidelines published to date include crash tests involving tracking vehicles. In
other words, all tests are configured to have the impacting vehicle rolling forward into a roadside safety
feature without any yawing or sideslip motion. However, crash data analyses have shown that approxi-
mately half of all run-off-the-r'oad crashes involve non-tracking vehicles in a yawing or sideslip motion
at the time of impact (41 ). Fufthermore, crash data studies also appear to indicate that the impact perfor-
mance ofroadside features can be adversely affected by non-tracking behavior (84,127).
Although the effects of non-tracking impacts on the impact performance of many roadside features are
not fully understood, it is clear that side impacts have the highest potential for severe injuries for many
nalrow roadside features, such as large sìgns, luminaire poles, narrow crash cushions, and teminals
( I 1 6). The side impact scenario with the greatest potential for injury involves the door of an impacting
vehicle striking the safety feature with the vehicle in a broadside skid. In this situation, the safety fea-
ture can penetrate a long distance into the occupant comparlment with a high likelihood of striking and
injuring vehicle occupants.
Although two sets of crash testing guidelines have been developed for side impact testing (114, I l5),
few roadside safety features have been designed to accommodate side impacts. Furthermore, it is antici-
pated that most large structural supports and terminals and crash cushions would likely fail either of the
side impact testing guidelines referenced above.
In recognition of the preliminary state of roadside safety feature design to accommodate side impacts,
the lack of compatibility with vehicle design, and the ever changing vehicle fleet, the impact perfor-
mance evaluation guidelines contained herein do not specifically recommend any side impact testing.
Additional research is needed to identi$ strategies for accommodating side impacts into narrow road-
')
side features. The latest side impact testing guidelines can be found in Me asurement of Heavy Vehicle
Impact Forces and Inertial Properties (13). Users are encouraged to conduct full-scale crash tests under
these guidelines whenever practical in order to build a better understanding of the efrcacy of the pro-
posed procedures and the performance of modem safety, features during side impacts.
\I
J i
Test lnstallation
3
3.1. GENERAL
ff ighway safety features are evaluated for a particular test level through a series of vehicular crash
I Itests under idealized impact conditions. Many important test parameters have been standardized
to arrive atapractical stratification oftests (test matrices) and to enhance the degree oftest replica-
tion. Care must be exercised in the interpretation of test results and in the projection of the results to
in-service performance. Good performance under ideal test conditions does not ensure comparable
perfotmance under in-service conditions. As discussed in Chapter 1, the evaluation process should not
stop with successful completion of tests recommended herein. Crash testing and in-service evaluation
are both critical to evaluating the performance of a safety feature and in-service evaluation should be
pursued as recommended in Chapter 7.
A flat surface, preferabìy paved, should be used when accelerating the test vehicle to the desired speed
and to provide for unrestricted tlajectory of the vehicle following impact. The surface should be free
of curbs, swales, ditches, or other irregularities that could influence impact or post-impact behavior of
the vehicle except when test conditions require such features. Ifnecessary a flat compacted soìl or sod
surface should adjoin the paved approach area to replicate conditions for safety features normally sur-
rounded by an unpaved surface, so that post-impact vehicular behavior can be properly assessed. Test
documentation should include the type of pavement or soil surfaces used on the approach to the test
article, surrounding the test article, and on the post-impact trajectory of the vehicle.
It may not always be possible to have totally flat and smooth surface for the test vehicle to accelerate
a
to the desired speed, and the vehicle may be bouncing somewhat down the track. However, it is critical
that the vehicle be stabilized at impact. A compressed suspension at impact may cause the vehicle to
nose dive and result in underriding of the test device. A.n extended suspension at impact may accentuate
the upward movement of the vehicle and result in overriding and vaulting of the test device. In either
case, the impact performance of the test device will likely be affected. Therefore, it is important that the
test vehicle not be bouncing excessively just prior to impact and that the actual bumper height at impact
be within 2 in. (50 mm) vertically of the nominal bunrper height, i.e., when the vehicle is at rest.
3.3 SOIL
The impact performance of some soil-mounted features depends on dynamic soil structure interaction.
Longitudinal bariers with soil-embedded posts and soil-embedded support sûuctures for signs and lumi-
naires are such features. When feasible, these features should be tested with soil conditions that replicate
typical in-service conditions. Soil conditions are known to vary with time, location, and environmental
factors, even within lelatively small geographical areas. Therefore, except for special test conditions, it
is necessary to standardize soil conditions for testing. In the absence of a specific soil, it is recommended
that all features whose impact performance is sensitive to soil-structure interaction be tested in a soil that
conforms to the performance specification as described in Section 3.3.1. However, product developers
and user agencies should assess the potential sensitivity ofa feature to foundation conditions. Ifthe fea-
ture is likely to be installed in a soil that could be expected to degrade its performance, testing in one or
more of the special soils described in Section 3.3.3 may be appropriate.
3.3.1STANDARD SOIL
It is recommended that the standard soil meet AASHTO standard specifications for "Materials for
Aggregate and Soil Aggregate Subbase, Base, and Surface Courses," designation M 147, grading
A or B B, Section Bl). It should be compacted in accordance with Sections 304.05
(see Appendix
ofAASHTO's Construction Manualfor Highway Construction (2) (see Appendix B, Section B3).
The soil should be re-compacted, as necessary before each test to rneet density requirements of the
Construction Manual (2). The soil should be well drained at the time of the crash test. The test should
not be performed if the ground is frozen or if the soil is saturated unless the test is specifically designed
to evaluate these conditions.
The basis for determining the recommended performance characteristics of the foundation soil was
developed frorn the existing material specifications used in NCHRP Repon 350 (129).In a round
robin test of testing facilities, the dynanric perfonnance of the cun'ent soil standard was evaluated.
This standard dynamic test, as detailed in Appendix B, Section 85, showed that the existing material
specífications are acculately represented by a minimum dynamic load (applied with a bogie vehicle
or a pendulum) of 7,500 lb (3,400 kg) at deflections between 5 in. ( 125 mm) and20 in. (500 mm) of a
standard steel post installation. Each test facility is free to define the soils and installation procedures
provided that this dynamic performance criterion is met. For example, the typical steel guardrail post
driven into a stiff cohesive soil may well exhibit the required dynamic performance and would therefore
be an acceptable installation process.
It is important to keep in nrind that the parameters associated with the soils and installation procedures
should be applied consistently to all installed systems to assure similar foundation performance. These
parameters include, but are not limited to native soil, backfill soil, compaction, moisture content, and
installation configuration. Since the properties of the native soils may vary at a given test facility, the
installation procedure required to meet the minimum dynamic performance may also vary.A recom-
mended summary sheet, for inclusion in each test report involving soil-mounted installations, is shown
in Figure 3-1.
Once a testing facility has developed an installation procedure that provides sufficient dynamic capacity,
it is important to make sure that the installation is not affected by conditions beyond the control of the
testing facility. For example, a significant rain may affect the performance of the installed system. To
date, these environmental influences have been difficult, if not impossible, to quantif,. Also, the cost of
performing dynamic tests to quantifl soil performance for each crash test installation is relatively high.
llhus, a static test procedure is recommended as a surrogate. Afrer the dynamic performance of a given
soil and installation procedure is determined to be acceptable and certified, a static test utilizing the
same steel post installation will be performed. The measured static load from this static test will form
the basis for evaluation of future test installations. More detailed descriptions of this calibration process
are presented in Appendix B, Section 86.
The recommended static test procedure is as follows. During the installation of each test system, two
additional standard steel posts will be placed in a convenient location near the test installation where
they will be subjected to the same environmental conditions as the installed system. The placement of
these posts, as well as all system hardware, should accurately reflect the standard installation method
certified by the testing facility.
On the day of crash testing, a static test willbe performed on one ofthese two test posts to determine
if outside factors have affected soil strength. The installation is determined to be acceptable if the static
resistance of the soil at 5, 10, and l5 in. (125,250, and 375 mm) of deflection, as defined in Appendix
B, Section 86, is at least 90 percent of the value determined during certification testing. If the measured
soil strength is lower than the required limits, the crash test will have to be postponed until the soil
conditions ale such that the soil strength is acceptable. The second test post is intended for this situation
in which the test has to be postponed and the soil strength re-measured. It is important that each testing
agency establishes the relationships between soil strength and various influencing factors for the local
standard soil in order to know when the soil strength is within the specifications.
Dynamic Setup PosþTest Photo of Post Static Load Test Post-Test Photo of Post
Pel@t Fin6r vs. GÉln Slzo ot Fill SollforOynámlc añd Sl,ouc Losd Tssls
,41\
(+!.J
--aÊ Oyr#Td
¡ Fl sd
! --.1- s¡ic h
ilsdt
Ê
qù Stq O (m)
ol
Hvdreullc
---- LdEOkphffirl cyllndet '5ln'
bMÑDù
td6.obpt.c.mt
bLdcd
- 72h.l182sffil
R6qtu ld
E Ohplåc.mnt 2+h.
--- - Ldnqspbcffinr &¡uþ Fü
nm SdcLdTd
0
r Static Loâd
40 h.(t0t€ ßh-(1ß2mm)
TesUlnstallâlion
Dlsplåæmonl (ln.)
Details
)
41412005
Test Facllity & Site Location Midwest Road Side Safety Facility (see attached site map)
ln situ soil descriptlon IASTM D2,l87l ....,......,,.. Sandy Gravel with silty fines
Fill mater¡al descriptlon (ASTM D24871 & sieve analysis .,,.. AASHTO Grade B Soil-Aggregate (see sieve analyses above)
Descr¡ptlon of fill placement procedure 6-in. (152-mm) lifts lamped with a pneumatic compactor
Bog¡e Weight.... 2108.8 lb (ss4 kg)
Figure 3-1. Recomrnended Summary Sheet for StronÉ Soil Test Results
Comparison of Static Load Test Results and Required Minimum: Load vs. Displacement
6000
5000
S-
3
E
)I
4ooo
sooo
v v \ \r
\
2000
1000
z \
0
0 5 't0 l5 20 25 30 ,t0 Static Load Setup
D¡splacement (ln,)
Percent Finer vs. Grain Size of Fill Soil from Static Load Tests
100
90 -
80
irP70 oo
=bs0
Ëoo
lL Post-Test Photo of Post
3o
20
't0
0
100 t0 1.0 0.1 0.01
Date 5t23t2005
Test Facllity & Slte Location Midwest Road Side Safety Facility (see attached site map)
ln situ soil descrlption IASTM D2487) Sandy Gravel w¡th s¡lty fines
Flll material description (ASTM D2487) & sleve analysis ..,.. MSHTO Grade B Soil-Aggregate (see sieve analyses above)
Description of fill placement procedure 6-in. (152-mm) lifts tamped with a pneumat¡c compactor
Note that the specified soil strength is intended as a minimum value and that higher values are ac-
ceptable. However, it should be emphasized that the testing agency should use the same installation
procedure used in the standard certification test for all test installations. It is unacceptable to have
different installation procedures for different safety hardware evaluation programs. In situations where
the native soil varies significantly at a given test facility, the installation methods established for each
native soil should have similar dynamic performance. It should be noted that the measured soil resis-
tance is associated with a specific combination of in-situ soil, standard soil, compaction condition,
and depth and lateral extent of the fill material. Variation of any of these factors would be expected to
change the soil strength and therefore necessitate a recalibration effort.
Weak Soìl.It is recommended that the weak soil meet AASHTO standard specification for "Fine
Aggregate for Hydraulic Cement Concrete," AASHTO designation M 6. The soil should be compacted
in accordance with Section 304.05 ofAASHTO's Construction Manualfor Highwry Construction (2).
The soil should be re-compacted, as necessary before each test to meet density requirements of the
Construction Manual (2). The soil should be well drained at the time of the crash test.
Saturated Soil. The "standard soil" and "weak soil" previously described may be used to evaluate the
impact performance of a feature under saturated soil conditions. Moisture content of the soil should
replicate expected in-service conditions.
Frozen,Sa¡7. The "standard soil" and '\veak soil" described may be used to evaluate impact performance
of a feature under frozen soil conditions. The degree to which the soil is frozen, measured in terms of
depth and temperature, and its moisture content should replicate expected in-service conditions.
The lateral extent of the fill material (if utilized) should be at least that of the minimum lateral extent
established during the dynamic certification test. However, it is generally acceptable to utilize a greater
lateral extent for the fill material if the stiffness of the fill material is greater than the native soil. Again,
variation in embedment for the purpose of tuning soil performance for a particular full-scale test is
not acceptable. The depth of the structural fill (if utilized) should always extend below the installed
appurtenance.
A sign support is typically embedded by driving the post or stub directly into the soil, by inserting the
support in a drilled hole and then backfilling the soil, or by placing the support or stub in a concrete
footing. Similar methods are used for embedment of a longitudinal barrier post. Most luminaire support
poles are supported by a concrete footing, and most utility poles are placed in a drilled hole. The soil is
then backfilled.
Some testing agencies have developed universal foundations for testing breakaway devices consisting
of multiple-use base plates supported on a very rigid concrete footing. While these foundations reduce
testing costs, they effectively eliminate soil-foundation interactions. They also raise questions about
feature-to-foundation interface friction and anchor bolt rigidity. If it cannot be shown that these effects
are insignificant, the test reporl should, at a minimum, alert the user agency of potential problems and
provide recommendations for foundation systems that will ensure proper breakaway performance of the
supporl.
Testing of a bridge rail will, in general, require a special support structure, i.e., a simulated bridge deck.
For a non-rigid bridge rail where lateral deflection is of concern, the structure to which the railing is
attached should simulate edge conditions so that the effect of vehicular penetration beyond the edge
of the deck can be properly evaluated. Regardless of the rail's strength, it may be desirable to evaluate
structural adequacy of the deck itself for the impact conditions, in which case the deck structure should
have the same strength and properties as the in-service structure.
If a universal or generic deck is used in a brìdge rail test, it is desirable that impact loads imposed on
the deck be measured or computed froln measured vehicular response and reported. By so doing, a
user agency will have some guidance on how to design a deck that differs from the one used in the test.
Procedures that may be used to estimate impact loads from measured vehicular accelerations are present-
ed in Reference I12. For tests of a flexible-to-a-rigid longitudinal banier transition, a prototype bridge
end structure or wìng wall structure must be constructed. Length, strength, and geometry of the prototype
should be sufficient to approximate the impact response expected of the in-service bridge end or wing
wall. Section 3.4.2.1contains further recommendation on the prototype bridge end.
3.4.1GENERAL
The material characteristics of all key elements in the test alticle that contribute to its structural integlity
or impact behavior should be documented in the test report. Physical and chemical material propefties
can generally be obtained from the supplier providing the test article(s). When mill certification test
results or other documentation are not available, materials from key elements should be sampled, tested,
and reported. To ensure that all critical elements are considered, a careful after-test examination of the
test article is essential. Materials should be tested independently when a failure occurs.
Material specifications such as AASHTO, ASTM, and so forth, should be reported for all key elements.
R.esults of random sample tests should confirm that stated specifications have been met and that key
elements in the test arlicle were representative of normal production quality. The testing agency should
offer a judgment on the potential effects of marginal materials or materials that significantly exceed
minimum specifications on the performance of the test alticle. In addition, speciñed but unverified prop-
erties of all other materials used in the test article should be reported.
The test article should be constructed and erected in a manner representative of in-service installations
and should conform to specifications and drawings of the manufacturer or designer. To assure uniformi-
ty and integrity of structural comections, current Aurerican Weldiug Society specifications for highway
bridges (7), Aluminum Association specifications for aluminum bridges and other highway structures
(l), American Institute of Steel Construction bolting procedures (l l), and other relevant documents
should be used as appropriate. Deviations from fabrication, specification, or erection details should be
delineated in the test report.
For tests examining performance of the length-of-need section, rails or barrier elements should be
installed straight and level and then anchored. Horizontally curved installations, sloped shoulders, su-
perelevation, embankments, ditches, dikes, and curbs should be avoided for general performance tests.
Barriers intentionally being tested for ditches, curbs, and roadside slopes should be tested under the rel-
ative conditions. The use of any nonstandard features should be documented and reported. As a general
rule, length ofthe test section, including terminals or end anchorage devices, should be at least three
times the length in which deformation is predicted, but not less than 75 ft (23 m) for a rigid banier (i.e.,
one for which little, if any, lateral displacement is anticipated), 100 ft (30 m) for a semi-rigid banier
(e.g., a metal beam-and-post roadside barrier), and 600 ft (183 m) for a flexible barrier (e.g., cable barri-
ifutilizing system lengths shorter
er). For serni-rigid barriers, testing laboratodes should use discretion
than 175 ft and be able to provide adequate analysis to support the use ofthe shorter system length. It is
recommended that semi-rigid barrier lengths only be tested at lengths less than 175 ft if the laboratory
can ensure that there will be minimal contact length and deformed barrier length during the impact, that
similar rail and post parts are used in both the body of the barrier system and the attached terminals,
and that the main section of the barrier and the terminal sections have equivalent stiffness and dynamic
deflection characteristics. Length ofthe test section should be such that: (l) terminals or end anchor-
age devices do not significantly influence the dynamic behavior of the bamieç and (2) the ability of the
barrier to contain and redirect the test vehicle in the recomrnended manner can be clearly ascertained.
Exceptions to recommended lengths can be made if it can be demonstrated that the installation satisfies
these two requirements. Demonstrátion of an exception to the recommended barrier length requirements
should include an analysis of anchor loading, cable or rail extension, and membrane action effects that
might reduce lateral deflections. Note that it is desirable to measure anchor loads during testing of
flexible (e.g., cable barrier) and semi-rigid barriers to enhance understanding of the transmission of rail
tension along these barriers.
When a flexible barrier (e.g., cable barrier) is tested in combination with a roadsîde slope or median
ditch, the length of the sloped terrain (e.g., slope or ditch) should exceed the predicted vehicle contact
distance or length of defomed barrier to permit evaluation of vehicle containment, redirection, and
stability prior to the vehicle encountering any tmnsition to level temain. The testing laboratory shall
locate the sloped terrain segment within the minimum required 600-ft ( I 83-m) length of barrier. Based
on historical cable ban ier testing in ditches, a 360-ft ( I I 0-m) to 400-ft (122-m) long segment of sloped
terrain should provide adequate length to evaluate barrier performance and vehicle behavior. However,
the testing laboratory is responsible for determining the length and position of the sloped terrain region
within the overall length of barrier system. At the upstream and downstream ends of slope or ditch sec-
tion, grade transitions may be used to transition the barrier system to level tenain before terminating the
barrier ends with an appropriate cable end terminal or end anchorage system.
'When
evaluating cable barriers installed on slopes or within median ditches, crash test vehicles may
become airborne before and/or after impacting the barrier system. When the airborne vehicle returns
to the ground, the vehicle-ground contact may involve one or more wheels and possibly other parts of
the vehicle. The soil properties and level of compaction in the region of vehicle-ground contact can
influence vehicle behavior and vehicle-barrier interaction. To achieve a higher degree ofconsistency
and uniformity in testing, it is recommended that an adequatély-sized region of well-compacted base
material (e.g., AASHTO M 147 Grade B material) measuring approximately 6 in. (150 mm) to 12 in.
(300 mm) thick be placed as a lining in the ditch in the predicted location of initial vehicle contact with
the sloped terrain through the location of maximum lateral barrier deflection.
Roadside slopes and median ditches will have slope break points defined by the intersection of two
different grades at one or more specific locations. For example, the front slope break point is formed by
the intersection of level approach terrain with the fill slope. In the real world, these discrete grade inter-
section locations (slope break points) will be rounded to some degree rather than be defined by a shàrp
angle. To account for this field rounding, the front and back slope break points of median ditches and
roadside slopes used for testing and evaluation of cable barrier systems may be constructed with radii of
3 ft (0.9 m) to 5 ft (1.5 m).
For a barrier system that is dependent on rail tension to perform properly, e.g., tensioned cable barri-
er, the barrier system should be tensioned to that specified for a temperature of 100.0"F (37.8'C) for
testing. This minimizes the effect of varying ambient temperature on the test results of a tensioned
banier system. Fufthermore, a cable barrier is encouraged to be tested with both the largest and smallest
recommended post spacings. Testing with the largest post spacing allows for assessment of maximum
barrier deflection as well as its greatest working width. Testing with the smallest post spacing allows
for assessment of vehicle instability. Barrier deflection and working width associated with intermediate
post spacings may be estimated from computer sinrulation or supplenrental crash testing.
A fi'ee-standing, unanchored ban'ier, such as a pre-cast, segmented concrete barier whose impact
performance depends in paft on frictional resistance between it and the surface on which it is resting,
should be tested on a surface that replicates the type on which it will be placed in service. If it will be
placed on more than one surface in service, it should be tested with the surface that is likely to have the
most adverse effect on performance, usually the one having the least frictional resistance. The type of
surface as well as end anchorages or terminals used should be reported.
The banier system used for a transition test should be oriented as it would be in service. As a general
rule, transitions of most concerÍì are those that serve to connect a less stiff ban'ier on the upstream side
to a stiffer barrier on the downstream side, such as the transition from a flexible roadside or median
barrier to a rigid bridge rail. In such cases, length, strength, and geometry ofthe prototype bridge rail or
wing wall should be sufficient to approximate impact response expected of the in-service bridge rail or
wing wall. It is recommended that the length of the prototype bridge rail or wing wall be at least l7 ft
(5 m). A minimum of 50 ft (15 m) of the more flexible barrier is recommended, exclusive of a properly
anchored end.
Lr cases where the tlarnition sel'ves to conlìect longitudinal bart'iers with sinrilar lateral stiffness, but
with different geometry a minimum of 50 fr ( I 5 m) is recommended for each of the adjoining barriers,
exclusive ofa properly anchored end.
A rigid, non-yielding backup structure should be used to simulate a highway feature (such as a bridge
pier, elevated gore, or bridge end) when appropriate. The shape ofthe backup structure should be select-
ed to represent the worst case condition for each of the recommended tests. For example, a redirective
crash cushion designed to be used in the median should be tested with a rectangular-shaped backup
structure with the maximum allowable width for tests involving the vehicle approaching from the crash
cushion side of the obstacle. In this way, the risk of a vehicle snagging on the end of the backup struc-
ture during Test 36 should be maximized. However, during reverse direction testing, Test 37, a naffower
backup structure, such as a concrete safety-shaped barrier should be utilized to maximize the risk of
the test vehicle snagging on the end of a cushion's fender panel system. The crash cushion should be
anchored as required by specifications or drawings.
3.4,2.3 Support Structures, Work-Zone Trafic Control Devices, and Breakøvay Utility Poles
Reference should be made to Section 2.2.4 for recommendations relative to the manner in which a
support structure, a work-zone traffic control device, or a breakaway utility pole should be oriented
with respect to the vehicular approach direction. Testing should verifo breaka\ilay or yielding features
designed to function identically when impacted from specific directions, such as a breakaway base de-
signed for front or rear impacts, or those features designed to function identically when impacted from
any direction, e.9., a breakaway base designed for omni-directional irnpacts.
Supports should be fully equipped with full-height struchrres, including signs, mailboxes, call boxes, and
mast arms for luminaires. For tests of a luminaire support, it is preferable that an actual luminaire be used
rather than a substitute equivalent weight as has often been used in the past. Tests have shown the lumi-
naire/ballast can break loose during impact and consequently may present ahazard to other motorists or
to occupants of the impacting vehicle. A full-lengtli utility pole should also be used together with associ-
ated cross-arm(s), down guy wires, and conductors (lines). For road closure gates, the typical length of
the gate arm anticipated to be used in the field should be installed. Road closure gates are typically tested
with the gate arm in the up position. Nevertheless, an assessnrent should be made as to whether the test is
more clitical with the gate arm in the up or down position, and test accordingly.
The test may involve multiple supports such as multiple mailbox supports or multiple, closely spaced
drums in a work zone. Orientation and spacing of these supports should be representative of in-
service conditions.
Occasionally, a work-zone traffic control'device, such as a barricade or a plastic drum, will overturn
or will intentionally be placed in an overturned position along the highway shoulder in a work zone.
Altematively, a barricade may typically be placed so that its panels are parallel rather than perpendic-
ular to traffic. Devices placed in these types of "out-of-seryice" orientations may pose a greater risk to
an errant motorist than the upright or normal position. If there is a reasonable expectation that a device
will be commonly found adjacent to traffic in an "out-of-service" orientation and this orientatión poses
a grealer risk to the motorist than in the normal position, it should be tested in the altemate position. If
it cannot be determined which position is more critical, tests in both the normal and "out-oÊservice"
orientations should be conducted.
For tests of a sign support system, the area of the sign panel should approximate the largest panel that
would normally be used on the support system. Sign panel material should be the same as that normally
used or to be used on the support system. If panels of different materials, such as plywood, sheet metals,
or fiber reinforced plastics, are used with the support system, the test should be conducted with the
material expected to pose the greater risk to occupants of the impacting vehicle. If it cannot be deter-
mined which material is most critical, it is recommended that the test be conducted with the panel of
greatest mass. The aspect ratio of the sign (height-to-width ratio) should be typical of the largest panel
that would normally be used on the support system. Mounting height of the sign panel (distance from
ground to bottom of panel) should be the minimum height the panel would normally be mounted in
service unless it can be shown that a highel mounting height would pose a greater risk to occupants of
the impacting vehicle.
For tests of a mailbox supporl system with a single mailbox, the mailbox should be the largest that
would normally be used on the suppoft system. For tests of a mailbox support system with multiple
mailboxes, the number and size of mailboxes should be the largest that would normally be used on the
support system.
appropriate, an infinitely heavy support truck can be simulated by blocking the support truck against
forward and lateral motion.
Barrier geometrics that should be reported include: mounting heights; lengths of rail elements; length of
test installation; lateral placement of barrier relative to embankment or ditch; length and longitudinal po-
sition of sloped temain within the overall test installation; alignment and orientation of the barrier relative
to vehicular approach; and targeted point of impact relative to the end of the barrier. Other parameters
to be reported include: barrier tension for cable barriers; type and location of temperature compensators;
cable connectors, splices, and fittings; and the layout and configuration ofsections on the slope or within
the ditch with compacted base material. The manner in which the barrier is terminated, anchored, or both
should also be presented. All connection details should be described, including precast concrete barrier
joints; nuts, bolts, and washers used to anchor a bridge railing post to a bridge deck or abutment; and all
hardware used to connect rail elements or attach them to a supporting post. Bolt torques or tightening
procedures should also be identified. Detailed construction drawings sufficient for complete reconstruc-
tion ofthe as-tested system should also be reported.
Reported foundation details should include the manner in which the barrier was supported, including
description of the structural details of the attachments between the bamier and its foundation or asso-
ciated bridge deck or pavement. Details of any simulated wing wall and or brìdge end should also be
reported for tests involving approach guardrail transitions. Embedment procedures and depths for posts
and other components should also be adequately described. The type ofsurface and the frictional char-
acteristics of pavements placed under free-standing temporary barriers should also be reported.
As described in Chapter 6, specifications for the as-tested materials should be reported for all irnport-
ant components and for associated hardware (fasteners, nuts, washers, backup plates, etc.) used in the
barrier. Actual sizes and physical and chemical properties should be repofted. Often this information is
available on mill certification documents that can be obtained from the supplier.
The repoft should also include the manner in which the device was supported, including embedment
procedures for posts embedded in soil (driven, drilled and backfilled, placed in driven inserts, concrete
footing, etc.); description ofthe soil and soil properties as described in Section 3.3.1; and description of
the surface on which a free-standing device was resting, including the type and condition of pavement,
when relevant.
Specifications for the as-tested materials should be reported for all important components and for asso-
ciated attachment hardware (fasteners, nuts, washers, backup plates, etc.) used in the test arlicle. Actual
sizes and physical and chemical properties should be reported. Often this information is available on
mill certification documents that can be obtained from the supplier.
Foundation data to be reported include: embedment procedures for posts or poles embedded in soil
(driven, drilled and backfilled, placed in driven inserts, concrete footings, etc.); embedment depth of
test article; description ofsoil and its properties ifrelevant; and description ofthe surface on which a
free-standìng device was resting, including the type and condition of pavement, when relevant.
Specifications for the as-tested materials should be reported for all irnportant components and for
associated attachment hardware (fasteners, nuts, washers, back-up plates, etc.) used in the test article.
Achral sizes and physical and chemical properties should be reported. Often this information is avail-
able on mill ceftification documents that can be obtained fì'om the supplier. Dynamic impact properties
of materials that are designed to fracture or tear, such as brittle sign suppofis and frangible bases, should
also be reported.
Specifications for the as-tested attenuator, other structural elements, associated hardware (fasteners,
nuts, washers, backup plates, etc.) used in the test afticle, and sizes of hardware elements should be
reported.
The make, model, mass, and ballasting of the strpporting truck should be reported. For Tests 50, 51, and
52, the maximum allowable weight of the support truck should be used in the testing while the mini-
mum allowable weight should be used for Test 53. The support truck may be braced against forward
and lateral motion if there is no upper limit on the weight of the support truck that can be used with the
TMA. The manner in which the supporting truck is braked (i.e., parking brake set, engine in second
gear, etc.) or blocked against forward motion should also be described in detail.
Test Vehicle
Specifications
4.1. GENERAL
p ull-scale crash tests are the primary method for measuring the impact performance of a roadside safety
I feature. Proper selection, preparation, and instrumentation ofthe test vehicle are critical
components of the impact performance evaluation process. The test vehicles are chosen to be representa-
tive of a wide range of vehicles. Procedures for preparation and instrumentation oftest vehicles have been
standardized as much as possible to improve consistency in the evaluation process. This chapter identifies
the recommended procedures for selecting, preparing, and instrumenting test vehicles.
For most devices, only two test vehicles are used to represent the entire fleet of passenger cars operating
on the nation's highways. These vehicles are selected based upon the philosophy that, if a safety feature
performs satisfactorily for both the smallest and largest passenger vehicles, it should perform adequate-
ly for all vehicle sizes in between.
Surrogate test vehicles are sometimes used to evaluate the impact performance of breakaway systems
and small traffic control devices. Pendulum and bogie vehicle tests are used to evaluate the low-speed
impact performance of breakaway signs and luminaire poles and high-speed bogie tests can be used to
evaluate the impact performance of work-zone traffic control devices. The mass and stiffness of these
vehicles must be controlled to properly match existing test vehicles.
The following sections describe the five production test vehicles and the surrogate test vehicles in great-
er detail, provide guidance for preparing the vehicles for use in the testing, and present techniques for
instrumentation and data collection during the testing.
The small test vehicle is selected to be representative of the 2nd percentile in terms of vehicle weight for
all passenger type vehicles sold in 2002. This vehicle was found to be a small sedan weighing approxi-
mately 2,4201b (1,100 kg). A review of all gasoline-powered sedans in this weight category showed that
the vehicles were similar in length, height, width, and wheelbase. Therefore, any small passenger car
weighing near 2,420lb (1,100 kg) is deemed to be acceptable as the small car test vehicle.
The large test vehicle was selected to be representative of the 90th percentile in terms of vehicle weight
for all passenger vehicles sold in 2002. As may be expected, all vehicles in this weight category are
light trucks, including many four-door sport utility vehicles and a variety of pickup trucks. Large sport
utility vehicles (SUVs) make up the highest proportion of vehicles with a mass of 5,000 lb (2,270 kg)
or greater. Large SUVs typically have high centers of gravity (c.g.), ranging from a low of 28 in. (710
mm) to a high of 29.5 in. (750 rnm) (65). In light of the lower costs for used pickup trucks compared
to large SUVs, a two-wheel-drive, four-door, half-ton pickup truck, weighing approximately 5,000 lb
(2,270 kg), was selected to represent the large passenger vehicle class. These pickup trucks were found
to have masses and center of gravity heights similar to those of large SUVs. In an effort to ensure that
test vehicles have a center of gravity in the same range as large SUVs, a minimum c. g. height of 28 in.
(710 mm) was selected for the pickup truck test vehicle.
A mid-size test vehicle has also been selected for use in evaluating the performance of staged ener-
gy-absorbing systems. This alternate test vehicle, designated 1500A, will only be required when the
analytical procedures for evaluating crash cushion and terminal performance described in Appendix G
indicate a potential failure. A sedan weighing 3,300 lb (1,500 kg) was selected as the most representa-
tive vehicle body style and as the most appropriate mass to explore these occupant risk measures.
It should be noted that the passenger vehicle fleet is changing constantly as a function of consumer
preferences and other unforeseen circumstances, such as fuel availability and pricing. The selected test
vehicles should be reviewed periodically to make sure that they remain representative of the current ve-
hicle fleet. Appendix H outlines a recommended procedure for re-evaluating and selecting test vehicles
when changes in the vehicle fleet are deemed significant enough to revise the test vehicles.
The remaining three test vehicles are heavy trucks selected to provide strength and containment tests for
median barriers and bridge railings. These vehicles include a single-unit van truck weighing 22,000lb
(I 0,000 kg), a tractor-van trailer weighing 80,000 lb (36,000 kg), and a tractor-tank trailer weighing
80,000 lb (36,000 kg). All heavy truck test vehicles should incorporate a cab-behind-engine configura-
tion, not a cab-over-engine design. The single unit truck was selected to provide a reasonable increment
in barier capacity between the passengervehicle strength test and the tractor-trailer strength tests. The
masses of the two tractor-trailer tests were selected to be representative of the most comrnon tractor-trail-
er weight combinations. The van trailer test vehicle should have a semi-monocoque structural design
and should include a sliding undercarriage (slide axles). The trailer rear tandems and axles should not be
modified in any way for the test. The tank-trailer test vehicle was selected to provide a maximum con-
tainment test for median barriers and bridge railings that would prevent most tractor-trailer combinations
from penetrating or rolling over the top of the bamier. In order to fulfill this objective, tank-trailers should
have a minimum height to the center of the tank of 80.5 in. (2050 mm).
Surrogate test vehicles, including pendulum and bogie vehicles, are used to evaluate the behavior of
breakaway devices during low-speed impacts and the behavior of work-zone traffic control devices.
Vehicle mass and the characteristics of the crushable nose of surrogate vehicles are the most critical
parameters associated with these tests.
Tires used in testing should match the manufacturer's recommendations found on the vehicle certifi-
cation plate. However, when necessary to meet recommended c. g. height requirements, vehicle tires
may be changed to one of the factory-installed optional tire sizes available at the time of vehicle manu-
facture. All-season highway tires should be used on test vehicles, not mud or snow tires. Vehicles with
aftermarket modifications to the vehicle body, suspension, or wheels should be restored to original man-
ufacturer's specifications prior to testing. Factory-installed optional equipment, such as power brakes
and steering, antiJock and anti-skid braking systems, and specialty wheels, are acceptable, provided
the vehicle meets the overall guidelines presented below. In addition, fuel tanks should be purged and
the battery removed to reduce exposure to needless hazards. However, with the requirement to have the
povier on for test vehicles in order to collect Event Data Recorder (EDR) and airbag deployment data, a
supplemental battery that is sealed and non-volatile will have to be used.
All passenger vehicles used in crash testing should be no more than 6 model years old on the day the
test is conducted. It is recognized that some research projects can experience extensive delays. To elim-
inate the potential for these delays to require replacement oftest vehicles purchased in anticipation of
testing, it is acceptable to utilize test vehicles that are within 6 model years of the date when the original
research project was initiated. Although it is cost-prohibitive to apply the 6-year limit to heavy truck test
vehicles, it is desirable to utilize vehicles of recent vintage. Heavy truck test vehicles should be repre-
sentative of widely used designs.
Test vehicles are specified by basic body style and vehicle mass properties. All test vehicles should fall
into the recommended ranges on basic vehicle mass properties. For fhe 2270P vehicle, a minimum vehi-
cle c. g. height is also specified. Although it is recommended that vehicle c. g. height be measured after
test preparation has been completed, measurement of a base vehicle is acceptable, provided the nleasured
height is properly adjusted for mass changes associated with test preparation. Vehicle c. g. height should
be measured using a suspension method, such as that described in Volume 4 of the 1986 SAE Handbook
(133). Alternate methods for measuring vehicle c. g. height may be employed, provided the testing
agency verifies that the accuracy and precision of the altemate procedure is comparable to the suspension
method.
In addition to the body style and mass parameters that must be me| ranges for other vehicle propefties,
such as wheelbase, track width, and overall length, are presented to serve as a guide for test vehicle
selection. \Mhen practical, test vehicles should be selected to conform to all the parameters shown in
Tables 4-1 and 4-2. Also, when possible, passenger vehicles used in crash testing should be selected
from one of the top three best-selling vehicle models within each test vehicle classification. Automobile
sales data may be obtained from the Automotive Yearbook (159), published by V/ards Reports, Inc., or
the Market Data Book (39), published by the Automotive News magazine
The2270P test vehicle should be a half-ton, 2-wheel-drive, pickup truck with four full-size doors.
Note that even though some extended-cab pickups have four doors, including two small rear doors that
cannot be opened independent ofthe fì'ont door, these vehicles do not conform to the 2270P vehicle rec-
ommendations. The test vehicle should not have a special bed, such as a "Sportside" or "Stepside." The
2270P vehicle may include a heavy-duty suspension supplied by the manufacture¡ provided the vehicle
mass properties and c. g. heights fall in the proper range.
The 2270P test vehicle must have a minimum c. g. height, in the ballasted configuration, of 28 in.
(710 mm). This requirement will necessitate that the c. g. height for all2270P test vehicles be measured
and reported. Note that the vehicle c. g. height of a light truck can often be raised by as much as l/2 in.
(13 mm) by replacing worn tires, after-market tires, or both with tires meeting the size recommendation
on the vehicle certification plate.
Note that there is little consistency in the mounting of truck boxes to single-unit truck frames. Weak
attachments can allow the cargo box to separate from the truck frame, thereby greatly reducing bari-
er loading during a full-scale crash test. To assure that the truck box attachment meets the minimum
capacity requirement, testing agencies should evaluate the existing attachments relative to the attach-
ment guidelines contained in Ford's 2005 Body Builder Layout Book (56).If the existing attachments
are found to be inadequate relative to the referenced guidelines, truck box attachments should be added,
sufficiently reinforced, or both to meet the guidelines.
) TABLE +t. Recommended Properties of 1100C, 15OOA, and 227OP Test Vehicles
1100c 1500A 2270P
Propefty
(Small Car) (lntermediate Car) (Pickup Truck)
MASS, lb (kg)
Test lnertial 2420 * 55 (1100 r 25) 3300 + 75 (1 500 t 35) 5000 r 110 (2270 +50)
Dummy 1 ô5 (75) Optionala 0ptionala
Max. Ballast 75 (80)
1 440 (200) 440 (200)
Gross Static 2585 t 55 (1175 t 25) 3300 t 75 (1500 ¿ 35)a 5000 t 110 (2270 t 50)a
a lf a dummy (surogate occupant) is used, the gross static vehicle mass should be increased by the mass of the dummy.
b Average offront and rear axles.
c For "test inertial' mass.
d 2270P vehicle must meet minimum c. g. height requirement.
TABLE 4-2. Recommended Properties of 100005, 36OO0V and 36000T Test Vehicles
I 0000s 36000V (Iractor/Van Trailer) 36000T firactorlTank Trailer)
Property (Síngle-Unit
Van Truck) Tractora Traile¡b Combinalion TractoÉ TraileÉ Combination
Mass,lb (kg)
Curb
13,200 x. 2,200
N/Sd N/Sd
29,000 t 3,100
N/Sd N/Sd
29,000 t 3,100
(6000 t I 000) (13,200 + 1,400) (13,200 t 1,400)
Ballaste As Needed N/Af As Needed N/Af N/Af As Needed N/Af
Test lnertial
22,046 + 660
N/Sd N/Sd
79,300 + 1100
N/Sd N/Sd
79,300 t 1100
(10,000 * 300) (36,000 r 500) (36,000 t 500)
200 200
Wheelbase (max) 240 (6,100) N/Sd N/N N/Sd N/Af
(51 00) (5,1 o0)
Overall Length
3e4 (1 0,000) N/Sd 636 (16,155) 780 (19,850) N/Sd N/Sd 780 (19,850)
(max)
Trailer Overhangc
N/N N/Af 87 (2,200)h N/Af N/Af 73 (1,850) N/AI
(max)
49 x.2 50t2
Cargo Bed Heighti N/Af N/AT N/Af N/Af N/AT
(1,245 * 50) (1,270 t 50)
Curb mass-mass of test vehicle in its standard manufacturer conditions \ilith all fluid reservoirs filled
and without occupants or cargo. In general, curb mass should not vary significantly.
Test inertial mass-mass of test vehicle including all items rigidly attached to the test vehicle, includ-
ing secured ballast and test equipment. Test inertial mass does not include any loose ballast or surrogate
occupants.
Gross static mass-total mass of the test vehicle, including all surrogate occupants; ballast, whether it
is secured or not; and all test equipment.
Passenger vehicles should be selected to require the minimum amount of adjustment to the curb mass
to achieve test inertial mass. Whenever possible, only seats, spare tires, fuel tanks, fluids, and optional
equipment should be removed from the vehicle in order to bring the vehicle into the proper mass range.
Note that bench seat frames have been shown to contribute to the rigidity of some light truck vehicles
and, as a result, the frames of these seats should not be removed from the 2270P test vehicles.
4.2.1.2 Ballast
Whenever test vehicles are found to have insufficient curb mass, fixed ballast may be added to 1100C,
15004, and2270P test vehicles, up to the maximum amount shown in Table 4-1. Ballast installed in
passenger vehicles can be either concrete or metal blocks placed in the passenger compartrnent and
rigidly attached to the vehicle structure by metal straps or other devices capable of withstanding a static
load of 20 times the ballast weight in both the longitudinal and lateral directions. Ballast should be
distributed to minimize changes to the horizontal and vertical locations of the vehicle center of gravity
in its curb mass condition. If necessary to minimize movement of the vehicle c. 9., some ballast may be
placed in the bed of the 2270P vehicle. It is important to note that the vertical c. g. location of lhe 2270P
vehicle should be 28 in. (710 mm) or greater after all ballast is added to the vehicle. In general, ballast
should not be used to raise the c. g. height of test vehicles that otherwise would not conform to the min-
i mum height requirements.
Heavy truck test vehicles, 100005 and 36000V, must be ballasted with simulated cargo to reach the
target gross static mass values shown in Table 4-2.Ballas| used in these vehicles can include sandbags,
concl'çte or steel blocks, bales of hay, or other means. Ballast should be as uniformly distributed across
the length and width of the cargo bed as possible. When heavy materials such as steel, concrete, or sand
are used as ballast, they should be placed on top ofa platform to raise the ballast c. g. to the proper
height. Wooden pallets can be used to create the platform, but any means that provides the proper total
cargo c. g. height is acceptable. Ballast should be firmly secured to prevent movement during and after
the test.
Water should be added to the tank trailer on the 36000T test vehicle to achieve the nominal test inertial
mass shown in Table 4-2.If lhe tank trailer has compaftments, the water should be distributed among the
compartments to provide a ballast c. g. location that is as close as possible in both the longitudinal and
vertical directions to the c. g. location achieved when all tanks are filled. Furthe¡ it is desirable to fill as
many compartments as possible in order to minimize sloshing effects during the crash test.
Any ballast used in a test vehicle should be thoroughly documented. At a minimum, this documentation
should include the ballast type, mass, location, center of mass, attachment method, and, when appropri-
ate, photographs ofthe ballast.
Any reliable method for guiding test vehicles into the test article is acceptable, provided the guidance
system can be removed from the test vehicle prior to impact and any components remaining attached
to the vehicle do not influence the outcome of the test. Many agencies utilize a cable guidance system
attached to one of the front tires. Although this system has proven to be reliable, a small hub normally
remains attached to one of the front tires during the crash test. In the past, some of these hub com-
ponents have contacted the ground during violent vehicle motions and stabilized some vehicles that
appeared to be on the verge of rolling over and tripped some vehicles that otherwise seemed likely to
remain upright. Special care should be taken to place this hub or any similar guidance system compo-
nent that remains attached to the vehicle in a location that will not allow the element to contact the test
article or the ground.
It is also necessary to measure occupant compartment deformation during the test by accurately mea-
suring vehicle interior dimensions prior to and after the test. A procedure for measuring the occupant
compaltment interior has been developed and is described in detail in Appendix E. This procedure in-
corporates a three-dimensional cool'dinate system that is established using points on the vehicle interior
that are not expected to deform during the test. The coordinate system is used to identifo the position
of the interior surface with respect to the vehicle reference points, both before and after the full-scale
crash test. The pre- and post-test positions can then be compared to measure the extent of vehicle de-
formation. Pre- and post-impact configurations of the vehicle's interior should also be well documented
photographically.
4. 2. I. 5 Surrogate Occupflnts
Although occupant risk during full-scale crash testing is to be determined by application of the "flail
space" model, a sun'ogate occupant is recornmended to be used during tests involving the I l00C
vehicle. The surogate occupant is specified to evaluate unsymmetrical vehicle mass distribution and
its effect on vehicle stability. V/hen utilized without instrumentation, a surrogate occupant should be
a Hybrid I or II dummy, representative of a 50th percentile male with a mass of approximately 165 lb
(75 kg), and should be placed into either the driver or right front passenger seat. For most I 100C tests,
the location of the dummy should be selected to maximize vehicle roll, pitch, and yaw motions during
the testing. The dummy should be secured in the seat with a seat belt, including both lap belt and shoul-
der harness.
A 50th percentile male dummy should also be used when there is a potential for an occupant to ex-
tend out of the vehicle and come into direct contact with the test article. This situation primarily arises
with tall barrier systems without sufficient set-back at the top of the banier or with luminaire and sign
supports placed on top ofbarrier systems. In this case, the head ofan occupant placed in the front seat
on the impact side of the vehicle can extend out of the vehicle window and contact the top of the barrier
or the support structure. It is recommended that dummies be placed in the front seat on the impact side
ofboth2270P and ll00Cvehiclesduringtestingofanylongitudinalbarrierwithaheightgreaterthan
or equal to 33 in. (840 mm) or if luminaire and sign supports are placed on top of the barrier'. For this
application, the dummy should be restrained with a lap belt and shoulder harness. An onboard camera
or tightly zoomed external cameras placed parallel with the barrier should be used to document the
dummy's head movement during the test.
Although not required for impact performance evaluation of common safety features subjected to
conventional impact conditions, instrumented dummies are recommended for use in roadside safety
research efforts, particularly when there is potential for occupant compartment deformation or intrusion.
Side impact testing is one area where extensive vehicle deformation is expected and further research
into the refinement and application of impact performance evaluation guidelines is needed. Researchers
are encouraged to utilize the instrumented dummy most appropriate for each type of testing. Side
impact dummies are likely to be most appropriate for side impact testing while Hybrid III or THOR
dummies may be more appropriate for oblique impact testing.
4.2.1.6 Documenlation
The pre-test conditions of all test vehicles should be well documented. Figures 4- I through 4-5 present
vehicle data items that should be measured and recorded for any full-icale crash test. It is preferable
that these data elements be presented in the format shown in these figures in order to promote consis-
tency among test agency reporting. Any dimensions that fall outside of the recommended tolerances
should be noted in the test report, along with a brief assessment of what affect the deviation may have
had on the crash test. Note that frame height measurements on the 2270P vehicle should be taken at
the ends of the frame rail and measuled to the bottom of the rail element. In addition, the longitudinal
location of the center of mass of trucks, tractors, and trailers should be reporled. Pre-test photographs
of the test vehicle should be taken to sufficiently document both the exterior and the interior conditions
of all test vehicles. For passenger vehicles, in addition to the areas described in Section 4.2.1 .4, pre-test
photographs should include the interior floorboard and the rear trunk interior. The pre-test interior pho-
a b
a m te I
s h
o p
e d s
W
Wheel Center Height
Engine Type
Mass Distribution
Engine Size
LF RF Transm ission Type:
LR RR Automatic or Manual
FWD or RWD or 4WD
Weights
lb (ks) Curb (No Fuel) Test lnertial Gross Stetic
w tot"l Mass
GVWR Ratings
Seat Posit¡on
Front
Rear
Total
c d
am nt f
s h
\
j
Test lnertial C.M.
k I
Ï¡re
m n
p
o p
o q r
s
s t
Wheel Center Height Front
LR RR Engine Type
Engine Size
Weights
lb (ks) Curb Tesl nert al G ross Static Transmission Type:
Wfront Automatic or Manual
w tot"l
F ront Type:
Back M ass:
c
u
n
h
aa
e o d
wheel center
Weights - lb (ks) Curb Test lnertial Gross Stat¡c
height front
wheel center
height rear
Wfront axel wheel well
clearance (FR)
Wrear axel
wheel well
wrornl clearance (RR)
Engine Type
Engine Size
Ballast Transmission Type:
Automatic or Manual
Mass Distribution FWD or RWD or 4WD
LF RF
LR RR
Tra c to r:
T ra ile r:
At2
Wheel
Ballast
\-,
M H
B o E K
Ml 2 3
N T
o U
J P
v
L
M1
M2
M3
M4
M.þ
MT
Tractor:
VIN No. M ake: M odel
Year: Odom eter:
Tra iler:
Flflh Wheel
Ballast
c K
M1 M4 M5
D a
E L R
F s
M1
M2
M3
M4
M5
MT
Vehicular impact speed and angle, impact point on the vehicle and the test alticle, exit speed and an-
gle, vehicular accelerations, and the three-dimensional motion of the vehicle should be measured and
reported. Exit speed and angle are measured at the time the vehicle loses contact with the test article.
For some suppoft structures such as a yielding sign support, the test article, or parts thereof, may stay
with the vehicle for a considerable distance beyond impact. In these situations, the exit conditions can
be measured when the vehicle clears the foundation or the footing of the test article. Note however tlrat
the point at which vehicle exit conditions are reported must be clearly defined. It is also recommended
that vehicle deformation, both in the interior and on the exterior, be measured and recorded. Procedures
for measuring vehicle deformations are presented in Appendix E. Other vehicle parameters that should
be measured and recorded include contact length with the banier, airborne distance of the vehicle, and
maximum rise of the vehicle's bumper and wheels.
Due to drop height limitations, pendulums are restricted to low-speed tests. However, since low-speed
testing is critical for most heavy breakaway devices, pendulums have been widely used in the evalua-
tion of the impact performance of breakaway luminaire poles. Bogie vehicles can be used for low- and
high-speed tests of breakaway devices to evaluate their impact performance. For these tests, change in
vehicle velocity during the crash and the impulse applied to the impacting vehicle by the breakaway
device are the primary measures of impact performance. Velocity change during a test of a breakaway
device has been shown to be corelated to the stiffness of the front of a test vehicle (127).The crushable
nose system now used on most suffogate test vehicles was calibrated to replicate the stiffness of a 1981
Volkswagen Rabbit. This vehicle does not meet the requirements of Section 4.2.1, and therefore, a new
crushable nose mrìst be developed to comply with the guidelines contained herein. Note that, as indi-
cated in References 4l and 93, any new surrogate vehicles should be developed to replicate a specific
production vehicle, not a generic vehicle.
Pendulums and most bogie vehicles do not have the capability to assess the effects of test article impact
with the windshield, roof or undercauiage of a vehicle. However, these vehicles can be used to identifo
the potential for such contact to occur. If contact is indicated, the test should be repeated with a produc-
tion vehicle. However, if no opportunity for contact is observed during surrogate vehicle testing, it is
reasonable to assume that no contact would occur during a full-scale crash test with a production car.
Note that bogie vehicles are beginning to be manufactured with a hood and windshield from a produc-
tion vehicle. These new bogie vehicles can be used to evaluate the effects of test article contact with the
hood and windshield area. However, these updated bogie vehicles still cannot be used to evaluate roof
crush associated with heavy breakaway devices falling on top of the vehicle or contact with the under-
carriage ofa test vehicle.
Heavier support trucks reduce the roll-ahead effect and represent a more critical condition for TMA de-
sign. However, designing TMAs for heavier support vehicles increases the required length of the TMA
and thereby increases its cost. Developers must therefore determine the maximum weight of the support
truck that can be used with each TMA design. All TMA testing must then be conducted with the maxi-
mum allowable support vehicle weight. A TMA designed for an unlimited support truck weight should
be tested with the support truck blocked against any roll ahead.
Although any vehicle can be used to support a TMA, in practice, dump trucks are used most often.
Therefore, it is recommended that, unless the attenuator is designed for attachment to a very light sup-
port truck, TMA testing should be conducted with an appropriately sized dump truck. The support truck
should be placed in second gear with the parking brake engaged. Any necessary ballast should be placed
in a manner to preclude movement during a high-speed impact with the rear of the vehicle. Ballast
should either be securely anchored to the bed ofthe support truck or be placed at the back ofthe bed to
minimize shifting during the test.
Documentation of support trucks should include year, make, and model of the truck as well as ballast
mass, test inertial mass, and the horizontal c. g. location of the ballasted vehicle. The type, location, and
manner of securing the ballast should also be reported. Finally, the condition and type of brakes used to
reduce vehicle roll ahead should be documented.
Vehicle instrumentation is necessary to accurately measure vehicle accelerations and rotations during
and after impact with a roadside safety feature. Tri-axial accelerometers are used to measure vehicle
accelerations and a single-rate gyro is used to measure vehicle rotational rates. The following sections
describe the specifications for electronic and photographic instrumentation of test vehicles.
tions. These specifications are reproduced in Appendi4 C. Part I relates to electronic instrumentation
and Part 2 relates to photographic instrumentation. Paft I of the specifications applies to the collection
and analysis of vehicle acceleration and rotation data; determination of impact speed; and measure-
ment, recording, and reduction of surrogate occupant data, including specifications for data channel
performance requirements, data channel selection, transducer mounting, sign convention, digital data
processing, timing marks, time of contact, and presentation of results. Part 2 of the standard addresses
the use of photographic instrumentation to collect and analyze the same data.
Table I of J2 I I Part I presents a list of recommended Channel Frequency Classes (CFCs) for various
applications. Note that determination of vehicle velocity changes and occupant displacements are critical
to the impact performance evaluation process. As shown in Table l, a CFC of 180 Hz is recommend-
ed for integration for velocity or displacement. Hence, accelerometers placed in a vehicle should, at a
minimum, adhere to CFC Class 180 requirements and the data should be filtered at this level prior to
processing. It should be further noted that, in Section 8.1, the standard recommends a pre-sample filter
conesponding to CFC 1000. Further, under Section 8.2, the standard indicates that the sampling rate
should be no less than l0 times the pre-sampling ûlter frequency, or at least 10,000 Hz.
It is important to note that SAE J2ll-l'JUL2007 andJ2ll-2NOV2008 do not speciff maximum trans-
ducer amplitude, called the Channel Amplitude Class (CAC). Instead, the standard leaves the selection
of the CAC to the testing agency. Maximum transducer accuracy is achieved when recording near
the mid-range of its amplitude. However, peak accelerations are often measured to be many times the
average values during the primary impact sequence. Subjecting an accelerometer to accelerations higher
than the CAC can potentially damage the unit and often cause any accelerometer data collected there-
after to be comrpted. The CAC should be selected to provide maximum accuracy without exposing the
transducer to undue risk of damage. Users should examine the results of similar tests when selecting a
CAC to meet this goal.
Although SAE J2 I I - I JUL2007 and J2l I -2 NOV2008 do not speciSr specific coordinate axes, they do
recommend that a right-handed coordinate system be used. However, in order to rnaintain consistency
with NCHRP Report 350 guidelines, the following righrhanded sign convention is recommended:
These sign conventions should be followed whenever acceleration, velocity, or displacement data must
be presented. Note that some testing agencies tend to reverse acceleration signs in order to produce a
plot that is primarily in the first quadrant. Although these changes are relatively simple to implement,
the revisions can lead to confusion. Therefore, it is recommended that the sign convention shown above
and in Figure 4-6 be utilized whenever any acceleration, velocity, or displacement data is presented.
Note that, as shown in Figure 4-6, the origin of the coordinate system is placed at the vehicle center of
gravity. Further, it is recommended that vehicle accelerations be filtered aT 60 Hz for the purposes of
presentation of accelerometer data in reports.
*ßo\\
+*'
Ø
+Y +Pitch
*1a'ü'l
xL,
SAE J2l l-1 JUL2007 also includes recommendations for the collection of vehicle crush data. For
purposes of impact performance evaluation, it is recommended that vehicle crush data be collected and
reported in accordance with Section 4.2.1.4 of this chapter and guidelines shown in Appendix E.
Other exceptions to the application of SAE l2ll-l JUL2007 and l2ll-2 NOV2008 include
Most vehicle components vibrate during a high-speed crash test. Because structural vibrations do not
contribute to overall vehicle motion, it is desirable to avoid mounting accelerorneters in a location
where these vibrations will contribute significantly to the recorded data. Therefore, it is recommend-
ed that accelerometers be mounted directly to a structural element of the vehicle in order to assure a
high stiffness of the mounting point. Most agencies mount transducers on the vehicle floor board. Care
should be taken to avoid mounting transducers to poltions of the floor board that could be expected to
deform during the crash. It is inappropriate to mount transducers to a frame that offsets accelerometers
from the structural components of the vehicle. The inherent flexibility of such a frame creates a struc-
tural filter that produces an unknown phase shift and spreading ofthe acceleration trace. Further, it is
desirable to combine accelerometers on a common structure, such as a metal block. Howeve¡ the mass
of the metal block should be minimized in order to reduce the effects of the added mass on the vibration
of the structural element to which it is attached.
For all test vehicles , a rate gyro should also be mounted as near to the center of gravity as possible. Like
accelerometers, the rate gyro should be attached to a stiff structural element to prevent local rotations of
the vehicle structure from distorting the gyro data. It is recommended that two rate gyros be mounted on
heavy truck test vehicles, one on the cab or tractor and another in the cargo area or on the trailer. Rate
gyro data is collected about the local vehicle axes and therefore the measured roll, pitch, and yaw rates
are coupled. This data must be decoupled using one of the commonly available procedures, such as the
Test Risk Assessment Program (TRAP) (155). Note that the rate gyro transducers should be selected to
preclude the possibility of test rotational rates from exceeding the maximum range. Note that rotational
rates during full-scale crash tests can sometimes exceed 1200 degrees per second.
Evaluation Criteria 5
S.l GENERAL
o the extent possible and practicable, limiting values recommended for the respective evaluation
f
I criteria are based on current technology and, when necessary on the collective judgment of experts
in roadside safety design. Establishment of performance criteria has been based on a "state-of-the-
practical" philosophy since the late 1960s. That philosophy basically contends that as technological and
economic conditions permit, higher levels of impact performance should be expected from some safety
features than from others. Thus, impact performance requirements of a breakaway sign, luminaire
support, or work-zone traffic control device are more demanding than a crash cushion. Recommended
values were also made in consideration of the limitations of the recommended test procedures and
methodologies used to estimate occupant risk. As a consequence and in view of the very complex nature
of vehiculat collisions and the dynamic responses of an occupa¡t to the collision, as well as human
tolerances to impact, the recommended criteria should be treated as general guidelines and not as
absolute criteria. The adequacy of these or other criteria would ultimateìy be established by the agency
responsible for the implementation of the safefy device being evaluated.
Note that the evaluation criteria relate only to the impact performance of the safety feature. Costs,
aesthetics, maintainability, durability, and other service requirements are not evaluated. It is recognized
that user agencies seldom nrake safety hardware implementation decisions based solely upon impact
performance evaluation and that these other factors are also considered.
Recommended impact performance evaluation criteria for safety features are given in Tables 5-lA
through 5- I C. Three dynamic performance evaluation factors are given together with recommended
evaluation criteria and applicable tests. The factors are: (l) structural adequacy, (2) occupant risk, and
(3) post-i mpact vehicular response.
Evaluation
Evaluation Criteria Applicable Tests
Factors
A. Test article should contain and redirect the vehicle or bring
10, 11, 12,20,21,22,
the vehicle to a controlled stop; the vehicle should not
30a,31a,32a,33a, 34a, 35,
penetrate, underride, or override the installation although
Structural 36, 37", 38a
controlled lateral deflection of the test article is acceptable.
Adequacy
(see B. The test article should readily activate in a predictable 60, 61, 62, 70, 7'1, 72, 90,
Section manner by breaking away, fracturing, or yielding. 81,82
5.2.1)
30b, 31 b, 32b, 33b, 34b,
C. Acceptable test article performance may be by redirection,
37b,39b, 40,41,42,43,
controlled penetration, or controlled stopping of the vehicle.
44,50,51,52,53,90, 91
10 fVs 16 fUs
Longitudinal 60, 61, 62,70,71 ,72
(3.0 m/s) (4.9 m/s)
l. The occupant ridedown acceleration (see Appendix A,
Section A5.2.2 for calculation procedure) should satisfy the
following limits:
Evaluation
Evâluation Criteria Applicable Tests
Factors
J. through M. Reserved.
Post-lmpact 30b,31b,32b, 33b, 34b,
Vehicular 37b,39b, 40,41,42,43,
Response (see N. Vehicle trajectory behind the test article is acceptable 44,45,60,
Section 5.2.3) 61,70,71,72,80,
81, 82, 90, 91
a Non-gating terminals and crash cushions
b Gating terminals and crash cushions
Structural adequacy is generally the first factor to be evaluated, and the safety feature should perform
successfully according to the requirements presented in Table 5-lA. Depending on its intended func-
tion, the feature may satisf, structural adequacy by redirecting the vehicle, by stopping the vehicle in a
controlled manner, or by permitting the vehicle to break through the device.
It should be noted that structural adequacy criteria refer to the structural requi¡ements associated with
the impact itself and not other structural aspects ofthe device. For example, the criteria do not im-
ply that a sign support system that meets the structural adequacy requirements of a test will meet the
structural adequacy requirements of wind and ice loads or other environmental considerations when
applicable.
Wide use is now made of temporaly longitudinal barriers in wolk zones. Free-standing, unanchored pre-
cast concrete bamiers are most commonly used. Increasing use is being made of a "movable," precast
concrete barrier in work zones or to separate traffic in high-occupancy vehicle lanes. Vy'hen used for the
latter, the movable barrier is typically moved laterally from one lane to another, one or more times per
day. A primary concem for barriers of this type is the deflection they undergo during a vehicular im-
pact. Because the amount a given installation can deflect without adverse consequences depends on site
conditions, it is not feasible to establish limiting deflection values for crash tests of these barriers. Thus,
it is important to accurately measure and report barrier displacement that occurs during the test so that
a user agency can make an objective assessment of the appropriateness of the barrier for its intended
application.
Similarly, the distance that an impacting vehicle extends over a barrier or penetrates into the system can
affect its impact performance, and this effect is highly dependent upon the placement of other hazards in
the vicinity of the barier. A barrier's working width is therefore measured and reported as a means for
allowing highway designers to avoid placing a barrier too near an obstruction.
Risk of occupant injury during impact with a highway safety feature depends, to a large extent, on the
crashworthiness of the impacting vehicle. In tnrn, crashwofthiness depends on the design of the occu-
pant compartment, including factors such as structural integrity, padding, restraint system, and so on.
However, to the extent possible, the variability of vehicular crashworthiness has been removed from
safety feature evaluation. Occupant risk is appraised according to gross vehicular accelerations because
they are primarily functions of the safety feature design and the external structural design of the test ve-
hicle. V/hereas the highway engineer is ultimately concerned with safety of the vehicle's occupants, the
occupant risk criteria ofTable 5-1B should be considered as guidelines for generally acceptable dynam-
ic performance. Note that the optional Criterion J included in NCHRP Report 350 (129) on the use of an
instrumented Hybrid III dummy has been deleted from the evaluation criteria.
Detached elements, fragments, or other debris from the test article should not penetrate or show po-
tential for penetrating the occupant compartment, or present likelihood to impact with other traffic,
pedestrians, or workers in a construction zone if applicable. The degree to which detached elements,
fiagments, or other debris and the degree to which the displacement of a temporary barrier puts other
traffic, pedestrians, and workers at risk in a construction zone will depend on the location of the fea-
ture and the impact conditions. A sign that has a tendency for detached elements to scatter over a wide
area upon impact may not be the appropriate design for use in the median of divided highway since the
detached elements could potentially encroach into opposing traffic lanes. On the other hand, the same
sign when struck on the roadside may be of little concem except to occupants of the impacting vehicle.
Fragments and debris from an impact with a traffic control device in a construction zone may or may
not be a consideration for workers in the work zone, depending on their location relative to the device
and the impact conditions. Consequently, it is not practical to establish absolute limits on test article
trajectory, debris scatter, or barrier displacement. Rathe¡ ít is important to accurately record and repolt
test article trajectory and debris scatter so that a user agency can make an objective assessment ofthe
appropriateness ofthe safety feature for the intended application.
A clear distinction should be made between: (a) penetration, in which a component of the test article
actually penetrates into the occupant compartment; and (b) intrusion or deformation, in which the
occupant compaftment is deformed and reduced in size, but no actual penetration is observed. No pene-
tration by any element of the test article into the occupant compaftment is allowed. As for deformation
or intrusion, the extent of deformation varies by area of the vehicle damaged and should be limited as
follows:
' Wheel/foot well and toe pan areas < 9 in. (229 mm).
' Side front panel (forward ofA-pillar) < 12 in. (305 mm).
' Front side door area (above seat) < 9 in. (229 mm).
' Front side door area (below seat) < l2 in. (305 mm).
' Floor pan and transmission tunnel areas { l2 in. (305 mm).
Note that a tear in the windshield plastic liner is not relevant if there is no potential for a test article
component to penetrate into the vehicle. For example , a tear produced by a breakaway luminaire pole
falling on the vehicle's roof would not be considered a cause for failing the test. As another example, a
continuous, flexible cable element may contact and deform the A-pillar of an impacting vehicle within
acceptable limits and cause minor tearing of the windshield's plastic liner.
During oblique vehicular impacts into tall, flexible cable median barriers, side windows may fracture
due to contact with continuous cable elements. Contact with the longitudinal cable elements may also
result in localized deformations to the roof's structural support members (i.e., the A- and/or B-pillars).
In such circumstances, the risk of significant intrusion into the occupant compartment and contact
between the cable elements and the head of a vehicle occupant is small. The relatively low vehicle
decelerations experienced during a vehicular impact with a flexible cable barrier will typically result
in only minor lateral movement of the occupant, and the occupant's head is not expected to be par-
tially ejected outside of a side window. Thus, in these instances, it may be reasonable to relax the side
window fracture criteria as long as several conditions are met. First, there should be no demonstrated
risk for the ends of rail elements to spear into the occupant compartment. Second, the continuous cable
elenrents should not laterally protrude more than 3 in- (76 mm) into the side window region based on
observation of the plastic deformation to the A- and/or B-pillars. Finally, the A- and B-pillars should not
be completely severed.
Note that some vehicles now incorporate glued seams on the floor board as well as other areas. In the
presence of significant deformation, these bonded seams can separate and create an opening into the
occupant compaftment. There is no available data to relate occupant injury severity to the opening of
seams in the floor pan area. However, it is generally believed that an opening in the occupant compart-
ment by and of itself does not necessarily result in injury to the occupants unless it is accompaníed by
an object moving toward the occupant. Therefore, a seam separation by itself is not considered a test
failure unless (1) a cornponent of the safety device protrudes through the opening or (2) the deformation
limits of l2 in. (305 mm) is exceeded.
It is essential that adequate documentation in the form of photographs and measurements of occupant
compartment damage be made and reported. Detailed pretest measurements should be made from at
least two reference points in the vehicle. Reference points should be widely spaced and selected from
locations unlikely to be deformed during the crash test. Interior and exterior photographs, as described
in Sections 4.2.1.4 and 4.2.1.6, should be taken prior to the test to permit direct conrparisons of before-
and after-test conditions.
Although not a specific factor in assessing test results, integrity of the test vehicle's fuel tank is a
potential concern. It is preferable that the fuel tank remains intact and not be punctured. Damage to, or
rupture of, the fuel tank, oil pan, or other features that might serve as a suffogate of a fuel tank should
be photographed and reported.
For the majority of tests, a key requirement for occupant risk evaluation is for the impacting vehicle to
remain upright during and after the collision. As an indication of vehicle stability, the maximum roll or
pitch angle of the vehicle during and after the impact sequence should not exceed 75 degrees. Although
it is preferable that all vehicles remain upright, this requirement is not applicable to tests involving
the 100005,36000V and 36000Tvehicles. SeeAppendixA, Section A2.2.1, for a discussion of these
exceptions.
Occupant risk is also assessed by the response ofa hypothetical, unrestrained front seat occupant whose
motion relative to the occupant compartment is dependent on vehicular accelerations. The "point mass"
occupant is assumed to move through space until striking a hypothetical instrument panel, windshield,
or side structure and subsequently is assumed to experience vehicular accelerations perpendicular to
the contact surface by remaining in contact with the interior surface. The two performance factors are:
(l) the longitudinal and lateral component of occupant velocity at impact with the associated
interior surface, and (2) the highest lateral and longitudinal component ofresultant vehicular acceler-
ation averaged over any l0-ms interval for the collision pulse subsequent to occupant impact with the
associated interior surface. These two performance factors are commonly referred to as the occupant
impact velocity (OIV) and the ridedown acceleration (RA), respectively. Methods for calculating the
occupant impact velocity and ridedown acceleration components are given in Appendix A, Section
1^5.2.2. Generally, lower values for these factors indicate that the safety features are more forgiving to
the occupants of the impacting vehicles. While a sumogate occupant is recommended in tests with all
I l00C and some 2270P vehicles, its dynamic and kinematic responses are not required or used in occu-
pant risk assessment. Hypothetical occupant compaftment impact velocities and ridedown accelerations
are calculated from vehicular accelerations.
It is also necessaly to assess the risk of injury to the driver of a supporting truck in a TMA system.
Because the types of impacts in this case are primarily unidirectional and the supporting truck is
accelerated forward, the driver will not move forward, at least initially, and is restrained from flailing
rearward by the seat and headlest, which should be standard on these vehicles. As such, the primary risk
of injury would stem from ridedown accelerations as the vehicle is accelerated forward.It is therefore
recommended that ridedown acceleration criteria be used as the primary assessment of the risk of injury
to the driver of a supporting truck in a TMA system.
Recommended limits for occupant impact velocity and ridedown acceleration are given in Table 5-lB.
Note that two values are given for each parameter, a "preferred" limit and a "maximum" limit. As
implied, it is desirable that the occupant risk indices not exceed the preferred values, and it is recom-
mended that they not exceed the maximum values. Reference should be made to Appendix A, Section
y'.s.2.2, for the rationale used in selecting these values. However, these limits are only recommendations
and a user agency may select different occupant risk limits if it so desires.
It should be noted that the measurements are associated with certain uncertainties and tolerances, which
may vary among testing agencies. The testing agencies are required to report on these measurement
uncertainties and tolerances for consideration by the user agency. Also, there is a degree ofprecision
or rounding errors associated with the measurements. It is recommended that the maximum allowable
limits are 40 fVs (12.2 nls) for occupant impact velocity and 20.49 G for the ridedown acceleration.
Although not required, testing agencies are encouraged to calculate and report the Theoretical Head
Impact Velocity (THIV), the Post-Impact Head Deceleration (PHD), and the Acceleration Severity
Index (ASI), as described in Appendix F. The THIY PHD, and ASI have been adopted by the European
Committee for Standardization (CEN) (140-142) as measures of occupant risks. At some time in the
fi;ture, the U. S. and CEN may develop common impact performance standards for highway features.
By calculating and reporting the THIV PHD, and ASI, a database will be developed from which
comparisons can be made relative to the flail space model and from which decisions can be made as to
appropriate measures of occupant risk.
Post-impact vehicular response is a measure of the potential of the vehicle to result in a secondary colli-
sion with other vehicles and/or fixed objects, thereby increasing the risk of injury to the occupants of the
impacting vehicle and/or other vehicles. Note that NCHRP Report 350 (129) has fow evaluation criteria.
However, Criteria K, L, and M were found to be non-discriminating and served little purpose. Thus,
these evaluation criteria have been deleted, Ieaving only Criterion N. To avoid conñlsion and maintain
continuity from one set of guidelines to another, the evaluation criteria were not re-lettered.
Penetration ofthe vehicle behind the test arlicle is unacceptable, except for certain classes ofsafety
features as listed in Criterion "N", for which vehicular trajectory behind the test article is acceptable.
Note that acceptable post-impact behavior may also be achieved if the vehicle is decelerated to a stop
while vehicle-barrier contact is maintained, provided all other relevant criteria of Table 5-lC are satis-
fied. However, it should be kept in mind that, if the barrier is within a lane width of adjacent traffic, the
slowed or stopped vehicle may pose risks to oncoming motorists.
Excessive pocketing or snagging ofthe vehicle and the resulting post-inrpact trajectory such as a high
vehicular exit angle or spin-out of the vehicle, are not desirable and should be documented. It is prefera-
ble that the vehicle be smoothly redirected (for redirective devices), and this is typically indicated when
the vehicle leaves the barrier within the "exit box." The concept of the exit box, adapted directly from
the CEN standards, is defined by the initial traffic face of the banier and a line parallel to the initial
traffic face of the barrier, at a distance A plus the width of the vehicle plus l6 percent of the length of
the vehicle, starting at the final intersection (break) of the wheel track with the initial traffic face of the
barrier for a distance of B. All wheeltracks of the vehicle should not cross the parallel line within the
distance B. Figure 5- I provides a graphic definition of the exit box and defines distances A and B for
passenger cars and other vehicles.
Vehicle rebound has been noted for some reusable crash cushions. Given that reusable crash cushions
are typically used in narow gore areas with heavy traffic volume, excessive vehicle rebound could
increase the potential for secondary collisions. In order to provide user agencies with the necessary
information regarding use and placement of such crash cushions, testing agencies are required to docu-
ment and repo¡t the rebound velocity and point of final rest. The rebound velocity is the velocity of the
vehicle center of gravity at the instant any wheel of the vehicle encroaches on the rebound line, which
is defined as a line perpendicular to the centerline of the crash cushion and20 ft (6.0 m) in front of the
nose of the crash cushion. Braking of the vehicle should be at least 2 seconds after initial impact and
beyond the rebound line. The final resting position of the vehicle should be documented relative to the
nose and centerline of the crash cushion and video coverage should continue until the impacting vehicle
comes to a complete stop.
During crash tests into flexible barrier systems (e.g., cable baniers), vehicles may travel significantly
behind downstream support posts and either ride over or under one or more of the longitudinal cable
elements. In such a scenario, there is increased opportunity for a cable element and/or support post to
contact and/or become entangled with components of the vehicle guidance system and result in ve-
hicle instability or rollover. For example, when a cable guidance system is used, a steel hub is often
attached to the front wheel opposite impact to aid in vehicle guidance prior to impact with the barrier.
In a situation where an externally-affixed component of the guidance system (e.g., steel guidance hub
protruding from front wheel) clearly interacts with the barrier system and causes an undesirable event
(e.g., rollover) to occur as the vehicle is exiting the barrier after an otherwise successful containment
and redirection, consideration should be given to relaxing the criteria that was violated. As part of the
evaluation ofsuch an event, the testing laboratory should provide clear and indisputable evidence that
barrier contact with the external vehicle guidance component led to the undesirable outcome.
Vw = Vehicle Width
Vl = Vehicle Length
B
Exit Box
nitialïraffic Face of Barrier
Tests of a geometric feature, such as a ditch, driveway, embankment, or curb, typically involve three-di-
mensional vehicular motions. Also, the duration of the test is usually long (up to 5 s or more) in
.. comparison to a test of a barrier or other highway safety feature (typically 0.30 s or less). As a con-
sequence, an unrestrained occupant can be expected to flail about the occupant compartment in three
dimensions over extended time duration, possibly contacting a given surface(s) more than once. For
these reasons, the OIV and RA values are generally not applicable. Fortunately, in tests of most geomet-
ric features, there are no design elements that would cause sudden and large vehicular velocity changes.
The primary concem is overtum of the vehicle as it traverses the feature.
In the absence of more objective criteria, the following procedures and evaluation criteria may be used
for a geometric feature:
Test
Documentation
(^) f Rrimary importance in the guidelines presented herein is the preparation of a comprehensive test
\r.-,f report. The test(s) should be documented in sufficient detail so that, if necessary others could re-
peat the test(s) and obtain similar results. Liberal use of photographs is encouraged to document before,
during, and after test conditions. Reference should be made to Chapter 3 for key pre-test, test, and post-
test parameters.
The test report should be prepared as a formal technical report. The guidelines presented herein are in-
tended as minimum requirements. A testing agency may, at its own discretion, include additional details
in its test reports. Also, individual user agencies may have different reporting requirements, and it may
be advisable to review such requirements to ensure that they are covered in the test repofts.
The title page should include the title of the report, name(s) and affiliation(s) of the authors, name and
address of the test facility, report number, sponsoring agency, and the date of the report. The fed-
eral Technical Report Docunlentation Page should be completed and included as part of the reporl.
However, disclaimer statement and acknowledgments are optional and may be included if applicable.
System Details" The test article and test installation should be fully described with engineering draw-
ings and material specifications. Key parameters that should be recorded are given in Section 3.4. The
engineering drawings shall be drawn to scale with appropriate dimensions and labeling to facilitate their
use by others. It is desirable to have the drawings prepared in an electronic format, suitable for inclu-
sion into the updates of A Guide to Standardized Highray Barrier Rail Hardware (9) or A Guide to
Small Sign Support Hardware ( I 0). For test articles with multiple design drawings, only the system and
installation drawings should be included in the main report. The remaining design drawings should be
included as an appendix to the repoft.
As appropriate, any revisions to the design made during the course of the test program should be fully
documented. The documentation should include a detailed description of the revisions and which tests
are conducted with the revised design. Also, any special fabrication and installation procedures (such as
heat treatment, weldments, bolt tension, galvanizing in critical stressed areas, etc.) that may influence
impact performance should be delineated.
For test articles installed in soil, the soil conditions should be carefully documented, including type of
soil used; why it was selected; its adherence to recommended specifications; installation procedures
(e.g., drill and backfill, trench, etc.); dimensions of backfill (length, width, and depth; or diameter and
depth); and any special installation details. Furthermore, the soil properties and strength as measured at
the time of test should be documented with each crash test as well as the latest soil strength calibration
results (see Figures 3-l and 3-2 for details on reporting requirements on soil strength).
Test Requírements tnd Evalustion Critería. A discussion of the test requirements should be presented,
including the required crash test matrix and the actual crash tests conducted. Ifan abbreviated crash test
matrix is used, the rationale for not conducting certain crash tests should be explained. Any prior testing
of the device or feature should be referenced and the results summarized.
Also, the evaluation criteria for assessment of the crash test results should be delineated for each of the
crash tests conducted.
Test Conditions. Detailed documentation of the test conditions is presented in this chapter, including
test facility, vehicle tow and guidance system, propefties of the test vehicles, and the data acquisition
systems (both electronic and photographic).Any dimensions of the test vehicle(s) that fall outside of the
recommended tolerances should be noted in the test report. The data acquisition systems should be fully
desclibed, together with the procedures used in calibrating and processing the data.
Crosh Test Descriptiorts. Detailed documentation of each crash test conducted should be provided
in separate chapters, including the impact sequence and post-irnpact vehicle trajectory system and
component damage, vehicle damage, occupant risk values, and a discussion summarizing the test and
evaluation results. Test documentation should include a sunìmary sheet of the test results that presents
an overview of the crash test results with four or five sequential photographs. Additional docunrentation
should include more closely spaced sequential photographs of all important camera views; documentary
photographs showing the impact location, system damage, and vehicle damage; and tables and figures
illustrating the extent of test article and vehicle deformation. It is desirable that weather conditions that
may affect test results be reported, including those at the time of the test as well as those preceding the
test (e.g., the extent of subfreezing or rainy weather).
Conclusíons and Recommendations.A summary of all test results in both narrative and tabular forms
should be provided with appropriate conclusions and recommendations.
References. A list ofany references used or cited in the test report should be provided.
Appendíces. Materials too bulþ to be included in the main report should be presented in the appen-
dices. Some typical items include detailed design drawings of the test article, occupant compartment
deformation measurements, accelerometer data analysis, and documentation of in-situ soil test results
for each ofthe crash tests.
It is recommended that the repoft contain, for each crash test, a summary page with the following infor-
mation, as shown in Figure 6-1. An example of the use of Figure 6-l is shown in Figure 6-2. Note that
data items that are not applicable may be omitted from the summary page, as deemed appropriate.
2. Plan view-Provide the plan view of the installation, showing the overall layout of the installation,
the impact point of the vehicle with the test article, and the post-impact trajectory of the vehicle and
the test article.
4. General information-Report the test agency, test number, and date of the test.
5. Test Article--Describe, to the extent possible and as space permits, the test article:
' Type-Identiff basic type of article tested (e.g., longitudinal barrier/bridge rail).
' Installation length-Report the installation length of the test article. For a longitudinal banier
test, this is the length of the standard barrier section, exclusive of the end terminals. For a ter-
minal test, it is the length of the terminal and the adjoining longitudinal barrier; lengths of each
should be given separately. For a crash cushion test, it is the length ofthe cùshion and backup
structure, if necessary; lengths of each should be given separately. For a TMA, it is the length of
the cushion and the srìpport truck; Iengths of each should be given separately. It is not applica-
ble for tests of support structures, work zone traffic control devices, and breakaway utility poles.
' Key elements-Provide descriptions of the key elements of the test afticle, such as rails, posts,
support structures, etc., and include, as appropriate, sizeldimension and material properties.
6. Soil conditions--Report, if applicable, the soil conditions for the test article:
' Type of soil-Report the type of soil used and a note should be made if there are unusual soil
conditions different from those recommended in Section 3.3.1.
' Soil strength-Repoft the measured soil strength at (l) initial calibration tests and (2) time of
the crash test.
9. Exit conditions-The exit conditions should be measured at the time the vehicle loses contact with
the test article (see Section 5.2.3 for fuilher discussion), including:
' Speed-actual exit speed film analysis.
as determined from
' Angle-actual exit angle as determined from film analysis. Also, report the vehicle heading
angle if the vehicle is in a non-tracking mode at exit.
' Report whether the exit box criterion is met or not.
10. Post-impact trajectory-Report the stability and the stopping distance (or point of final rest rela-
tive to the initial point of impact) of the vehicle. Also, report the rebound velocity of the vehicle if
applicable.
I l. Occupant risk values-Occupant risk values are computed as described in Section A5.2.2 of
Appendix A and in Appendix F, including:
' Occupant Impact Velocity (OIV), and
' Ridedown Acceleratìon (RA).
12. Test article damage-Provide a brief summary of the damage to the test afticle.
I3 . Test article deflections-Report the permanent and dynamic deflections of the test article plus the
working width during impact. These measurements normally apply to longitudinal ban'iers, termi-
nals, crash cushions, and TMAs. Permanent deflection is the residual lateral displacement of the
test article remaining after the impact. Dynamic deflection is the maximum lateral displacement
of the test article on the traffic side that occurs during the impact. The working width is the maxi-
mum dynamic lateral position of any major palt of the system or vehicle. These measurements are
all relative to the pre-impact traffic face of the test article. For the working width, the height of the
maximum working width should also be documented and reported.
14. Vehicle damage-Repoft damage to the vehicle (see Section 4.2.1.4 for discussion ofVehicle
Damage Scale (VDS) and Collision Damage Classification (CDC). See Section 5.2.2 andAppendix
E for discussion of measurement of occupant compaftment deformation), including:
If available, report Event Data Recorder (EDR) data and airbag deployment information as an appendix
to the test report.
As part of the documentation, a video of the test(s) should be prepared. The video should include a
title block identiSing the test, test conditions, date, sponsoring agency, before-and-after documentary
coverage of the test article and vehicle, and high-speed views of the impact (both profile and overhead).
Also, it is important that the test report contain an ample number of photographs of the pretest, test, and
post-test conditions.
6.1.4 ASSESSMENT
For each crash test conducted, the impact performance of the test article should be discussed with
regard to the three evaluation factors: structural adequacy, occupant risk, and post-impact vehicular
trajectory. It is recommended that an assessment summary page, as shown in Table 6-3,be prepared to
address each relevant evaluation criterion of Table 5-1. Note that example entries shown in Table 6-3
are for illustrative purposes only and are not all-inclusive.
In case of multiple crash tests, a table summarizing the results of the individual tests should be provid-
ed, as shown in Table 6-4. Again, the example entries shown in Table 6-4 are for illustrative purposes
only and are not all-inclusive.
Finally, a conclusion should be presented as to acceptability of the impact performance of the test
article based on results of the crash test(s). Recommendations should be offered as to modifications
that may improve the impact performance and cost-effectiveness of the test article. Recommendations
should be categorized as either desirable or essential. Known or predictable limitations of the test
article, such as sensitivity to foundation conditions or effects of improper orientation of the test article,
should be discussed. Recommended applications may also be identified.
Photography
Before and after of test vehicle Photos
stiil
and test article
Selected sequential frames (8 Photos
High-Speed CineMdeo
minimum) during impact
Vehicle Accelerations
x,y, z components; filtered
Plotsa
(see Section 4.3.2)
Damage
TestArticle
Deformation
Permanent Length ofcontact Narrative
Profile of deformation Plotaffable
Dynamic Magnitude and location of maxi- Photos and Narrative
mum deformation
Key Elements Damage to key elements Photos and Narrative
Vehicle
Exterior Damage to exterior Photos and Narrative
VDSBb and CDSb Scales and
NASSb Measurement Tech
lnterior Damage to interior Photos and Narrative
Deformation Measurement
Magnitude and location of maxi- Photos and Narrative
mum deformation
Optional
Dummy
Accelerations x, y, z components of head and Plotsa
chest; (see Section 4.3.2)
Femur Load Femur load cell Plotsa
Seat Belt . Seat belt load cell Plotsa
Other lnstrumentation Load cell, strain gauge, etc. Plotsa
TABLE G3. Example of Recommended Assessment Summary Page for lndividual Crash Tests
Test No.: EXP-1 Date: April 12,2005 TestAgency: Midwest Roadside Safety Facility
Vehicle Trajectory
N. Vehicle trajectory behind the test article is acceptable. Not applicable Pass
Note: Table entries are not all-inclusive and are for illustration purposes only. This table will typically require more than one page to
complete.
TABLE G4. Example of Recommended Assessment Summary Page for Multiple Crash Tests
Evaluation
Evaluation Factors Test EXP-1 Test EXP-2 Test EXP-3
Criteria
A N/A S N/A
Structural Adequacy
c S N/A S
D S S S
F S S S
Occupant Risk
H S N/A N/A
I S N/A N/A
L N/A S S
Post-lmpact Vehicular
M N/A S S
Response
N S N/A S
o . TestNumber Angle
0) o Date ¡ Exit Box Criterion
an 5. Test Article: I 0. Post-Impact Trajectory:
J
. Type oVehicle Stability
-'l
o oStopPing Distance
Installation Length
an
Key Elenrents 11. OccupantRisk:
,) 6. Soil Conditions: ¡Longitudinal OIV
o Type of Soil oLateral OIV
at,
É Soil Strength . LongitudinalRA
U, 7 Test Vehicle: . LateralRA
Type/Designation 12 Test Article Daruge:
Make and Model 13. Test Article Defleotions:
Test Inertial
¡ Permanent
Gross Static
. Dynamic
8. ImpactConditions: . Working Width
t SPeed 14. Vehicle Damage:
. Angle . VDS
¡ Locatior/Orientation . cDc
o Max. Deformation
o
? . --..-¡{--}t-:--
!.,
m
xÐ
g3
o 0.000 sec 0.078 sec 0.2 l6 sec 0.400 sec 0.628 sec
o
Ð
o 258 ft [78640.04 mm]
o
o
3 2 3 I ott2fJl{ l6
3
o
--=-
9 fi 10 in. [3006.70
J
CL ïre Marks
o l'ire Marks o
a. LF ïre
Ø
3
. Test Agency MwRSF . Posþimpact Trajectory
3 . Test Nunrber ...-22l4TB-2 Vehicle Stability Satisfactory
ql . Date tDlt2/04 Stopping Distance..........-.. 25 ft(78.6 m) downstream
(t . Test Article.. Free-Standing Temporary Barrier 9.8 fr (3.0 m) laterally behind
J o Total Length 200 ft (62.34 m) ¡ Vehicle Snagging...................... None
o . Key Elements - Barrier o Vehicle Pocketing..................... None
o
Description F-Shape Barrier . Occupant Impact Velocity
o Length.. 150 in. (3810 mm) Longitudinal..... 16.99 fl/s (5.18 rn/s) <39.4 ftls (12 m/s)
C) 22.5 in. (572mm) ¡ Lateral(notrequired).., 17.29ftls(5.27 mls)<39.4ft|s(t2mls)
o¡
th
32 in. (813 mm) r Occupant Ridedown Deceleration (10 msec avg.)
J Lotrgiturlinâ1..... 7.17 <20G
2270P
d
Ø 2002 Dodge Ram 1500 Quad Cab Pickup
Lateral (not required)........ 11.37 G
¡ TFIIV............ c)
f
....22ß0 fr/s (6.89 m/s) fl)
ut 4,777 tb (2t67 ke) . PHD............. .... ll.52C
Þ
o
o
UI
Test Inertial...... 5,000 tb (2268kg) ¡ Test A¡ticle Damage ................. Moderate
c Gross Static...... 5,000 lb (2268ke) O)
1¡ . Impact Conditions
o Test Article Deflections
Set.
Permanent 73 in. (1854 mm)
{o
I
o
Speed 62.0 mph (99.7 km/h) Dynamic........... 80 in. (2023 mm) o
Angle 25.4 deg Working Width................. 102 in. (2595 mm) o
o
Locatior/Orientation......... 4 ft (1.2 m) upstream barriers 8 & 9 joint ¡ Vehiole Damage........................ Moderate c
f
¡ Exit Conditions vDs l I-LFQ-3 o
J
Speed 48.5 mph (78.I km/h) cDc II.LYES4 0)
Angle............... 15 deg (estimated) Maximum Deformation ....4.25 tn. (108 nm) at front floor pan
4.
o
;N
l¡
ln-Serv¡ce
Performance Eval uation
T.l PURPOSE
As mentioned in Chapter l, in-service performance evaluation (ISPE) is a very important step in the
assessment of the impact performance of a new or extensively modified safety feature. The purpose of
in-service performance evaluation is to determine and document the manner in which the safety feature
performs during a broad range of collision, environmental, operational, and maintenance situations for
typical site and traffic conditions. The in-service performance evaluation remains an important fol-
low-up to the crash test experiments described in previous chapters. Testing and analysis only partially
assess the efficacy of a feature and a more thorough and in-depth knowledge of the feature is important
to its proper implementation.
Although the crash testing guidelines set forth in this report assure that safety devices function well for
the specified test conditions, there are many unknowns and concerns about the impact performance of
the roadside features under real-world conditions. Differences between field performance and crash test
results can arise due to many factors, including:
' Field impact conditions that are not included in crash test guidelines, such as non-tracking and side
impacts;
' Site conditions, such as roadside slopes and ditches, that adversely affect vehicle kinematics
before, during, or after impact with the safety device; and
' Sensitivity to installation details, such as soil resistance or barrier flare configuration.
Therefore, if necessary conduct an in-service performance evaluation to assess and monitor field per-
formance of roadside safety featules. In-seruice performance evaluation allows user agencies to identify
the overall impact performance of a feature as well as identifl potential weaknesses or problems with
the design.
The following sections describe goals and suggested procedures for in-service performance evaluation
However, the random and extremely complex nature of vehicular crashes coupled with resource lim-
itations of transportation agencies greatly restrict the extent to which these goals can be met and the
procedures can be carried out.
7.2 oBJECTTVES
l. Demonstrate that design goals are achieved in the field and identiff modifications that might im-
prove performance.
3. Identiff factors tha|may compromise or defeat a feature's perfiormance. Examples of such factors
include vulnerability of the feature to pilferage or vandalism, accelerated corrosion or degradation
of materials due to de-icing salts and other contaminants, and susceptibility to damage during snow
plowing or mowing operations.
5. Examine the influences that the feature may exhibit on other highway conditions that, in turn, may
adversely affect highway operations and traffic. Such factors to be monitored are traffic
congestion, change in crash rates or pattems, disruption of surface drainage, or the cause of snow or
debris buildup.
6. Acquire routine maintenance information. As a part of this effort, the feature's design and layout
should be examined for possible modifications that would lower installatìon, maintenance, and
damage repair costs. Problems encountered during routine maintenance and damage repair should
be documented and reported. Note that frequency of lepair and repair demand (after both nominal
and severe impacts) are critical factors. Systems that can sustain numerous or severe impacts while
remaining serviceable offer substantially better p'rotection to motorists than those that are rendered
out of service by virtually every impact. This becomes especially critical on high-volume roadways,
on roadways where maintenance activities cause congestion and increased risks of crashes, and at
problem or high-crash locations. Information of this type can become the primary consideration in
selection ofa barrier system for such locations.
Depending on the questions posed, in-service performance evaluation could involve different
approaches with varying degrees of detail. NCHRP Repoft 490,Ln-Service Perforntance of Traffic
ì Barriers, (l I 8) presents one such approach with detailed step-by-step procedures on conduct ofan
in-service performance evaluation. The procedures described in this report are intended mainly for
evaluation of a specific roadside feature; however, it can be expanded to include continuous monitoring
of several types of features as part of a long-term safety management system. This approach utiliz-
es maintenance forces as the main source of data collection, supplemented by data from police crash
repofts. Figure 7-1, reproduced frorn NCHRP Report 490, shows the various steps of this in-service
perfonnance evaluation process. Detailed procedures for each ofthese steps are outlined in the report
and will not be repeated herein.
Instead, a more general discussion on the conceptual framework of a comprehensive in-service perfor-
mance evaluation is presented herein. The conceptual framework covers not only evaluations using the
procedures detailed in NCHRP Report 490, but also other aspects of in-service performance evaluation.
In general, in-service performance evaluation includes two separate, but integrated, programs that
address different aspects of in-service performance evaluation:
The in-service performance evaluation of a new or extensively modified feature presents some unique
issues. First, the number of initial installations is typically very small for a new feature. Consequently,
the number of crashes involving these installations will also be very small. The small sample size
would limit the statistical significance of the results of the evaluation, and render it more anecdotal and
subjective in nature. To increase the sample size of crashes, the alternatives are to increase the number
of initial installations or increase the evaluation period. There is a practical limit to the length of the
evaluation pedod, e.g., three to five years. The more logical approach is to increase the number of initial
installations. However, as discussed previously, a state transportation agency may not want to wide-
ly deploy a new feature without any in-service evaluation. This dilemma is best resolved by pooling
resources among state transportation agencies that are interested in the same feature. Each agency could
install and monitor a small number of initial installations. The results can then be combined to provide
a larger sample size. This approach allows the agencies to keep the number of initial installations small
for each state, yet provides a sufficient sample size for evaluation.
Defìne Objectives
Develop Sarnpling
Profile
Examine
Crash Data
Estimate Hardware
Estimate Number of
Cases Needed and Planning &
Preparation
Analyze Data
Analysis
Evaluate Hafdwate
-
Pelformance
_l
Second, other than design drawings and prototype installations for the crash tests, there is no
real-world experience with installing or maintaining the feature. Even with the best designs, it is
reasonable to expect that there will be some unforeseen problems that need to be ironed out with the
initial installations. Thus, one of the objectives of the in-service performance evaluation for a new
feature is to identifu and resolve any problems associated with the installation and maintenance of the
feature.
Given that the true impact performance of a new or extensively modified feature under real-world
conditions is not known, it recommended that in-service performance evaluation of the feature be
conducted prior to widespread deployment of the feature.
The in-service performance evaluation of a new or extensively modified feature should include the
following components:
' Installation and maintenance checklist-The checklist includes a list of items related to the con-
struction, installation, and maintenance of the device. Any identified problem should be investigated
and documented, and the information should be forwarded to the designers for appropriate corrective
action.
' Inventory-The inventory should include location and design details of the installations so that the
information can be matched to the crash data for evaluation.
' Crash monitoring-The monitoring should include both reported and unreported crashes involv-
ing the installations. The crashes are identified from police notification or crash reports for reported
crashes and maintenance notification or logs for unreported crashes. For reported crashes, the
investigation should include: obtaining a copy of the police report; visiting the site to document any
available scene evidence, such as tire marks, damage to the installation, final rest position, etc.; and
taking photographs of the site. For unreported crashes, the investigation will be limited to documen-
tation of the site and system damage.
' In-depth investigation-Crashes involving the new feature that resulted in a fatal or serious injury
should be investigated in-depth. In addition to obtaining the police crash report and documenting the
site, the involved vehicle should be examined and efforts made to reconstruct the crash in terms of
impact configuration and conditions, e.g., point of impact, speed, angle, vehicle orientation, etc., and
to assess the impact performance of the feature, i.e., whether the device performed as designed and,
if not, whether there are extenuating circumstances.
The results of the in-service performance evaluation should be summarized in a report which should
include, but not be limited to:
However, the setup of the continuous monitoring program is very different from that of the new feature
evaluation system. With a new feature, the number of initial installations and the resulting crashes are
expected to be relatively small. Thus, monitoring of the new feature evaluation system is typically small
in scope and can even be managed manually. In comparison, the number of installations and associated
crashes are likely to be much greater with the full-scale deployment of a feature. This provides a much
larger sample size suitable for statistical analysis. Thus, the continuous monitoring system should be
computerized to keep it manageable and to minimize manpower requirements.
The continuous monitoring system would consist of the following three subsystems
The backbone of the continuous monitoring system is a computeized dalabase created by merging the
following data files:
' Highway and traffic data-Items of interest include such information as: highway type, firnctional
class, number of lanes, lane width, shoulder width, average daily traffic, percent truck, etc.
' Maintena¡ss ¡sss¡fls-There are two general areas of interest regarding the maintenance records.
First, the records are reviewed to identifu problems associated with maintenance of the device. Any
identified problem should be investigated and documented and the information forwarded to the
designers for appropriate corrective action. Second, the records are compiled to determine the extent
of unreported crashes, which is part of the evaluation of the impact performance of the feature.
' Roadside feature inventory data-The inventory data should include location and design details of
the installations so that the information can be matched to the crash data for evaluation.
' Crash data-Police repofted crashes involving the feature of interest are matched to the roadside
feature inventory data or by location on the highway.
The computerized database should be analyzed periodically, e.g., annually, for generalized trend analy-
sis and ploblem identification. The analyses could be route-specific (i.e., analyze crash or maintenance
records for all roadside devices on selected sections of highways), device-specific (i.e., analyze uash
or maintenance records for selected devices regardless of highway type), or a combination of both (i.e.,
analyze crash or maintenance records for selected devices on selected highway sections). Examples of
such analyses may include:
' FrequencylraTe and severity ofreported crashes and frequency/rate ofunreported crashes involving
various roadside features, broken down by year, highway type/functional class, and traffic volume
for each district and statewide;
' Trend analysis of fi'equency/rahe and severity of reported and unrepofted crashes involving various
roadside features.
The database can also be used to conduct comparative analysis on an ad hoc basis for selected roadside
safefy features and highway sections. Examples of ad hoc comparative type of analyses that may be
addressed with this database include:
' Comparison offrequency/rate and severity ofreported crashes and unreported crashes before and
after installation of median barriers,
' Trend analysis offrequency/rate and severity ofreporled crashes and unrepofted crashes involving
various roadside safety features for specific highway sections.
The supplemental data collection subsystem is intended to supplement the computerized database for
analyses in which the level of detail of the computerized database may not be sufficient. The supple-
mental field collection may include specific data on the selected roadwa¡ roadside, and safety feature
or manual review of hard copies of police crash reports to obtain information otherwise not available
from the computerized database, or both. Studies under the supplemental data collection subsystem
will be conducted on an ad hoc basis for selected roadside safety features, e.g., comparison of impact
performance between different guardrail types as a function of highway type, speed limit, lateral offset,
mounting height, etc.
The in-depth investigation subsystem involves thorough examination of selected crashes, including
reconstruction of the crashes to estimate impact conditions and to assess the performance of roadside
safety features. This subsystem will be used in selected studies where the highest level of detail is
deemed necessary. This subsystem requires resources typically beyond what user agencies have cur-
rently or will have in the foreseeable future. Thus, this subsystem is likely limited to ad hoc studies
conducted by outside contractors.
7.4 DISCUSSTON
V/hile there is no formal requirement for in-service performance evaluation, it is highly recommended
that some form of an in-service performance evaluation program be implemented, perhaps as part of the
safety management system. NCHRP Report 490 ( I l8) presented detailed procedures for one approach
to the conduct of in-sen¿ice performance evaluation. The conceptual framework presented above cov-
ers additional aspects and approaches for an in-service performance evaluation program. However, it
should be emphasized that it is intended as a conceptual framework and user agencies should select the
specific aspect or approach that best fits the needs and resources ofthe agency. Ideally, the in-service
performance evaluation program would include both new feature evaluation and continuous monitoring.
The new feature evaluation system would assess the impact performance and operational characteristics
of any new or extensively modified feature to make sure that the feature is performing as designed. The
continuous monitoring system would monitor the operational performance of various safety features
in case there are changes in the vehicle fleet or highway operating conditions that adversely affect the
performance of roadside safety features.
Also, in today's environment of limited manpower and increased workload, it would be a good idea to
pool resources among several states with interest in the same safety features in order to obtain larger
sample sizes and reduce the workload of individual states. It is further recommended that a national
center for in-service performance evaluation be established as a clearinghouse to disseminate the
information and to coordinate such efforts.
Commentary
A
urther discussion and elaboration are provided on certain sections in the text. Those sections for
F which commentary is given correspond to section numbers in the text preceded by the letter "A."
For example,uA3.2-1" refers to Section 3.2.1 in the text.
CHAPTER ONE
Vehicle crash tests are complex experiments that are not easily replicated because of difficulties in
controlling critical test conditions such as speecl, angle, and condition of test vehicle and the some-
times random and unstable behavior of dynamic crush and fracture mechanisms. Testing guidelines
are intended to enhance precision of these experiments while maintaining their costs within acceptable
bounds. User agencies should recognize the limitations of these tests and exercise care in interpreting
the results.
It is impractical to attempt to duplicate the innumerable site and safety feature layout conditions that
exist along the nation's highways in a limited nunrber of standardized tests. Accordingly, the aim of the
guidelines is to normalize or idealize test conditions. Hence, straight longitudinal barriers are tested,
although curved installations exist; a flat grade is recommended, even though installations are some-
times situated on sloped shoulders and behind curbs. These normalized factors have significant effect
on the petformance of many safety features and may obscure serious safety deficiencies that exist under
more typical but less ideal conditions. However, these normalized factors are thought to be second-
ary in importance when the object of a test program is to compare the results of two or more systems.
Moreover, the normalized conditions are more easily duplicated by testing agencies and help to assure
consistency from one lab to the next. Nevertheless, when the highway engìneer suspects that a system
willbe particularly sensitive to some specific site conditions such as a unique soil or roadside geometry
it is important that the feature be tested under these "more critical" conditions instead oi or in addition
to, the idealized conditions recommended herein.
These guidelines are intended for use with highway safety features that will be permanently or tempo-
rarily installed along the highway. Temporary features are generally used in work or construction zones
or other temporary locations, and their duration of use is normally relatively small. An important addi-
tional characteristic of
a work zone is the exposure of construction personnel to errant traffic. Thus, a
banier in a work zone may be required to (l) redirect errant traffic away from a roadside hazard or other
traffic and (2) to shield workers from errant vehicles. Depending on specific site conditions, the impact
severity in construction zones may equal or even exceed conditions found at typical non-construction
zone sites.
CHAPTER TWO
A2.1. GENERAL
The multiple service level (MSL) concept for highway safety features was first introduced for bridge
railings in NCHRP Report 239 (22). NCHRP Report 230 (92) also incorporated the MSL concept to
some degree. Table 3 in NCHRP Report 230, "Crash Test Conditions for Minimum Matrix," provided
testing for an MSL of 2. Table 4 of Report 230,"Typical Supplementary Crash Test Conditions," pro-
vided test conditions for MSLs of I and 3. The supplementary matrix applied primarily to longitudinal
barriers. Section 20 ofAASHTO's Standard Specifications for Highway Bridges (5) also incorporated
the MSL concept by including four different performance levels for bridge railings. The MSL concept
was formally introduced for all safety features with the publication of NCHRP Report 350 (129), which
included 6 levels of seryice or "Test Levels." This document also includes 6 test levels, largely modeled
after the test conditions recommended by NCHRP Report 350.
Unfortunately, there are no widely accepted warrants or criteria that identify roadway classifications,
traffic conditions, traffic volumes, etc., for which a safety feature meeting a given test or performance
level should be used. Given the choice, it would be preferable to first establish conditions or warrants
for which features having given capabilities would be cost-effective and thereby define appropriate
test levels. Instead, it is necessary to first establish a set of test levels with the uncertainty as to where
features developed to meet these levels have application. When warrants for multiple test level features
are developed, it is possible that some of the levels will prove to have little application and other levels
are needed.
Errant vehicles of all sizes and classes leave the travelway and strike highway safety features with a
wide range of speeds, angles, and attitudes. It should be a goal of transpofation officials to design safe-
ty features that will satisfactorily perform over as wide a range of impact conditions as can practically
be accommodated. Combinations of vehicle speed, mass, and approach angle that occur are unlimited.
However, impact conditions must be reduced to a very limited number to keep an evaluation test series
within economic and practical bounds. The approach used in formulating the recommended test con-
ditions is to evaluate the devices for cases that are believed to represent the worst practical condition.
Accordingly, there is no assurance that a safety feature will perform acceptably with other vehicle types
presently in service or those vehicle types that may come into use during the normal service life of the
device. This "worst practical condition" has been defined as the combination of the 5th percentile light-
est and heaviest passenger vehicles striking a safety feature at the 85th percentile highest speed and 85th
percentile highest angle. This combination of nearly worst case weight, speed, and angle is believed to
produce an extremely rare impact event. Nevertheless, these impacts do occur and have been designated
as representative of the most severe impact conditions that can be practically accommodated. This defi-
nition of the worst practical impact condition was originally implemented for large passenger vehicles
with the first set of evaluation guidelines presented in Highway Research Boatd Circular 452 (153).
The precedent established with the first set ofguidelines for full-scale crash testing has been extended
through Transportation Research Circular 191 (154), and NCHRP Reports 230 (92) and 350 (129).
The only significant revision to passenger vehicle testing philosophy incorporated with the cunent
guidelines is application of the 85th percentile inpact angle to the small passenger vehicle. All available
accident data shows that impact angles for small cars are at least as high as those associated with large
passenger vehicles and SUVs. Further, accident investigations appear to indicate that higher impact
angles significantly increase accident severities for all sizes ofpassenger vehicles. Therefore, the rec-
ommended impact angle for tests involving small car redirection matches that for the light truck test
vehicle.
Unfortunately, there is limited crash data available with which to quantiff heavy truck crash severities.
Heavy truck impact conditions recommended in NCHRP Report 350 were primarily based on Section
20 ofAASHTO's Standard Specifications for Highway Brídges (5). However, the increased severity of
the limiting TL-3 test now exceeds the severity of the limitin gTL-4 test from NCHRP Report 350 by
approximately l8 percent when measured in tems of Impact Severity, IS, which is defined as follows:
19= y,ttt(Vsin0)'
Where
IS impact severity, kip-ft (kJ)
M mass of impacting vehicle, kip-sec2/ft (kg)
l/ velocity of impacting vehicle, fusec (m/sec)
0 impact angle (deg)
It is logical to expect that TL-4 barriers should be capable of withstanding higher impact severity
levels than TL-3 baniers. Although there is insufficient data available to identifli the full distribution of
impact angles and speeds for heavy truck impacts on roadside barriers, the data that is available clearly
indicates impact severities can be as high as or higher than what has been proposed for TL-4 and TL-5
in NCHRP Report 350 (129) (84). The TL-4 impact conditions incorporated into NCHRP Report 350
originated with Section 20 ofAASHTO's Standard Specifications for Highway Bridges (5) and were se-
lected as a replacement for several bus tests included in NCHRP Report 230 (92). These bus tests were
replaced due to the inflammatory nature of some test videos showing surrogate bus occupants being
ejected from the vehicle's windows, even though the vehicle was successfully contained and redirected.
ThefourbustestshadlSvaluesrangingfromalowofl12kip-ft(l52kJ)toahigh of323 kip-ft(438
kJ). Unfortunately, when the bus tests were replaced with single-unit trucks, the IS value for TL-4 was
reduced to 98 kip-ft (132 kJ), well below even the least severe bus test included in NCHRP Report 230.
These reduced impact conditions were originally selected because the single-unit truck was deemed to
be less stable and would, therefore, place additional demand on barrier performance in order to prevent
rollover. However, after the TL-4 impact conditions were selected and approved, the evaluation criteria
for all heavy truck tests were revised to allow the impacting vehicle to roll over on the traffic side of the
railing. In light of the increase in the severity of TL-3 testing and the history of the TL-4 impact con-
ditions, this test was revised to significantly increase the impact severity so that there is some increase
in capacity going from TL-3 to TL-4 baniers by raising the mass and impact speed for the TL-4 test to
22,046Ib (10,000 kg) and 55.9 mph (90 km/h), respectively.
Note that cable barriers have traditionally been tested without any cable splices in the impact region.
However, cable splices must be used in long runs of cable barrier and to repair cables damaged during
a crash. Hence, any splice that is expected to be used in the field must be incorporated into the critical
impact region during crash testing.
Note that guardrails and median bariers may occasionally be flared relative to the travelway such that
the effective impact angle is increased. This document does not recommend that every barrier system
be tested under the highest possible flare rate condition. However, decisions regarding appropriate
banier flare configurations must be based upon a careful evaluation of the consequences of increasing
or decreasing the flare rate. Increasing a barrier flare rate is believed to increase the severity ofbarrier
crashes. However, increasing flare rates also reduces the number of barrier collisions and total barrier
costs by reducing the barrier length. Optimal barrier ffare rates should be chosen based upon a cost-ef-
fectiveness analysis that provides the lowest total societal cost, including accident costs and barrier
construction costs. Optimal flare rates chosen in this manner may produce barrier or terminal installa-
tions that cannot meet the full-scale crash testing requirements described herein under the conditions in
which they are installed. The guidelines contained in Chapter 2 are intended to assure a minimum level
of impact performance for barriers installed parallel to the travelway, not for every possible barrier flare
configuration.
A transition between nvo longitudinal barriers with differing lateral stiffness, such as a rigid concrete
bridge rail and a W-beam guardrail, can pose a difficult design problem. The most common method
for constructing such a transition is to build an intermediate barrier section with stiffness somewhere
between the approach guardrail and bridge rail. Testing has shown that vehicles impacting upstream of
the intermediate stiffness section can pocket behind the stiffer barrier and either roll over or rupture the
rail element ( I 10). In this situation, it is important to conduct transition testing at both critical locations,
i.e., the transition between the intermediate stiffness section and the bridge rail as well as the transition
between the approach guardrail and the intermediate stiffness system. Note that small car testing has not
indicated a significant problem for either impact location. Thus, when approach barriers have geom-
etries very similar to previously tested systems, it may not be necessary to conduct small car tests at
either impact location.
While it is preferable that the test vehicle remain upright after each test described herein, èxceptions are
made for all heavy vehicle tests. A one-quarter roll is permitted in the heavy vehicle tests because the
primary goal in these tests is to demonstrate that the longitudinal barrier being evaluated can contain
and redirect the vehicle. Further, analysis of truck accident data does not show the same strong link
between vehicle rollover and injury and fatality that is found with passenger vehicle data. Note that
even though ovefturn is permitted for all heavy vehicle tests, evaluation criterion D of Table 5-lB must
be satisfied, i.e., the overturn must not result in deformations of the occupant compartment that could
cause serious injuries.
In recognition that teminals and redirective crash cushions perform the same function and for rigid barri-
ers the terms have become interchangeable, these devices are subjected to the same set of full-scale crash
tests. Non-redirective crash cushions are not designed to perform as a longitudinal banier when struck
downstream from the nose ofthe system and therefore are subjected to a reduced set oftests. Note that
NCHRP Report 350 also identified different test conditions for "gating" and "non-gating" terminals and
crash cushions. Although the definition of "gating" and "non-gating" has been retained from the prior
publication, the two types of systems are now subjected to the same set of full-scale crash tests, Tests
30 through 38. In order for a terminal or crash cushion to be classified as non-gating, it must capture the
impacting vehicle during Tests 32 and 33. Note that the impact angle is given as ranging from 5 to 15
degrees for these tests. Each device must be tested at the impact angle that will maximize the risk of test
failure. Devices expected to qualif, under the gating category should be tested near the lower end of this
range and devices designed to function as non-gating should be tested near the upper end ofthe range.
Note that if a non-gating device fails to capture the impacting vehicle during Tests 32 or 33, it must be
retested at the lower impact angle in order to qualiff as a gating system. Similarly, systems that capture
impacting vehicles when tested near the lower end of this range must be retested at the higher impact
angle before qualifying as a non-gating design.
Tests 38 and 45 have been added to ensure that staged attenuation systems perform adequately for all
sizes of passenger vehicles. Note that the purpose of this test is to identif, occupant ridedown accel-
erations for mid-sized vehicles. Since this behavior is primarily related to the mass of the impacting
vehicle and the critical vehicle weight may vary from system to system, the acceptable mass range was
increased to +220 lb ( I 00 kg). Further, removal of up to 441 lb (200 kg) of test vehicle components to
reduce mass and the addition of up to 441lb (200 kg) of ballast to meet the mass guidelines is accept-
able. This test can be waived if the analysis shown in Appendix G indicates that Test 38 or 45 will meet
occupant risk critelia.
Note that the recommended test matrices cannot and should not be expected to be an all-inclusive set of
standardized procedures. When appropriate, testing agencies and/or user agencies should devise other
critical test conditions consistent with the range of expected field impact conditions.
It is important to assess both the risk to occupants of impacting vehicles when a TMA is mounted to a
heavy truck and the anticipated roll-ahead distance when mounted to a light truck. Therefore, Tests 50,
5 l, and 52 are lo be conducted with the maximum allowable support truck weight while Test 53 is to
be conducted with the minimum allowable support truck weight. Support truck roll-ahead distances
should be carefully documented for all four tests. It is noted that roll-ahead distances can be accurately
estimated from the "conservation of momentum" principle of mechanics based upon an estimate of the
frictional resistance of the suppoft truck to forward movement (70). Results of recommended full-scale
crash tests can be used to assess effective friction values. Using these techniques, testing agencies
should evaluate expected roll-ahead distances over the entire range of support truck weights ranging
from the minimum to the maximum.
Risks to the support truck occupants increase as the truck weight is reduced. Howeve¡ as long as the
support truck weighs significantly more than the impacting vehicle, truck accelerations will be signifi-
cantly less than measured on the test vehicle. For example, if the support truck weighs 20 percent more
than a 2270P test vehicle and the attenuator mass is 992 lb (450 kg), the support truck would experience
approximately 30 percent lower accelerations than the impacting 2270P test vehicle. Further, during
an impact, the support truck is accelerated forward, pushing its occupants backward into their seat.
Rearward occupant movement is generally less dangerous than forward movement, primarily because
well supported by the seat and head-rest system. Therefore, provided the support truck
the body is
weighs significantly more than the impacting vehicle and proper head rests and seat-belt systems are
utilized, the risks to suppoft truck drivers should be significantly less than risks to the occupants of the
impacting vehicle.
Test 54 is implemented to assess if TMAs with staged energy absorption would perform satisfactorily
during impacts with mid-sized vehicles. Clearly all TMAs should be designed to accommodate the
full range of passenger vehicles, including mid-sized cars. However, there is concern that the staging
of some TMAs may be too finely tuned to accommodate only the small passenger car and the heavy
pickup tluck that they may not perform satisfactorily for a mid-sized vehicle, particularly the ridedown
acceleration. On the other hand, thele is also concern over the possibility that many existing TMAs
might fail this test. This would require re-design of these TMAs, which may lead to increased costs and
operational and durability problems. Thus, this test is considered optional. Manufacturers and users are
strongly encouraged to develop and implement TMA designs that can be both operationally efficient
and exhibit proper impact performance during mid-sized vehicle testing.
Tests 50 and 5l are recommended for use in evaluating the impact performance of portable work-zone
traffic control trailers. These same tests are also recommended for evaluating the safety risks associated
with any other Iarge work-zone devices that begin to be recognized as a potential safety hazard.
The energy or force required to fracture a breakaway device or support structure, in general, may be
sensitive to its orientation with respect to direction of impact or the impact angle. For example, tests
have indicated a breakaway transformer base breaks more readily when struck on a colner than on a flat
side. Because errant vehicles may approach a support structure, work-zone traffic control device, or a
breakaway utility pole at various angles, it is recommended that the device be tested assuming the most
severe direction of vehicle approach consistent with expected traffic conditions or at the critical impact
angle (CIA) discussed in Section 2.2.4. For instance, the transformer base should be oriented so the ve-
hicle strikes a flat side. Moreove¡ because the energy required to fracture a device can be increased due
to buckling of the support at the point of contact with the vehicle, the handhold in the luminaire shaft
should be positioned during a test so that the probability of local collapse of the shaft is maximized.
Energy-absorbing, yielding support structures have been developed as potential replacements for
conventional breakaway systems. These devices are designed to decelerate the vehicle to a controlled
and safe stop, similar to a crash cushion, rather than permitting the vehicle to break through and con-
tinue with minimal reduction in speed. Rigid, non-breakaway supports are often used in wban areas
where encroachment of the vehicle beyond the pole could endanger pedestrians or other innocent
bystanders. While this practice may offer protection for the innocent bystander, it also increases risks
to errant motorists. The yielding pole may have application in these areas, and/or areas where trees or
other hazards existjust beyond the pole line that could endanger occupants ofthe encroaching vehicle.
However, since such a design would not pass occupant risk criteria recommended for breakaway sup-
port structures, it should be evaluated according to criteria recommended for a crash cushion. Although
recommendation on the use of such features is beyond the purview of this document, the appropriate
applications for this type of device should be identified based upon a benefilcost analysis of the various
altematives.
Work-zone traffic control devices are subjected to similar testing as structural supports. However, due
to their light mass, most of these systems will meet all vehicle velocity change criteria. In these cases,
the primary concerns are related to windshield damage and the potential for structural components to ìn-
trude into the occupant compartment. In fact, instrumentation of test vehicles is not required for testing
of work-zone devices, provided the system is free-standing and its total weight is less than 220Ib (100
kg). Note that many testing agencies impact two work-zone traffic control devices in a single run. The
devices are placed to impact opposite quarter points of the front of the vehicle. Device spacing is select-
ed such that the fir'st device is usually completely disengaged from the test vehicle before it strikes the
second device. In some cases, the first device does not disengage or it produces sufficient damage that
it is impossible to determine the extent of windshield damage for the second device. In these situations,
the second device should be retested.
Breakaway utility poles are tested and evaluated somewhat differently from other breakaway support
structures. A higher occupant impact velocity is permitted in a utility pole test because the substantial
mass of a typical utility pole produces a much higher velocity change than allowed for a conventional
breakaway device. Since a higher occupant impact velocity is permitted, the impact speed for the "low
speed"test'¡/assetat3I mph(50km/h), or47ftls(13.9nr/s).Notethatforanimpactspeedof l9mph
(30.0 km/h) or 27 ft/sec (8.3 m/s), as used for other breakaway support structures, the vehicle could
come to an abrupt stop and still pass the 39 ft/s ( I 2.0 m/s) maximum occupant impact velocity criteri-
on. Recommended tests and assessment criteria notwithstanding, it should be a goal of the designer to
develop breakaway utility pole systems that minimize vehicular velocity change and, when possible,
limiting occupant irnpact velocities should equal those for other support structures. Replacement of
solid timber poles with lighter sttuctures, if feasible, could reduce or eliminate problems associated with
the relatively large mass of timber poles. Utility poles could then be expected to meet the same safety
standards as other support structures.
Other objects that are placed near high-speed roadways, such as fire hydrants, electrical transformers,
etc., should be subjected to the same crash test matrix as a breakaway support structure. Most of these
elements can be made to meet the criteria associated with breakaway supports. If a device is incapable
of meeting these criteria simply due to its mass, evaluation criteria presented for breakaway utility poles
should be utilized.
Longitudinal channelizers do not function as a longitudinal barrier, and testing agencies should clearly
state this fact in all test repofts. Because these systems are designed to allow an impacting vehicle to
penetrate through the line of elements, the critical impact angle appropriate for each system is difficult
to determine. I1 is anticipated that most longitudinal channelizers will function as a baruier when sûuck
at extremely low approach angles. Further, these systems are expected to allow rapid gating during
high-angle impacts. The impact angle most prone to cause a vehicle to roll over is generally believed to
be somewhere in between, where the risk of a vehicle contacting and possibly overriding the end of one
of the system's segments is maximized. Another potential failure mode for longitudinal channelizers
involves an impacting vehicle rotating as it penetrates into the system and the side of the vehicle con-
tacting the end of one of the segments. In this situation, occupant compartment defomation can become
excessive. Designers and testing agencies must attempt to identifr the expected system behavior and
the appropriate critical impact angle to maximize the risk of undesirable performance through computer
simulation or evaluation of tests of similar systems.
Longitudinal baniers generally fail due to structural inadequacies that allow snagging or pocketing on
stiflpoints in the barrier systems or rupture of one of the "weak points" in the baniet'system, such as
at a splice. Thus, most barrier systems have one or more critical locations where failure is likely to take
place, whether through wheel snag or rupture of a barrier element. The potential for every failure is
affected to some extent by the selected impact point.
NCHRP Report 350 (129) produced procedures for identi$ring these critical intpact points on any type
of longitudinal barrier. These procedures have been updated for the test vehicles recommended herein
and are presented in Chapter 2. Nevertheless, as recommended previously under NCHRP Repon 350,
Barrier VII or another simulation program should be used, whenever possible, to identisr CIPs for lon-
gitudinal barier tests. Techniques for utilizing simulation programs to identif, critical impact points are
summarized in References 82 and 120.
Connection loading is another important test parameter that is affected by impact location. Fortunately,
impact locations that maximize wheel snagging or pocketing at one point in the barrier will also max-
imize connection loads near that same point in the barrier. Therefore, whenever rail splices or other
critical connections fall at or near (withìn 5 ft (1.5 m)) a snag point such as a barrier post, the impact
location can be chosen to maximize both the potential for snagging and connection loadings. Since
banier loadings are generally higher upstream ofthe snag point, critical connections should be placed
at orjust upstream ofthe snag point, provided the connection locations are consistent with in-service
locations. Rail tensile loads are maximized all along the length of the first span upstream from the snag
point. Thus, the potential for rail splice tensile failure can generally be maximizedby choosing the CIP
for snagging if the connection is placed at the snag point or anywhere within the first span upstream
from the snag point.
However, when a ban'ier connection is not located within approximately 5 ft (1.5 m) of a snag point,
bending moment and shear in the connection will not be maximized by an impact location chosen to
maximize snagging. When barrier connections are not within 5 ft (1.5 m) of a snag point and when wheel
snagging or pocketing as well as connection loading in bending and/or shear are significant concerns,
the designer may consider conducting two tests with different impact locations. Barrier VII or a similar
simulation prograrn is recommended to investigate the need for two tests and to select CIPs.
It has been found that the CIP with regard to snagging is sensitive primarily to dynamic yield force of
banier posts, plastic moment of rail elements, and post spacing (137). Post yield forces and spacing
were then combined into a single parameter, Fp,by dividing the dynamic post yield forces by the post
spacing. CIP selection curves were then developed as a function of plastic moment of rail elements, Mp,
and post yield force per unit length of barrieç $. Refelence 8 I contains a more detailed description of
the development of the CIP selection curves shown in Figures 2-6 through2-17.
The plastic moment of a barier rail element is merely the product of the beam's plastic section modulus
and the material yield stress. Procedures for calculating plastic section modulus are presented in many
textbooks on plastic design of steel structures (I 34). The plastic section modulus can be estimated with a
reasonable degree of accuracy by multiplying the elastic section modulus by a form factor. Form factors
for common beam shapes vary from a low of about I .l to a maximum of 2.0. As the fraction of a be4m's
cross section located nearthe neutral sul'face increases, the form factor ofthe cross section increases.
'Wide flange beams have very little material near the neutral surface and, as a result, generally have form
factors less than I .18 with an average near I . 14. Form factors for square box beams range from a low
of 1.13 for a very thin-walled tube to a high of 1.5 for a solid rectangular rod. Forrn factors and plastic
moments for some common barrier rail elements are shown in Table A-1.
Elastic Sec-
Plastic Sec- Yield Plastlc Mo-
tion Modu- Form
Raila tion Modulus, Strength, ment, kip*ft
lus, ln.3 Factor
(cm3)
in.3 (cm3) ksl(MPa) (kNrm)
12-gauge W-beam "t.37 (22.4s) 1.41 1.93 (35.63) 50 (345) 8.0 (10.e)
1O-gauge W-beam 1.76 (28.84) 1.41 2.48 (40.64) 50 (345) 10.3 (14.0)
12-gauge Thrie-Beam 2.1e (35.8e) 1.4 3.07 (50.31) 50 (34s) 12.8 (17.3)
10-gauge Thrie-Beam 2.80 (45.88) 1.4 3.e2 (64.24) 50 (345) 16.3 (22.1\
TS 6 by 6 by 3/1.
7.s3 (129.5) 1.17 e.24 (151.4) 46 (317) 35.4 (48.0)
Box Beam
TS6by6by3/.
13.s0 (227.8) 1.21 16.80 (275.3) 46 (317) 64.4 (87.3)
Box Beam
TSBby6bylla
1s.00 (245.8) 1.2 18.00 (2e5.0) 46 (317) 6e.0 (s3.5)
Box Beam
a Rail sizes are in English units.
Barriers with multiple rail elements complicate the selection of an appropriate plastic moment for the
barrier. When this type of barrier deflects during an impact, the upper rail deflection is much higher
than that of lower rail elements. A simple energy analysis indicates that the total energy absorbed by
each railelement is roughly proportional to the mounting height of the element. E,quation 2-4 was then
developed to estimate an equivalent plastic mouìent for rnultiple rail systems. A limited sensitivity study
using Bamier VII revealed that the CIP determined by use of Equation 2-4 accurately estimates the
CIP for most multiple rail barrier systems. This study indicated that the procedure was somewhat less
accurate for barriers that have relatively stiff rail elements well above the impacting vehicle. For this sit-
uation, barier posts will yield above the impacting vehicle and the upper rails will not deflect as much
as the lower rails. Although the CIP selection procedures do give reasonable estimates of critical impact
Iocations for most of these barriers, a simulation program should be used when possible to verifli the
findings.
When barrier posts are rigidly anchored, yield forces are controlled by the material properties of the
post. A dynamic magnification factor is normally applied to the plastic section modulus of metal posts
to estimate the dynamic yield force for a post as given in Eq. A3- I .
(Eq. A3-1)
"='(+)
Where
Fy dynamic post yield force for a rigid anchor;
D dynamic magnifi cation factor;
cy post yield stress;
zp post plastic section modulus; and
H, height ofhighest rail above base ofpost.
The accuracy of Eq. A3-l can be demonstrated by comparing a measured value of $ for a rigidly an-
chored W6 by 9 (Wl52 by 13.4) steel post with the calculated value. A dynamic magnification factor of
1.5 is typically used for steel posts and a W6 by 9 (V/152 by l3.a) beam has a plastic section modulus
of 6.3 in.3 (103 crn3) and a yield stress of 36 ksi (248 MPa). For a 1.7-ft (0.53-m) mounting height, Eq.
A3-l gives an F, of 16.2 kip-force (71.9 kN) compared to a measured value of 16.8 kip-force (74.7 kN)
from Development of a Cost-Effectivness Model for Guardrail Selection (29).
Wood materials exhibit a brittle failure mechanism, and therefore, the plastic section modulus in
Eq. A3-l is replaced by the modulus of rupture. Reference 29 reported that pendulum tests of 6- by 8-in.
(15.2- by 20.3-cm) Douglas Fir posts have an average failure force of 16.2 kip-force (72.1 kN) when
mounted in a rigid support. Southern Douglas Fir has an average modulus of rupture of 6800 psi (46.8
MPa) (84). Using a dynamic rnagnification factor of I .0, Eq. A3- I predicts failure forces of 20.7 kip-
force (91.9 kN) and 16.6 kip-force (74.0 kN) for rough cut and finished posts with a nominal 6- by 8-in.
(15.2-by 20.3-cm) size. Although it is unclear whether posts used in the pendulum tests were rough cut
or finished size, the test results do indicate that the dynamic magnification factor from Eq. A3- I should
be no more than 1.0 for wood materials. Table A-2 shows the modulus of rupture for some common
wood post materials.
Dynamic yield forces for posts embedded in soil are generally more difficult to estimate. Soil yield
forces are usually measured through pendulum or instrumented cart testing at speeds near 20 mph
(32 km/h). A number of guardrail posts have been tested for various soil embedment conditions (24,
29,42,44,149,150). Dynamic yield forces for common guardrail posts embedded in strong soils are
shown in Table A-3. The testing programs referenced above have shown that post yield forces can be
approximated as a function of the square of the embedment depth. Thus, yield forces from Table A-3
can be extrapolated for other embedment depths by multiplying the forces shown by the square of the
ratio of the two embedment depths as given in Eq. A3-2.
'Where:
Some pendulum tests have been conducted in soft soils and are reported in Development of a Cost-
Effectivness Modelfor Guardrail Selection (29). Analytical procedures for estimating the yield forces of
other post sizes and soil conditions are discussed inl
Study ofthe Soil-Structure Interaction Behavior
of Highway Guardrail Posts (42).
'O 2016 by [he Anrerican Association of Srate Highway an(l Transportat.ion Ollicials
All rights reseryed Duplication is a violation ofapplicable lau,
142 | Manual forAssessingSafety Hardware
for the truck to roll over the top of the barier, the impact point should be selected to maximize lateral
loading at mid-span where the top barrier rail would be expected to deflect downward and increase
rollover potential. Note that since heavy trucks spread impact loads over a larger area, a single test can
usually be devised to apply near maximum loadings on all critical connections and adequately investi-
well as rollover. Table 2-7 presents findings from full-scale testing
gate the potential for post failure as
of rigid barriers and instrumented walls (13,27, 68).
CHAPTER THREE
The attitude of the vehicle at the point of impact can have a significant effect on the performance of the
feature being tested. For example, a nose-diving vehicle can result in underride of the barrier while an
upwardly bouncing vehicle can promote override of the barrier. Thus, it is critical to have a relative-
ly flat surface in the approach area so that the vehicle attitude is stabilized at impact. The runout area
for the post-impact vehicle should be long and wide enough for the vehicle trajectory to be properly
assessed.
A3.3 SOIL
The impact performance of many longitudinal bariers and breakaway or yielding support structures
depends on the strength and fixity of the soil foundation. Thus, soil foundation is an integral part of
such systems. For example, displacement and/or rotation of a breakaway device footing during collision
can adversely affect the fracture mechanism. Insufficient soil sr.rpport can lead to excessive guardrail
post movements and guardrail lateral deflection during vehicle collision and result in a lower system
capacity to contain and redirect en'ant vehicles. Insufficient soil strength can also be a critical and limit-
ing factor for the anchoring function of a longitudinal barrier terminal. On the other hand, an unusually
fiml soil can increase the lateral stiffness of a longitudinal barrier and subject occupants of a colliding
vehicle to undue risk.
Soil conditions along the highway are variable and may be affected by many factors. Soil type could
range from soft sand materials to hard rock materials; moreover, the soil type may vary considerably
within a locale well as from region to region. In addition to soil type, soil strength may also be a
as
function of the season as it can be significantly affected by moisture content and whether the soil is
frozen. Other significant factors may include compaction, density, and consolidation of the soil. The
testing agency should be aware of the importance of soil strength and select the most appropriate soil
type consistent with potential application of the feature.
Recommended soils are well-graded materials that should be readily available to most testing agencies.
The standard soil of Section 3.3.1 is a selected AASHTO material that compacts to form a relatively
strong foundation. The weak soil of Section 3.3.2 is a typical AASHTO fine aggregate. These soils are
essentially the same as the "strong" and "\Meak" soils of NCHRP Report 350 (129).
The following general guidance in soil selection is offered to the user agency and the testing agency:
A3.3.1STANDARD SOIL
Unless the test article is linlited to areas of weak soils, the standard soil should be used with any feature
whose impact performance is sensitive to soil-foundation or soil-structure interaction. A large percent-
age of previous testing has been performed in similar soil and a historical tie is needed. Although it
is probably stronger than the average condition found along the roadside, it is still representative ofa
considerable amount of existing installations.
43.3.2 SOILSTRENGTH
The soil specifications in NCHRP Report 350 (129) are intended to provide consistency to soil strength
and test results of soil-based installations among the testing agencies. Howeve¡ even though the soil
types available to the various testing agencies locally all meet the material specifications, they vary wide-
ly in their characteristics, including soil strength. The installation procedures speci$ the width and depth
of the fill material for embedment of posts to ensure that the test results are not affected by the character-
istics of the native soil. However, these recommendations were seldom followed since they required an
inordinate amount of fill materials for a typical test installation. There are also other factors affecting soil
strength that are not clearly covered by the specifications, such as moisture content. These factors led to
concems over the repeatability of test results among the test agencies.
Each testing agency is asked to conduct an initial set ofcalibration tests to establish the baseline val-
ues for the in-situ soil strength testing. The purposes ofthese calibration tests are to define the range of
soil strength among the testing agencies and to establish the minimum acceptable value. However, it is
recommended that the individual testing agencies also use this opportunity to establish the relationships
between soil strength and the various influencing factors, such as native and fill rnaterial soil type, width
and depth of fill material, gradation, compaction, soil density, and moisture content. The information
would allow the testing agencies to judge if sufficient soil strength is available for testing without actual
conduct ofthe in-situ tests.
The in-situ test should be conducted immediately prior to each crash test to veri$r that the soil strength
exceeds the minimum acceptable value. If the test results indicate insufficient soil strength, the crash
test should be postponed until the soil conditions improve. For each test installation, a minimum of two
posts should be installed for in-situ testing in case the first in-situ test results are not acceptable. Unless
more than two posts are instailed, there will only be two oppoftunities for the in-situ soil strength tests.
Thus, it is important for the testing agencies to understand the relationships between soil strength and
the various influencing factors so that the in-situ test is conducted only when the soil conditions are
favorable.
In addition to soil selection, tle footing or foundation used in a test ofa breakaway support structure
should be designed for the minimum wind conditions permitted, thus yielding a minimum footing mass
and size; a larger footing will yield a greater fixity for a breakaway device and is, hence, less critical.
The standard soil of Section 2.2.1 .l is especially sensitive to moisture content. The testing agency
should sample and test the soil to ensure moisture content is within recommended limits given in the
speciflcation at the time of the test.
Failure or adverse performance of a highway safety feature during crash testing can often be attribut-
ed to seemingly insignificant design or construction details, something as innocuous as a substandard
washer. For this reason, it is most important to assure that the test article has been properly assembled
and erected and that critical materials have the specified design propefties. Details of most concern are
those that are highly stressed (such as welded and bolted connections, anchor cables, cable connections,
and concrete footings) or those that must fracture or tear away during impact (such as breakaway sign
bases or weakened barrier posts). Compressive tests of concrete cylinders, proof tests of cable assem-
blies, and tests of physical and chemical properties of matedals, in general, should be performed on a
random sample of the test article elements or obtained from the supplier of the material. Even though
well-defined material specifications and appropriate fracture modes may not be fully developed, the
properties of all components and materials used in the test article should be documented in detail in the
test report or at least be traceable should any questions arise.
simulate a long barrier. Recommendations in the minimum installation lengths are provided in the docu-
ment and should be followed unless there are extenuating circumstances. Barrier lengths may also need
to be extended to ensure that heavy trucks are fully contained priorto reaching the end ofthe barrier.
Also to be considered is the possible need to extend the barrier installation to observe a second collision
between vehicle and barrier.
CHAPTER FOUR
Two test vehicles have traditionally been used to represent the entire fleet ofpassenger vehicles op-
erating on the roadside. The philosophy behind this approach has been that, ifa roadside feature can
safely accommodate both ends of the vehicle size spectrum, it should provide good performance for
almost all vehicle sizes in between. For most classes of safety features, the same approach has been
maintained in this document. Over the last two decades, vehicle masses have increased dramatically as
the popularity of large sport utility vehicles (SUVs) has grown. It is possible that the size and nature of
the vehicle fleet may change significantly over the next 10 to I 5 years, with vehicle weights declining
to more closely match histol'ical averages. In such a circumstance, it would be desirable to revise test
vehicle specifications to more accurately reflect the new vehicle fleet. A brief summary of tlris process is
presented below.
A mid-sized test vehicle, designated 15004, is included in the test matrix in order to evaluate cable
barriet'systems and energy-absorbing terminals, crash cushions, and truck-mounted attenuators. This
vehicle will be used to determine if staging in an attenuation system is designed properly to safely
accommodate high-speed, head-on impacts with mid-sized vehicles. In this situation, the mass of the
mid-sized vehicle will carry it beyond the point where the I l00C vehicle is brought to a stop and likely
enter into the high-energy dissipation ranges of an attenuator where deceleration forces may become
excessive for mid-sized cars. Hence, the primary concern is that this test will cause excessive ridedown
accelerations. Because activation of attenuation systems is primarily related to vehicle mass and the
test is a head-on impact, where spin-out and rollover are not a factor, total mass is the only important
vehicle parameter for the 15004. The 3,3 l4-lb ( 1500-kg) vehicle mass was chosen after evaluating the
potential for excessive occupant ridedown acceleration in recently tested energy-absorbing terminals
and crash cushions. This analysis showed that test vehicles weighing between 2,872 and 3,755 lb (1300
and 1700 kg) would be most likely to cause excessive ride-down accelerations.
Note that safety haldware designers are expected to consider the full range of vehicle sizes when
designing attenuation devices. The mid-sized vehicle test is merely added as a minimal check to verif,
that designers are achieving this goal. The test does not indicate that designers do not need to consider
other vehicle weights between the small car and light truck test vehicles. Attenuation systems should be
designed to accommodate all vehicle masses between 2,425lb and 5,004 lb (1100 kg and 2210kg).
NCHRP Report 350 included an exception that permitted the 820C and 2000P vehicles to be more than
six model years old at the time of the test. The exception was predicated upon the testing agency being
able to demonstrate that:
"...key properties of the test vehicle are essentially the same as those of a vehicle meeting
all of the recommended requirements. Key propefties include those given in Table 2-1 plus
unspecified properties that may change with succeeding model years such as dynamic force-
deformation properties of the bumper and frontal structure of the vehicle and vehicular profile
as defined by bumper height, hood height, hood sweep, windshield srileep, and height of the
windshield."
Based upon this exception, many testing agencies have utilized older test vehicles. Unfortunately, there
has been little effort to veriff that the force-deformation characteristics are similar to vehicles within
the six model year limit. A detailed analysis, including full-scale validation testing, would be required
to properly ascertain similarity in vehicle force-deformation characteristics. Further, the "unspecified
properties" that are listed above do not encompass all of the important vehicle characteristics that can
affect the outcome of a test. For example, suspension characteristics and failure modes have been shown
to be important to many banier and transition tests. With the increased frequency of design changes in
the light truck market, it is ímpractical to even attempt to asceftain that an older vehicle is structurally
similar to one within the recommended age limit.
Therefore, the exception for using older light trucks and automobiles in full-scale crash testing was
revised. The current exception now allows test vehicles to be within 6 model years of the date that the
research and development project was undertaken. This change was adopted in recognition that some
research and development efforts experience delays that would potentially require a testing agency to
sell a test vehicle and repurchase a newer one. Under the current guidelines, a testing agency can pur-
chase six year old test vehicles without fear that the project may extend into the next calendar year.
The high cost of heavy trucks precludes implementing the six-year limit for these vehicles.
Nevertheless, heavy truck test vehicles should be structurally sound and representative of widely
used designs. rühenever possible, it is recommended that heavy trucks not be more than 12 model
years old.
Test vehicle design and overall condition at the time of testing can have a major influence on the
impact performance of a feature. Among the more important parameters are vehicle bumper height,
configuration, and stiffness; vehicle mass distribution; suspension system; and vehicle structure. In
order to assure proper safety feature performance and consistency in test results from one test to the
next, test vehicles should correspond closely to the recommended vehicle properties. Test vehicles
should be in sound structural condition without major sheet metal damage. Further, tires used during
full-scale crash tests should be in good condition and match manufacturer specifications shown on the
vehicle identifi cation plate.
Use of a vehicle for more than one crash test without repairs should be avoided because vehicle dam-
age may affect performance in a subsequent test. This is particularly important in evaluating safety
features such as a breakaway suppoft where vehicle crush can significantly affect fracture mechanisms.
The three heavy test vehicles, 100003,36000V and 360001 were selected for evaluation ofthe
capacity of longitudinal barriers where higher levels of containment are necessary such as on
high-volume bridges, overpasses, and in medians of high-volume freeways. In these situations, pen-
etration or overiding of the barrier would produce a high risk of driver fatality and could pose a risk
to other traffic below. Full-scale crash testing indicates that heights of approximately 32 in. (81 cm),
42 in. (107 cm), and 80 in. (203 cm) will be required for rigid barriers for the 100005, 36000V, and
36000T vehicles, respectively, when ballasted as recommended.
Testing and user agencies should be aware of potential problems that may occur with a test using the
36000V test vehicle. In pafticular, the undercarrìage attachment of the trailer tandems to the trailer
frame may not be of sufficient strength to provide necessary restraint during the specified test. This
problem is believed to be peculiar to sliding undercarriage or sliding axle designs. In at Ieast one
test, the attachment (which was the sliding undercarriage type) failed due to an inability to transfer
lateral impact loads, and the trailer went over the ban'ier. In a similar test with a fixed undercarriage
attachment, no such failure occurred and the trailer did not go over the barrier. A sliding attachment
is recommended for the test trailer since it is widely used in the industry. While it is desirable to test
with widely used vehicles and equipment, the primary purpose of the test is to demonstrate structural
adequacy of the barier, not the trailer. A barrier capable of containing a trailer with a sliding axle may
have to be considerably taller than one capable of containing a trailer with a fixed axle. Neveftheless,
public safefy requires effective containment of vehicles on the road. If testing reveals this type of trailer
design will cause significant increases in the cost of effective barrier systems, supporl should be sought
from appropriate officials and agencies to develop improved trailer designs to eliminate this problem.
occur when the impacting vehicle mass was between 2,872To3,7551b (1300 and 1700 kg). The mid-
sized vehicle mass was set to 3,314 lb (1500 kg) because it fell near the center of what was believed to
be the critical range and it provided some consistency with European safety hardware evaluation pro-
cedures. In recognition of the wide range of vehicle mass that produce near critical occupant ridedown
accelerations, a largervariation in vehicle mass, a221 lb (100 kg) was deemed to be appropriate.
A4.2.1.2 Bullasl
Ballast for test vehicles that is free to shift or break loose during impact may be totally ineffective or
only partially effective in loading of the feature because it tends to move independently of the vehicle.
Unless specifically designed to evaluate effects of cargo shifting, tests with the 100005 and 36000V
vehicles are to be conducted with firmly secured ballast. The tie-down system should be capable of
resisting a lateral load equal to approximately ten times the weight of the ballast.
It must be noted, however, that test experience has shown that it is quite difficult to design a ballast
tie-down system for a van truck or trailer with sufficient strength to resist typical impact loads for two
reasons: (l) the absence of lateral stiffness in the walls of the van and (2) the height the ballast must be
placed above the floor of the van to achieve the recommended center of mass of the ballast. For reasons
of economy and convenience, sandbags on pallets are commonly used as ballast in tests of van trucks or
van trailers. While this achieves the required mass and center-of-mass height, it is difficult to secure this
type of ballast and it creates a concentrated lateral load at some height above the floor of the van during
impact. It would be preferable to use a ballast material with a density as low as possible so that the
ballast would be uniformly distributed along the length, width, and height of the van, thus minimizing
the need and structural requirements of the tie-down system. Bales of hay have been used as a relatively
low-density ballast.
the NASS vehicle measurement procedures will provide a link between full-scale crash test results and
real-world crashes.
Occupant compartment deformation is used as an indicator of the potential for occupant injury. The risk
of ínjury is believed to be related to the extent of deformation and thus special care must be taken to
accurately measure occupant compaftment deformations. This rneasurement is complicated by the fact
that severe crashes can distort the entire occupant compartment. As described in Appendix E, pre-test
measurements should be made using two different reference axes in order to minimize the errors associ-
ated with reflected damage to the vehicle interior.
Sophisticated collision victim simulation (CVS) mbdels are also gaining expanded applications. CVS
models incorporate three-dimensional simulations of an occupant's motions during a crash. Evaluation of
occupant risk during a crash test involves using vehicle trajectory and acceleration data from the test as
input for the CVS program. The program then computes the dynamic response of an occupant positioned
anywhere in the passenger compartment in either a restrained or an unrestrained condition. However,
the amount and complexity of input data required for CVS programs, the cost of running the program
and, more importantly, the absence of any past record of performance and demonstrated efficacy of these
programs to predict occupant risk essentially precludes their application at this time.
Breakaway Signs and Luminaire Supports-The energy dissipated during an impact with a break-
away system is influenced by the stiffness of the front structure of the striking vehicle. In general, a
softer nose structure will produce greater energy dissipation prior to activation of the breakaway mech-
anism. Fortunately, crushable nose systems used on today's surrogate vehicles are believed to be much
softer than vehicle structures found on small cars today. As a result, safety features tested with the older
crushable nose systems are still believed to be acceptable. However, with the mass increase associated
with the test vehicles presented herein and the greatly improved crashworthiness of modern passenger
cars, it is no longer appropriate to continue to test with outdated surrogate vehicles. In order for a sur-
rogate vehicle to be utilized in place of a production vehicle fol testing of large breakaway structures,
the fi'ontal crush stiffness must be re-calibrated to match a vphicle meeting the requirements of Section
4.2.1 . Vehicle mass characteristics, such as weight distribution and yaw moments of inertia, have also
changed significantly. Hence, surrogate vehicle mass characteristics must also be adjusted to match an
appropriate production vehicle.
Work-Zone Traffic Control DeVices-\üork-zone traffic control devices include a wide range of
free-standing features that are used to channelize traffic and to r¡/arn or instruct motorists. The small
mass of these free-standing systems greatly limits the nraximum deceleration that could be produced
during a vehicular impact. Instead, the primary concern during an impact with these features is that a
component of the device will penetrate into the occupant compartment or cause damage to the vehicle's
windshield that obstructs a driver's ability to see other objects in the work zone. Hence, the trajectory
of the traffic control device during an impact with a surrogate vehicle is used to evaluate the potential
for test failure. It is necessary to mount a windshield and vehicle roof area on top of the bogie vehicle
in order to determine the potential for a work-zone device penetrating into a vehicle. If the device is
shown to clear the top of the vehicle structure without contacting the windshield region of the surrogate
vehicle, the test is considered a success. Before a bogie vehicle is used in place of a production vehicle
to evaluate the performance of work-zone traffic control devices, it must be calibrated against full-scale
crash tests of similar systems.
It is recommended that any surrogate test device be configured to model a specific production vehicle,
as opposed to a generic vehicle, with the stipulation that the vehicle being modeled meet specifications
for production model test vehicles, i.e., specifications that define tolerances on age, weight, etc. This is
by far the less expensive of the two options since properties of only one vehicle have to be measured
and the validation process involves crash testing with only one vehicle model. It would be desirable for
FHWA or NCHRP to establish a project in which all bogie properties would be updated and validated
periodically to keep cument bogies within specifications. This would be the most efficient approach
since each testing agency would not have to do it independently, and it would also ensure uniformity
and consistency among the testing agencies.
Guardrail Posts and Blockouts-Bogie vehicle tests are often used to assess post and blockout
strength when considering altemative materials for use in established banier systems. Dynamic testing
of posts and blockouts must assure that the proposed replacement components are structurally equiv-
alent to the original materials and shapes. Similarities in the results of bogie vehicle and/or pendulum
tests of the proposed replacement and original components are often used as the basis for assefting
stluctul'al equivalency. Impact forces are nolmally calculated fì'om measured bogie or pendulum ac-
celerations. Unfortunately, post inertia has been shown to contribute significantly to measured bogie
accelerations. Utilizing a crushable nose system on the front of the impacting vehicle has been shown to
virrually eliminate the inertial contributions to the measured accelerations. Note that the crushable nose
system does not need to be the same as that used for breakaway support testing. Instead, the soft nose
system merely needs to have a crush initiation stress that corresponds to an impact force that is below
the expected post strength and a divergence stress that corresponds to an impact force that is significant-
ly above the estimated strength (63).
In order for two post and/or blockout systems to be structurally similar, the force-deflection character-
istics must at least be comparable. A proposed replacement post that can sustain the same ultimate load
as a wood post, but only after deflecting twice as far, cannot be considered to be structurally similar.
Similarly, a blockout that exhibits 25 percent compression during bogie or pendulum testing cannot be
considered structurally similar to a wood block that exhibits only 5 percent compression under the sante
test conditions. Thet'efore, it is important to measure both impact force and hardware deflection during
component testing. Note that when a soft or crushable nose system is utilized, the deflection of the
structural component must be measured directly, utilizing a string potentiometer or some other equiva-
lent device.
Guardrail posts and blockouts are often loaded in a torsional manner and fail due to a combination
of bending and twisting during full-scale impact testing. It is important that surrogate vehicle testing
procedures attempt to assess structural similarity for this loading condition as well as for perpendicular
loading. Torsional loading characteristics can be evaluated by attaching a torsion bar to the post-and-
block system and allowing the sunogate vehicle to strike the torsion bar instead of directly contacting
the post or blockout. Alternatively, when evaluating components for a relatively flexible rail element,
such as W-beam, the surrogate vehicle can be configured to strike a rail element connected between two
post-and-block systems. In both cases, the force deflection and total deformation before failure should
be similar for both the proposed replacement and the conventional barrier components.
A4.2.3 TR UCK-M OU NTED AN D TRAI LER-M OU NTED ATTEN UATOR (TMA) SU ppORT VEH |CLE
Under previous evaluation guidelines, support vehicles for TMA tests were recommended to be bal-
lasted to a mass of 19,881 lb (9000 kg). The advantage of using a standardized support truck mass is
that all TMAs are tested under similar conditions. Unfortunately, in practice, TMAs are often mounted
on suppoft trucks weighing much more than 19,881 lb (9000 kg). Users cannot be assured of adequate
impact performance when a TMA is mounted to a support truck heavier than that used in the full-scale
crash testing. Users have also questioned the performance of TMAs when mounted on very light sup-
port trucks. In this case, a high-energy impact can propel the support vehicle a long distance forward
and potentially endanger workers. In order to alleviate these problems, the guidelines were revised
to allow designers to define both a maximum and a minimum support truck weight that can be safely
used with a TMA. Tests 50, 5 I , and 52 are to be conducted with the TMA mounted to a support truck
with the defined maximum weight, and Test 53 is to be conducted with the lightest allowable support
truck weight. Note that no upper limit has been designated for suppolt truck roll-ahead distance during
Test 53. The measured roll-ahead trajectory is to be reported for this test in order for potential users to
be able to place the support truck in a safe location relative to work crews.
Historical TMA testing has shown that ballast placed in the middle of a dump truck bed will shift signifi-
cantly during impact. This shifting ballast reduces the effective weight of the TMA support vehicle. In
order to assure that a TMA will perform adequately for vehicles weighing up to the assigned maximum,
ballast shifting must be controlled. This can be accomplished by utilizing rigid ballast elements that are
either secured to the truck bed or braced against the rear of a dump truck bed. If loose ballast is utilized, it
must be placed in the bed of the dump truck in a manner that will prevent any significant shifting.
The recommended braking of the support vehicle is believed to be representative of typical in-service
conditions. Test 52 is designed to assess both occupant risks and the roll-ahead distance ofthe support
truck. It is noted that roll-ahead distances for heavier support trucks can be estimated from the "consel'-
vation of momentum" principle of mechanics and simple friction calculations. Results from Test 52 can
be used as a baseline for êstimating roll-ahead distances for heavier support trucks.
High-speed cine is essential for the study of crash dynamics to determine behavior of the test vehicle
and the test article. In addition, high-speed cine has been used by some agencies as a backup system
for determining vehicular accelerations and kinematics. Guidance for this secondary system consists
of (l) intemal or extemal timing device, (2) stationary references located in the field of view of at least
two cameras positioned 90 degrees apart, and (3) Appendix C. Layout and coordinates of referenc-
es, camera positions, and impact point should be reported. Reference targets should be located on the
side and the top ofthe test vehicle and should be ofsufficient size and distance apart to allow accurate
interpretation of the film. The instant of impact should be denoted by a flash unit placed in view of data
cameras and should also be recorded on the electronic recording device(s).
Vehicular accelerations are used in the assessment of test results through the occupant flail space model
Accelerations may also be used to estimate impact forces between the vehicle and the test article.
Although it is recommended that accelerometers should be placed at the vehicle c. g., experience has
shown that this cannot always be done due to physical constraints within the vehicle. As a result, actual
placement of the set of accelerometers may be offset a significant distance from the center of mass.
Depending on the offset, major differences can occur between measured accelerations and those at the
center of mass for redirection impacts (such as impacts with a longitudinal barrier) or impacts which
cause angular vehicular motions.
The following procedure is recommended if accelerometers cannot be placed within 2 in. (5 cm) of the
center of mass as measured in the x-y plane. Although roll motions (rotations about the vehicle's x-axis)
of the vehicle are not accounted for, the method has been shown to give acceptable levels of accuracy
even for moderate roll motions.
Procedure:
I ) A tri-axial set of accelel'ometers, set I in Figure A- I , is mounted on a comrron block and placed
as close to the vehicle's center of mass as practical with the positive directions comesponding to
the positive sign convention given in Figure 4-6. Measurement of the vertical (z direction) accel-
eration is optional, but preferred. The set must be mounted along the fore-aft centerline (along x
axis) of the vehicle. Theoretically, it is not necessary that set I be placed near the center of mass;
however, this is recommended in the event accelerometer set 2 malfunctions. It is preferable that
distance ft1 be within +1.2 in. (3 cm) of distance 1L
2) A tri-axial set of accelerometers, set 2 in Figure A- l, is mounted as far as practical from set l,
preferably 23.6 in. (60 cm) or greater, either in front of or behind set L Note that both sets must
be mounted forward of the cab/bed interface for the 2270P vehicle. The separation distance of
the two sets should be as large as practical to reduce computational errors provided the acceler-
ometers are not placed in an area that would be expected to undergo significant local dynamic
deformations. Set 2 must also be mounted along the fore-aft centerline of the vehicle. It is prefer-
able that distance å2 be within +0.8 in. (2 cm) of distance å1.
3) Using output from the above two accelerometer sets and distances d¡ and d2,lateral,longitudi-
nal, and vertical accelerations at the center of mass are computed by Equations A4-3, developed
below. Note that fi and d2and their signs are measured with respect to the origin of the x-y-z
axes located at the center of mass. For positions shown in Figure A-1, both fi and d2 are positive.
However, it is not necessary that either be positive.
4) Values of d1, d2, hy, h2, and I/ should be recorded and reported as shown in Figure A- I
Accelerometer Set 1
-t \
C f
X QVehicle
C )
d1
d2
Accelerometer Set 1
Accelerometer Set 2
d2_ h2
av=aß+ddJ,
@q' Aa-l)
az = a,s - döJy
Where:
axg, ayg, azg longitudinal, lateral, and veftical accelerations at the center of mass;
0;,, o., oy, o, pitch and yaw rates, and pitch and yaw accelerations.
a,,=e,r-a,(rtt,+io1,)
arr=orr+d{JJ,
e4=err-dr6,
orr=arr+dz\J"
Qrr=Q"r-drõ,
Equations A4-2 can be solved to obtain the desired accelerations at the center of mass, exg, ayg, and azs,
as follows:
dzarr- dra *,
qrg =
dz-dt
dzarr- dn
,,
dz- dt
@q. Aa-3)
Note that the second and fifth equations and the third and sixth equations of setA4-2 canbe solved to
yield an explicit solution for pitch and yaw accelerations as follows:
dtr- o tz
av=
ù-dz
oyr- o y, @q. Aa-a)
0)
ù-dz
Pitch rate, @y, at any time, T, after impact can be obtained by adding the pitch rate at impact to the
integral of the first equation of set A4-4 with respect to time from irnpact to time Z. Yaw rate, o-, can be
similarly computed using the second equation of setA4-4.
Measuring and recording both the vehicle damage scale (VDS) (formerly the traffic accident data scale
(TAD)), and the collision damage classification (CDC), are recommended for the following reasons.
First, VDS has been in use for a number of years by various accident investigation agencies, and a
considerable bank of data exists relating VDS to occupant injuries. Hence, by not reporting VDS, the tie
of future tests with these historical data would be lost. And second, the National Highway Traffic Saf-ety
Administration (NHTSA) has standardized on the CDC for its multidisciplinary accident investigations.
Therefore, CDC is needed to tie test vehicle damage (in which vehicle accelerations are measured) to
real accidents in which occupant injury is documented.
CHAPTER FIVE
AS.1. GENERAL
Recommended evaluation criteria are limited to appraising impact performance of highway features for
idealized vehicle crash test conditions. The basic purpose ofcrash tests is to screen out those candidate
systems with functional deficiencies and to compare the relative merits of two or more promising candi-
date safety features. The test results are insufficient to project the overall performance of a safety feature
for in-service use or in an actual collision situation. Final evaluation ofa safety feature should be based
on carefully documented in-service use.
Criteria for evaluating a vehicular crash test of a safety feature are patterned after those in NCHRP
Report 350 (129) and consist ofthree interrelated factors: structural adequacy, occupant risk, and
post-impact vehicle trajectory. In comparison to NCHRP Report 350, the present criteria presented in
Table 5-l incorporate the following changes and/or modifications (further discussions of these items are
given in following sections):
(a) Criterion A-Additional language was incorporated to clarifr acceptable vehicle action
(b) Criterion D-Limits were specified for deformation or intrusion of occupant compartment.
(c) Criterion F-Limits were specified for maximum acceptable roll and pitch angles.
(d) Criteria H and l-Precision was specified for maximum acceptable occupant impact velocity
(OIV) and ridedown acceleration (RA).
(e) Criterion J on the use of the Hybrid III dummy as an optional measure of occupant risk for
frontal impacts was deleted.
The "structural adequacy" factor essentially assesses the feature from a structural and mechanical as-
pect. Depending on the feature, conditions to be examined include:
Strength-For longitudinal ban'iers, this requires containment and redirection of the test ve-
hicle. The condition of controlled stopping while the vehicle remains in contact with the rail is
also considered satisfactory and added to Criterion A. Terminals and redirective crash cushions
should develop necessary anchoring forces for anticipated site conditions.
2. Geometry-Longitudinal barrier rail members should engage the impacting vehicle at proper
height to prevent the vehicle from underriding or overriding the installation. As a general rule,
the vehicle-barier contact surface should facilitate a smooth redirection. Rail discontinuities
such as splices and transitions and other elements such as support posts should not cause snag-
ging to the extent that occupant risk criteria would not be met, or another failure mode would
occur. Shaped rigid baniers, such as the New Jersey concrete barrier, should be designed to
consider the stability oftest vehicles.
In general, a safety feature should perform its function of redirecting, containing, stopping, or permit-
ting controlled penetration of the test vehicles in a predictable and safe manner. Violent roll or rollover,
pitching, and spinout ofthe vehicle are unacceptable behavior, indicative ofunstable and unpredictable
dynamic interactions.
Relationships between occupant risk and vehicle dynamics during interaction with a roadside safety fea-
ture are extremely difficult to quantifli because they involve such important, but widely varying, factors
such as occupant physiology, size, seating position, attitude and restraint, and vehicle interior geometry
and safety features. There are sophisticated analytical and experimental tools available that can better
define these relationships, such as the crash victim simulator (CVS) computer program (54) and use
of instrumented anthropometric dummies. However, the use of these tools was considered unfeasible
for the present document because of: (a) costs associated with their purchase and/or use, (b) level of
instrumentation and expertise needed, and (c) the absence of experience by testing agencies involved in
evaluating highway safety features. Studies are needed to better define feasibility and effectiveness of
tools of this type in improving occupant risk assessment in crash tests.
First, a clear distinction was made between: (a) penetration, in which a component of the test article
actually penetrates into the occupant companment; and (b) intrusion or deformation, in which the oc-
cupant compartment is deformed and reduced in size. No penetration by any element of the test afticle
into the occupant compartment is allowed.
Second, for deformation or intrusion, it was recognized that the potential for serious injury varies by the
area of the vehicle damaged. For example, deformation in the roof area may potentially be more serious
than the wheel/foot well area due to the proximity of the head to the roof. Thus, the limiting extents of
deformation should also vary accordingly.
Third, the limiting extents for deformation were based on the recommended guidelines developed by
the Insurance Institute for Highway Safety (IIHS) for evaluating structural performance of vehicles in
offset frontal crash tests. The recommended guidelines were based on results from selected full-scale
offset frontal crashes, which are as follows:
. Roof < 4 in. (102 mmFA much lower limiting extent of deformation was selected for the roof area
since the headroom inside the vehicle is limited and impacts to the head are more likely to result in
serious or fatal injuries.
. Windshield-No tear of plastic liner and maximum deformation of 3 tn. (76 mm). A much lower
limiting extent of deformation was selected for the windshield area since an occupant, particularly an
unbelted occrìpant, would move forward toward the windshield. Thus, deformation of the windshield
would increase the potential of the occupant impacting the windshield and could lead to more severe
injuries. Also, tearing of the plastic liner could lead to penetration of the occupant compartment and
thus is not permitted. Note that a tear in the windshield's plastic liner is only precluded when there
is a potential for a test article component to penetl'ate into the vehicle. Tealing of the plastic liner
produced when a test article contacts windshield support structures is acceptable.
' Window-No shattering of a side window resulting from direct contact with a structural member of
the test article. In cases where the windows are laminated, the guidelines for windshields will apply.
It was observed that the occupants' head would typically strike the side window in redirectional im-
pacts. Thus,if the side impact was shattered from direct contact with a structural member of the test
article, it is logical to assume that the occupant's head could also strike the structural member and
result in serious injuries.
' Wheel/foot well and toe pan areas < 9 in. (229 mm¡-The limiting extent for deformation in these
areas coffesponds to the acceptable range as recommended by the IIHS. Due to the proximity of the
occupant's lower extremities in these areas, any deformation would likely result in injuries to the
lower extremities. rWhile such injuries are typically not life-threatening, they could be severe, result-
ing in long-term disabilities.
' Side front panel (forward of A-pillar) < 12 in. (305 mm)-The limiting extent for deformation in
this area corresponds to the marginal range as recommended by the IIHS. While deformation in the
side front panel still could result in injuries to the lower extremities, the likelihood is lower sinbe the
occupant's lower extremities are typically located further away from the side front panel.
' Front side door area (above seat) < 9 in. (229 mmFThe limiting extent for deformation in these
areas corresponds to the acceptable range as recommended by the IIHS. Due to the proximity of the
occupant's torso to this area, any deformation would likely result in serious injuries to the occupant.
' Front side door area (below seat) < 12 in. (305 mmFThe limiting extent for deformation in this
area corresponds to the marginal range as recommended by the IIHS. It is reasonable to assume that
the seat assembly would shield the occupant from some ofthe deformation and alarger limiting
extent is thus appropriate.
' Floor pan and transmission tunnel areas ( 12 in. (305 mmFThe limiting extent for deformation
in this area corresponds to the marginal range as recommended by the IIHS. Deformation in these ar-
eas is typically the result of induced damages, which has a much lower potential for serious injuries
than deformation from direct contacts.
Ttshould be emphasized that any occupant compartment damage should be carefully documented in
the form of photographs and measurements, particularly for area(s) where the maximum extent(s) are
exceeded. The same applies for any damage to, or rupture of, the fuel tank, oil pan, or other features that
might serve as a surrogate of a fuel tank.
Flail-Space Model
NCHRP Repoft 350 (129) continued to use the simplified point mass, flail-space model first developed
under NCHRP Report 230 (92) for assessing risks to occupants due to vehicular accelerations. Two
measures of risk are used: (l) occupant impact velocity (OIV)-the velocity at which a hypothetical
occupant impacts a hypothetical interior surface; and (2) ridedown acceleration-acceleration experi-
enced by the occupant subsequent to contact with the interior surface. The flail space model has served
its intended purpose well, and there are no indications that features designed and assessed thereby have
performed adversely in service. Thus, it was decided to retain the flail space model for the present
document.
All testing agencies are now using a standardized computer program for determination of the occu-
pant risk factors. This should promote consistency among testing agencies and assure accuracy of the
calculations.
One source of potential emor is inconsistencies in positioning accelerometers and rate gyros used in
measuring accelerations and rotational rates, i.e., they are not being placed at the vehicle's center of
mass. Recommendations contained in Section 4.3.2 should greatly reduce or eliminate this problem.
In the flail-space approach, lateral and longitudinal, but not vertical, vehicular accelerations measured
at the vehicle's center-oÊmass are used. By requiring that the vehicle in the occupant risk test remain
upright throughout the collision, it is believed that the vertical component of vehicle acceleration be-
comes of secondary importance with regard to occupant kinematics for the level terrain tests described
in this document and for most roadside features. Consequently, the vertical acceleration is considered an
optional factor at present and has been neglected in the flail-space calculations.
The performance design strategy for a feature should be to: (l) keep the occupant-vehicle interior
impact velocity low by minimizing average vehicle accelerations or vehicle velocity change during
the time the occupant is traveling through the occupant space; and (2) limit peak vehicle accelerations
during occupant ridedown.
Some questions were raised about the rounding of limiting values due to conversions from the English
units in NCHRP Report 230 (92) to the SI system in NCHRP Report 350. To avoid such arguments, the
maximum limiting values are set to the following: 40 frls (12.20 m/s) for the longitudinal and lateral
occupant impact velocity; l6 ff/s (4.9 m/s) for the longitudinal occupant impact velocity of breakaway
support structures; and 20.49 G for the ridedown acceleration.
The limiting value for the longitudinal occupant impact veìocity was originally set at 40 ft/s
(12.2 mls) in NCHRP Report 230, which was then hard converted to 39.4 ft/s (12.0 m/s) in NCHRP
Report 350. Since a number of safety devices were approved with the maximum limiting values at
40 fVs (12.2 mls), this limit is used to allow for continuity among the guidelines. The limiting value
for the longitudinal occupant impact velocity of breakaway support structures was originally set at 15
fls (4.6 m/s) with an impact speed of 20 mph (32.2 km/h). These values were hard converted to I 6 ftls
(4.9 m/s) with an impact speed of 21.8 mph (35.0 km/h). Thus, the maximum limiting value of l6 ff/s
(4.9 m/s) already represents a significant easing of the requirement and should be considered as the
maximum limiting value. NCHRP Report 230 specified 20 G as the maximum limiting value, which
was retained for NCHRP Report 350. It was argued by some hardware developers that, since the limit
is set as an integer, the actual maximum limiting values should be 20.49 G to account for rounding and
some devices were accepted with this limiting value. Again, to maintain continuity among guidelines, a
maximum limiting value of 20.49 G is recommended.
Calculation Procedures
The expression for occupant impact velocity is
where, x: 2 ft (0.6 m) and y: I ft (0.3 m). Acceleration in the x direction is integrated twice with
respect to time to find the value of time, f"*, at which the double integration equals 2 ft (0.6 m).
Acceleration in they direction is integrated twice with respect to time to find the value of time, /r", at
which the double integration equals I ft (0.3 m). Time /* is the smaller of f,' and /r*.
In tests of breakaway features, the impulse on the vehicle may be relatively small and of shott duration.
It is not unusual in such tests for x andy to be less than2 ft and I ft (0.6 m and 0.3 m), respectively,
during the period in which accelerations are recorded or up to the time brakes are applied to the test
vehicle. In such cases, it is recommended that the occupant impact velocity be set equal to the vehicle's
change in velocity that occurs during contact with the test article, or parts thereof. If parts of the test
article remain with the vehicle after impact, the vehicle's change in velocity should be computed at the
time the vehicle clears the footing or foundation of the test article.
For the ridedown acceleration to produce occupant injury it should have at least a minimum dura-
tion ranging from 0.007 to 0.04 s, depending on body component (158). Thus, vehicular acceleration
"spikes" ofduration less than 0.007 s are not critical and should be averaged from the pulse. An arbi-
trary duration of 0.010 s has been selected as a convenient and somewhat conservative time base for
averaging accelerations for occupant risk assessment. This is accomplished by taking a moving l0-ms
average ofvehicular "instantaneous" accelerations in the.x and y directions, subsequent to /".
The occupant impact velocity and the highest l0-ms average acceleration values are then compared to
recommended limits; it is desirable that both values are below the "preferable" limits; values in excess
of the "maxirlull" limits are considèred to be unacceptable.
Recommendations relative to the measurement of accelerations are given in Section 4.3.2 andin
Appendix C. Further, for purposes of standardization of occupant risk calculation procedures, the
fol lowing are recommended:
(1) Prior to integration using above formulas, accelerometer analog data should be digitized at
1,500 samples per second. This is consistent with recommendations of Appendix C, Section
8.2. It is recommended therein that the sample rate be, at a minimum, eight times F¡,, where
F¡, = I 80 for measurement of vehicular response. Note that F n
" I = 1,440, which is rounded to
1,500 for convenience and ease ofintegration.
(2) It is reconrmended the "linear acceleration" assumption or the equivalent "trapezoidal rule" be
used to integrate the digitized accelerometer data. As such, accelerations are assumed to vaty
linearly over each time step tito li+ l. Description of the trapezoidal rule can be found in most
numerical methods textbooks.
In general, the ideal after-collision vehicular trajectory performance goal for all features should be that
the vehicle trajectory and final stopping position should not intrude into the adjacent or opposing traffic
stream. For breakaway or yielding supports, the trajectory of a vehicle after it has collided with a test
article that satisfies structural adequacy and occupant risk requirements is generally away from the traf-
fic stream and is, hence, less critical. For end-on impacts into crash cushions and barrier terminals that
function as crash cushions, preferably the final position of the vehicle should be next to the test device.
For redirectional performance tests of length of need, transitions, terminals and redirective crash cush-
ions, the after-collision trajectory is more difficult to assess. The after-collision trajectory may be one of
the least repeatable perfotmance factors because of variation in method and timing of brake application.
Further, variables that are in part related to the specific model of vehicle selected for tests, such as dam-
age to vehicle suspension, tires, etc., may alter the vehicle's stability and path. Moreover, because driver
response in avoiding secondary collisions is not simulated in the crash tests, it seems inappropriate to
predict in-service performance based on the complete test trajectory.
User agencies should assess the post-impact vehicle trajectory of a roadside safety feature in light of
the actual field conditions. For many tests, a scaled diagram showing the post-impact trajectory of the
vehicle, including the point of final rest, should provide sufficient infonnation for the user agencies to
make their assessment. Additional information is provided for some features, such as the "exit bo{'
criterion assessment for redirective devices and the "rebound" velocity for reusable crash cushions and
attenuation systems.
For redirective devices, it is preferable that the vehicle be smoothly redirected. As shown in Figure 5- l,
the exit box is defined by the initial tlaffic face of the barrier and a line parallel to the initial traffic face
of the barrier, at a distance A plus the width of the vehicle plus l6 percent of the length of the vehicle,
starting at the final intersection (break) of the wheel track with the initial traffìc face of the barrier for
a distance of B. It is preferable for the vehicle to exit within the "exit box," i.e., all wheel tracks of the
vehicle should not cross the parallel line within the distance B. As a point of reference, the "exit box" is
equivalent to a maximum exit angle of 12.4 degrees.
Vehicle rebound has been noted for some reusable crash cushions and attenuation systems, which could
potentially lead to secondary collisions. In order to provide user agencies with the necessary informa-
tion regarding the use and placement ofsuch crash cushions, testing agencies are required to document
and report the rebound velocity and point of final rest, as outlined in Section 5.2.3. While these report-
ing requirements are intended mainly for reusable crash cushions and attenuation systems, they would
also apply to any feature resulting in rebound of the test vehicle.
Specific test and evaluation guidelines for geometric features are not provided due to the largely non-
standard and variable nature ofsuch features. However, it should be a goal oftransportation agencies to
design and implement geometric features that meet the spirit, if not the specifics, of safety recommenda-
tions for the more well-defined roadside safety features.
Evaluation guidelines given in this section were derived from a review of past practices and the collec-
tive expertise of those involved in preparing the document. They are, of necessity, general and may be
amended as necessary to accommodate special designs or test conditions.
CHAPTER SIX
It should be borne in mind that the leporting guidelines presented herein are intended as minimum re-
quirernents. The underlying philosophy is that the test(s) should be documented in sufficient detail such
that the test(s) could be repeated by anothel testing laboratory without further inquiry. It is not possible
to include all of the details necessary to achieve this goal. The actual level of detail needed will vary
depending on test objectives, test article, and test levels involved. If there is any question about wheth-
er some details need to be included in the report, it is recommended that the additional information be
included to assure that the testing is adequately documented.
CHAPTER SEVEN
AT.l PURPOSE
In-service performance evaluation guidelines are intended to encourage a more consistent, systematic,
and thorough implementation of new devices and to monitor the field performance of safety features on a
continuing basis. With careful monitoring of a new device, unanticipated problems and design deficiencies
can be identified before the feature has been installed in an excessive number of sites. Moreoveç all the af-
fected depaftments will have an opportunity to observe the performance of the device with respect to their
operations. For instance, there may be minor design changes recommended by the maintenance groups
that may reduce nomal maintenance or damage repair costs. However, care should be taken not to make
changes in design details that could adversely affect impact performance without verification of adequate
performance through full-scale crash testing or other acceptable means.
Continuous monitoring of installed safety features assures that changes in vehicle and traffic character-
istics do not adversely affect the field performance of the devices. This aspect of in-service performance
evaluation was not included in the guidelines set forth in NCHRP Report 350 (129), but added to this
document in recognition of the need to monitor field performance in light of changes in the vehicle
population and other safety developments.
FHV/A continues to serve as the key arbiter in establishing acceptability ofnew safety features, espe-
cially those used on federal-aid highways. The acceptability ofa new safety feature is based on design
details, specifications, and crash test results, but there is no assessment of the in-service performance of
the features. Thus, the responsibility of in-service performance evaluation of new and existing safe-
ty features would fall on the state transportation agencies. For proprietary safety devices, it may be
appropriate for the developer or manufacturer to sponsor or contribute to the in-service performance
evaluation.
The need for in-service performance evaluation is well recognized, but there has only been limited
implementation of in-service performance evaluation programs due to lack of resources, both in terms
of funding and manpower. State transportation agencies are faced with the problem of ever increasing
workload and shrinking budget, and it may be difficult to divert the resources from other, more pressing
needs. Nevertheless, the establishment of an in-service performance evaluation program, even if it is
on a limited scale, would be highly beneficial and strongly recommended. Altematives, such as pooling
of resources from multiple states and cooperation with developers and manufacturers on proprietary
devices, should be explored. Results from in-service performance evaluation studies should be made
available to other state transpoftation agencies and disseminated through such channels as National
Technical Information Services (NITIS), FHWA regional resource centers, and pooled fund consortiums.
The establishment of a new national center on in-service performance evaluation is also recommended
This center would serve to:
In summary the need for in-service performance evaluation of existing and new roadside safet¡r features
cannot be overly emphasized. More in-service performance evaluation sttrdies are needed to assess and
monitor field performance, and the results of the studies should be disseminated to the user agencies.
Soil Strength
Performance Test ]E
81 PURPOSE
h" purpose of this dynamic soil performance test is to assure consistency and appropriateness of
T
I the foundation conditions for soil-based installations. By basing soil conditions on dynamic load
response rather than a material and method specification for installation, consistency between tests, both
among and within different testing agencies, will be assured. Fill materials, installation methods, and
in-situ soil conditions will influence the results of this test. Also, in-situ soil conditions may even vary
with location at a given test facility, so several tests may be required to assure compliance of installa-
tions at various locations.
82 APPLICATION
While there are in-situ soil conditions that will provide the required minimum performance without
the use of structural fill surrounding the installation, a stiffer fill material surrounding the embedded
test post will be required in most situations to achieve the required soil resistance. The required lateral
extent (hole size) of this fill established utilizing this standard instrumented post will be the minimum
lateral extent required for testing of all future soil-based appurtenances. Note that it is acceptable to
utilize a greater lateral extent of fill as this test establishes a minimum soil resistance. Also, the depth of
the structural fill should always extend below the installed appurtenance.
83 INSTRUMENTED POST
The use of an instrumented post as a surrogate to dynamic testing with a bogie or pendulum is intended
to minimize the effoft and costs associated with dynamic testing. The test post detail shown in Figure
B-l is offered as an acceptable design. However, other variants incorporating similar load measurement
capabilities willbe acceptable. Use of a standard W6 by 16 (W152 by 23.S) structuralshape is required
to assure compatibility between testing houses. Any appropriate load cell is acceptable for testing.
However, the following notes of caution should be bome in mind:
' Second, it is not acceptable to utilize acceleration measurements from the impacting vehicle to deter-
mine post loading. The structural differences between different impactors can significantly influence
the interpretation of load, which could in turn lead to inconsistencies between testing facilities.
4 in.
n
zslsin. l2in.
zsla llzin.
c
11lt
7la-in.
n 211 tn
t 3/e n
¡-D
I
\,, -in. Dia.
Bolt
Washer
Load Cell
Washer
Nut
Note: Put nut on
hand
Ground Line
W6x15
2ft4
1l¿in.
'tl4
111¿in. jl4
1l¿in.
3fi8in.
84 POST PLACEMENT
As shown in Figure B-2, the standard instrumented post should be placed with a 32 in. (813 mm) top
height and 40 in. (1,016 mm) of embedment. The impact location should be at25 in. (635 mm) above
the ground line. Note that, because the objective of this performance standard is to establish a mini-
mum load/deflection relationship, the method employed to place the post is to be decided by the test
facility and may vary among testing agencies. Thus, it is important that the method employed for the
installation be documented in detail. Required documentation of the installation is shown in Figure 3-1,
Recommended Summary Sheet for Strong Soil Test Results.
2-ftDia
103 in.
32in.
25 in.
72in.
40 in.
43 in.
85 DYNAM¡C TEST
It is recognized that different testing facilities have various means with which to perform the dynamic
testing. The specifications for the dynamic test are purposely set to be as open as possible. The desired
impact conditions are:
As shown in Figure B-3, the minimum post load required for deflections from 5 in. (125 mm) to 20 in.
(500 mm), measured at the impact height of 25 in. (635 mm), is 7,500 lbs (33.36 kN). If this force level
is not achieved (as shown in Figure B-3) with a particular fill configuration, fill material, placement
method, and native soil condition, it will be necessary to increase the lateral extent of the fill material or
use a stiffer fill material. It would be most helpful if each testing facility establishes the unique relation-
ships between the post load and the lateral extent of the fill material and other influencing factors, such
as moisture content.
As mentioned previously, the native soil at a given testing facility may vary from location to location. In
order to assure adequate soil resistance, the dynamic compliance tests should be performed at a location
where the native soil is believed to have the lowest stiffness. In cases where the site of the dynamic test
has different native soil than sites where actual crash tests are conducted, the"artificial" native soil used
for the compliance testing should have lower stiffness than the actual native soil. In the case of cohesive
soils, a modulus determined from an unconfined compressive strength test (ASTM D2166) for both the
native and "artificial" soil can be compared. If the native soils are noncohesive, the recompacted "artifi-
cial" soil should have a lower density than that of the naturally occurring material. The native soil used
in the dynamic compliance test should be documented, particularly if the soil used is different from the
naturally occurring soi ls.
14000
12000 rl
tt
lr
I 0000 rl
ll
8000
Ît '/,
!
!(E
o
-J
6000
4000
iI I
it
2000
0
r I
-2000
0510152025303540
Displacement (in.)
---- 2214 PT-1 EDR-3 Load (lb) 2214 PT-1 Load Cell Load (lb)
-
Figure B-3. Dynamic and Static Test Results for Standard Post Test
Since soil performance can be influenced by a number of uncontrolled factors, it is necessary to assure
that appropriate soil conditions exist on the test day. This could be accomplished by running a standard
dynamic test in the vicinity of the anticipated full-scale test. However, there are obvious logistical and
economic problems associated with this approach. Thus, a surrogate static test was recommended in
place of the dynamic test to assure test day compliance.
The relationship between native soil and fill materials used in the standard dynamic test must be similar
to those utilized for full-scale crash tests to assure consistency. It is recommended that the soil classi-
fications of both the native soil and fill material from the standard dynamic test and the crash test be
the same as defined by ASTM D2487. This comparison is available from Figures 3-l and 3-2, both of
which are recommended for inclusion in each crash test report. In addition to this classification, it is
recommended that any noncohesive materials utilized in the construction of a soil-based system fall
within +10 percent of the results of the sieve analysis conducted on the fill material utilized in the initial
standard dynamic test.
In order to assure that the test installation will have sufficient soil strength, a static push/pull test may be
utilized, as shown in Figure B-4. To establish criteria for this static test, a second installation identical
to that of the dynamic test should be constructed. This second installation will then be statically tested
utilizing any convenient loading system, such as a winch or hydraulic ram. The standard post should be
loaded at the same height as in the dynamic test. A plot of the load/deformation either measured or de-
termined at the load height should be included on the soil test documentation, as shorvn in Figure B-3.
It is recommended to use the same load cell as utilized in the dynamic test for the static test to minimize
variability due to instrumentation. Deflection for the static test can be measured from any convenient
reference point by hand or electronically measured as desired.
2-ftDia.
Granular Fill
or
Hydraulic 32in.
Cylinder 25in
72in.
40 in.
43 in.
During installation of each article for full-scale crash testing, two standard W6 by l6 posts should be
installed in the immediate vicinity of the full-scale test article. These posts should utilize the same fill
materials and installation procedures used in the standard dynamic test, even if the fill geometry is
different for the full-scale test article. In addition, the static test posts should be installed at the same
time as the posts in the impact region of the tested system. This installation should be located in an
area where rainfall and other natural events will impact the test post in the same manner as the installed
system. On the same day, but prior to the full-scale crash test, a static test of one of the standard posts
should be performed. A plot of the load versus deformation should be compared to the standard estab-
lished in the initial testing, as shown in Figure B-5. If the load measurements are at least 90 percent of
l5 in. (125, 250, and 375 mm), the
those of the initial standard installation at deflections of 5, 10, and
soil condition is considered acceptable for full-scale crash testing. As shown in Figure B-5, a significant
rain the previous evening has had a significant effect on the performance in this particular installation,
and the soil conditions are not acceptable for testing. In this instance, additional time for drying was
lequired prior to full-scale testing. The second installed standard post is intended for the purpose of
retesting in case the soil conditions are not acceptable in the first test.
Comparison of Static Load Test Results and Required Minimum: Load vs. Displacement
6000
5000
-- S-
4000
4 \
3000
2000
't000
z v \
--È\ \
\
7 l-r¡-r,
')
0
0 5 10 '15 20 25 30 35 40 45
Displacement (in.)
. WetTest
Figure B-5. Test Day Static Load Test Compared to Standard Test
1
I
I
t
J
O 2016 by the American Association olState Highway antl Transportation Oflicials.
All rights reserved. Duplication ís a violation of applícable law.
174 | Manual forAssessingSafety Hardware
PRACTICE
Suoersedlno J211-1 DEC2003
RATIONALE
Revisions to SAE J211 are a continuing process and are considered at each five year review Changes were made as
part of an ongoing effort of harmonization with ISO 6487 and include additional definitions, tightening of the filter corridors,
procedures for verifring system performance, changes and additions to CFC recommendations, specification of
transducer equivalency, and changes to data format recommendations.
,I. SCOPE
This recommended practice outlines a series of performance recommendations, which concem the whole data channel.
These recommendations are not subject to any variat¡on and all of them shall be adhered to by any agenry conducting
tests to this practice. However, the method of demonstraling compliance with the recommendations is flexible and can be
adapted to suit the needs ofthe particular equipment the agency is using.
It is not intended that each recommendation be taken in a literal sense, as necessitating a single test to demonstrate that
the recommendation is met. Rather, it is intended that any agency proposing to conduct tests to this practice shall be
able to demonstrate that if such a single test could be and were carried out, then their equipment would meet the
recommendations. This demonstration shall be undertaken on the basis of reasonable deductions from evidence in their
possession, such as the results of partial tests.
ln some systems it may be necessary to divide the whole channel into subsystems, for calibration and checking purposes.
The recommendations have been written only for the whole channel, as this is the sole route by which subsystem
performances affect the quality of the output. lf it is difficult lo measure the whole channel peformance, whích is usually
the case, the test agency may treat the channel as two or more convenient subsystems. The whole channel performance
could then be demonstrated on the basis of subsystem results, together with a rationale for combining the subsystem
results together.
Part 1 of this recommended practice covers electronic instrumentation and Part 2 covers photographic instrumentation.
1.1 Purpose
The purpose of this SAE Recommended practice is to provide guidelines and recommendations for the techniques of
measurement used in impact tests. The aim is to achieve uniformity ¡n instrumentation practice and in reporting test
results. Use of this recommended practice will provide a basis for meaningful comparisons of test results from different
sources,
SAE Techn¡æl Slanderds Board Rules provide that 'Th¡s report is published by SAE lo ãdvance the slate of lechn¡câl end sclences. Tho use of lh¡s Eport is
entirely voluntary, ãnd ¡ts appl¡€tility ând $¡lebilily for any part¡@lâr use, indud¡n9 any palent ¡nfringemeDt arising lherekom, is
SAE reviews eách lecån¡€l report at least every live years ãt wh¡ch time it may be Gâlt¡med, revised, or €ncelled. SAE ¡nvitgs your witten comments and suggest¡ons.
Copyñght @ 2007 SAE lntemal¡onal
All rights reseryed. No part of lh¡s publical¡on may be reproducêd, stored in ã relrieval system or transmilted, in any fom or by any meâns, electron¡c, mechãn¡€|,
pholocopy¡ng, recording, or otheruise, without lhe pñorwitt€n pêmiss¡on ol SAE.
TO PLACE A DOCUMENT ORDER: Tel: 877{06-7323 (¡nsld6 USA and Canâda}
Tel: ?2t?76-4970 (outside ÌJSAI
Fax: 721-77G479O
Reprinted with permission from SAE J21111@.2007 SAE lnternational. Additional copies
may be purchased directly from SAE at http://sae.org or call (8771 606-7323.
1 .2 Field of Application
The instrumentat¡on as defined in this recommended practice applies in particular to impact tests for road vehicles,
including tests oftheir sub-assemblies, and occupant surrogates.
2. REFERENCES
2.1 ApplicablePublications
The following publications form a part of this specification to the elrtent specified herein. Unless otherwise indicated, the
latest version of SAE publications shall apply.
Available from SAE, 400 Commonwealth Drive, Wanendale, PA 15096-0001, Tel: 877-60È7323 (inside USA and
Canada) o¡ 7 24-7 7 6497 0 (outside USA), www.sae. oro
2.'1.2 NHTSAPublication
Available from Department of Transportation, The Ofñce of Crashworthiness Research, 408 7th Street SW Washington,
DC 20590.
NHTSAVersionSTestReferenceGuide: Volumel-VehicleTests,Volume2-Biomechanics,Volume3-Components,
and Volume 4 - Signal Waveform Generator
3. DEFINITIONS
The definitions in paragraphs 3.5 to 3.14 apply to the whole data channel, as defined in paragraph 3.1.
All of the instrumentation from and including a single transducer (or multîple transducers whose outputs are combined in
some specified way) up to and including any analysis procedures that may alter the frequency content or the amplitude
content or the t¡ming of data. lt also includes all cabling and interconnections.
3.2 Transducer
The first device in a data channel, used to convert a phys¡cal quantity to be measured into a second quantity (such as an
electrical voltage) which can be processed by the remainder of the channel. For transducer equivalency, see Append¡x B.
The maximum usable linear range of a data channel. For ATD transducers, see SAE J2570.
That value of a data channel determined by the component of the channel with the lowest full scale level. This is
expressed in terms of the measured variable (input). For example, F.S. = 50 G, 1 000 N, 1 m/s, etc.
The designation for a data channel that meets certain amplitude characteristics as specified by this recommended
practice. The CAC number is numerically equal to the upper limit of the measurement range (that is, equivalent to the
data channel full scale).
The channel frequency class is designated by a number indicating that the channel ftequenry response lies within limits
specified by Figure 1 for CFCs of 1 000 and 600, and by Figure 2 for CFCs of 60 and 1 80.
NOTE: Figure t has not changed in reference to J21 1 March 1 995 except to speci! nodal points instead of slopes. This
number and the value of the frequency FH in hertz are numerically equal.
The value measured and read during the calibration of a data channel (see paragraph 4.6).
3.9 SensitivityCoefficient
The slope of the stra¡ght line representing the best fit to the calibration values determined by the method of least squares
within the channel amplitude class.
The arithmetic mean of the sensitiv¡ty coeffìcients evaluated over frequencies which are evenly spaced on a logarithmic
scale between Ft and Fnl2.5.
The ratio, in percent, of the maximum difference between the calibration value and the corresponding value read on the
straight line defined in paragraph 3.9 at the upper limit of the channel amplitude class (data channel full scale).
3.12 Sensitivity
The ratio of the output signal (in equivalent physical units) to the input signal (physical excitation), when an excitation is
applied to the transducer. (Example: 10.24 nVIGN for a strain gage accelerometer.) For ATD transducers, see SAE
J2570.
The phase delay time of a data channel is equal to the phase delay (in radians) of a sinusoidal signal, divided by the
angular frequency of that signal (in radians per second).
3.14 Environment
The aggregate, at a given moment, of all external conditions and influences to which the data channel is subjected
NOTE: The transverse sens¡tivity is usually a function of the nominal direction of the axis chosen. For ATD transducers,
see SAE J2570.
The ratio of the transverse sensitivity to its sensitivity along its sensitive axis,
3.18 ATD
FL FH Fn zfH Fô Fr
0
AllenuôlionE
-10
lppor lowr \
FL {0.5 . 0.5
r¡ rO.5 -t0
:2O r0.5 -¡0
iO.5
Fo .3t,0
¡0 F, -40.0 -00 \
40 cFc
Frequ€nciss
FX F{ 2TH FO FJ \
800 600 1000 t200 zil9 3885
0.'r 1000 r650 2000 3498 6442
-50
FT FC Fil FO FC Fo
0
Allenual¡on9
t2 uppor lwer
-10
F! r0.5 - 0.5
FH r0.5 . 1.0
Fc .03 - r.8
Ftr .1.8 -3.8
FD - 5.2 . A.2
FE .9.2 .13.2
-30 -4f¡ -49.3
Fo
FrEqusnc¡e9
E. ç.- Ê- F.. Ê- E-
^tÃ
60 to.'t 60 75 t00 t30 t6ù t52
reolo.r reo 225 3oo rgo ago i3ro
-50
0.1
The frequency response of a data channel shall lie within the limiting curves given ¡n Figure 1 for CFCs of 1000 or 600.
For CFCs of 1p0 or 60 the frequency response of a data channel shall lie within the limiting curves given in Figure 2. The
0 dB line is defined by the calibration factor. Note: The Figure 2 envelopes have been derived to center the frequency
responses of the March 1995 J21 1 Appendix C CFC 60 and 180 filters in the conidors.
The phase delay time between the input and the output of a data channel shall be determined, and shall not vary more
than 1/(10-F¡) seconds between 0.03.FH and F¡. This includes the transducer, that is, the input is the excitation to the
transducer.
4.4 Time
A time base shall give at least 1/100 second resolution w¡th an error ofless than 1/10000 second.
The relative time delay between the sígnals of two or more data channels regardless of their frequency class, must not
exceed 1 millisecond. Two or more data channels of wtrich the slgnals are combined shall have the same frequency
class and shall not have a relative time delay greater than 1(1O.FH) seconds. This requirement appl¡es to analog signals
as well as digital signals.
The transverse sensitivity ratio of all transducers shall be less than 5% in any direction. For ATD transducers, see SAE
J2570.
4.6 Calibration
Values in this section apply to reference equipment or "standards" against which a data channel is "calibrated", that is, ¡ts
pelormance is determined.
@ 201 6 by the American Association ol State H ighway antl Transportat ion Oüisials.
All rights reserved Duplication is a viola¡ion ofapplicable law
Appendix C-Electronic & Photographic lnstrumentation Specifications | 179
4.6.1 General
A data channel shall be calibrated at least once a year against reference equipment traceable to known standards. The
methods used to carry out a comparison with reference equipment shall not introduce an error greater than 1% ofthe
CAC. The use of the reference equipment is limited to the range of frequencies for which they have been calibrated.
Subsystems of a data channel may be evaluated individually and the results factored into the accuracy of the total data
channel. This can be accomplished for example by an electrical signal of known amplitude simulating the output signal of
the transducer which allows a check to be made on the gain of the data channel, except the transducer.
The accuracy of the reference equipment shall be certified or endorsed by an approved metrology service (for example,
traceable to the National lnst¡tute of Standards and Technology).
4.6.2.1 StaticCalibration
4.6.2.1.1 Accelerations
The enor shall be less than 1 .5% of the channel amplitude class.
4.6.2.'l.2 Forces
4.6.2.1.3 Displacements
4.6.2.2 DynamicCalibration
4.6.2.2.1 Accelerations
The error in the reference accelerations expressed as a percentage of the channel amplitude class shall be less than
1.5% below 40OHz, less than 2% between 400 and 900 Hz, and less lhan 2.5o/o between 900 Hz and the maximum
frequency at which the reference acceleration is utilized (see paragraph 4.6.4).
See SAE J2570'Íor load cell specifications titled "Free Air Resonance" for force and moment info.
4.6.2.3 Time
The enor in the reference time shall be less than 'll(10 * sample rate).
The sensitivity coefficient and the linearity error shall be determined by measuring the output signal of the data channel
against a known input signal, for various values of this signal. (The input signal is referenced to well known physical data,
which is, a load or acceleration, but not voltage.)
The calibration of the data channel shall cover the whole range of the amplitude class. (This is between Fr_ and Fx/2.5.)
For bi-directional channels, both the posit¡ve and negative values shall be evaluated. lf the calibration equipment cannot
produce the required input, due to excessively high values of the quantity to be measured, calibrations shall be canied out
within the limits of these calibration standards and these limits shall be recorded in the report.
A total data channel shall be calibrated at a frequency or at a spectrum of frequencies with its significant values
comprised between Fr_ and Fn/2.5.
The response curves of phase and amplitude aga¡nst frequency for the data channel shall be determined by measurÍng
the output signals of the dala channel in terms of phase and amplitude against a known input signal, for various values of
this signal varying between F¡ and 10 times the CFC or 3000 Hz whichever is lhe lower value.
The presence of any environmental effects shall be checked (that is, electric or magnetic flux, electrostatic discharge,
cable velocity, etc.). This can be done for instance by recording the output of spare channels equipped with dummy
transducers. lf such an output signal is greater than 2% ot lhe expected data peak value, conective action shall be taken,
for instance re-allocation or replacemenl of cables.
5. DATACHANNELSELECTION
The selection of a frequency response class is dependent upon many considerations, some of which may be unique to a
particular test. The ultimate usage of the data and good engineering judgment will determine what portions of the
frequency spectrum are signifcant or useful. The various classes offrequency response in Figures 1 and2 are intended
to permit appropriate choices for different engineer¡ng requirements.
It is important to note that valid comparisons using different frequency response classes may be difi¡cult to make. lt is
useful to establish specific frequency response classes when comparing test results fom different sources. The
frequency response classes in Table 1 are recommended for that purpose. These recommendations reflect current
practices and equipment. However, it is recognized that other considerations (for example, biomechanics) may impose
special instrumentation requirements.
The channel class recommendations for a particular application shall not be considered to imply that all the frequencies
passed by that channel are significant for the application. ln several cases, such as occupant head accelerations,
headform accelerations, and femur force, the recommendation may be higher than necessary, but cunent biomechanical
knowledge will not permit a closer specification. All data are to be gathered at CFC 1000 or higher, for any purpose.
Thorax
Spine accelerations 180
Rib accelerations 1000('ì
Sternum accelerations IOOO(')
Deflections 6oo('?¡r
Lumbar
Forces 600(E
Moments 600(u'
Pelvis
Accelerations l Ooor')
Forces 600þ)
Moments 600(")
Femur/KneeÆibia/Ankle
Accelerations 1000
Forces 600
Moments 600(')
Displacements 180(')
Sled acceleration 60
Steering column loads 600
Headform acceleration 1000
'1. When overall acceleration of lhe frame or body in a given direction is desired and a hlgher kequency
response class ¡s used, readab¡llty ol lhe data may be improved by averaglng outputs of two or more
transducers ât d¡llerent locations.
2. UMTRI-86-32 and ISO/TC2ZSC12lWGg N 194.
3. Theso classlflcatlons aro needed to calculats head impact lorces basêd on n€ck lorces and head
accelerations when us¡ng an ATD,
4. SAE paper 930100.
5. sAE paper 2002-01-0806 or ISO/Tc22lsc1 2A^tGo N557.
6. Rsference Appendlx B.
7. When Force channels are multlpl¡ed by a moment arm, a CFC 600 l¡ller shall be used.
6, MOUNTINGOFTRANSDUCERS
Mechanical resonance associated with transducer mounting shall not distort readout data.
Transducers shall be mounted on dummies using a support specially provided for th¡s purpose. ln cases where
properties of non-mechanical test subjects preclude rig¡d transducer mount¡ng, an anaMical or experimental evaluation of
mounting effects on the data shall be provided.
Acceleration transducers, in particular, shall be mounted in such a way that the ¡n¡tial angle of the actual measurement
axis to the corresponding axis of the reference axis system is not greater than 5 degrees unless analyt¡cal or experimental
assessment of the effect of the mounting on the collected data is made. When multi-axial acceleratlons at a point are to
be measured, each acceleration transducer axis shall pass within 10 mm of that point, and the center of seismic mass of
each accelerometer shall be within 30 mm of that point.
7. SIGN CONVENTION
ln order to compare test results obtained from different crash test facilities, standardized coordinate systems need to be
defined for crash test dumm¡es, vehicle structures and laboratory fixtures. ln addition, recorded polar¡ties for various
transducer outputs need to be defned relative to positive directions of the appropriate coordinate systems. This section
describes the standardized sign convention.
7.1 Right-HandedCoordinateSystem
To assure consistent vector directions of moments and angular velocities and accelerations produced by vector
multiplications all coordinate systems used in vehicle testing will be "ríght-handed".
Ríght-handed coordinate system consists of an orderêd set of three mutually perpendicular axes (x, y, z) which have a
common origin and whose positive directions poínt in the same direct¡ons as the ordered set of the thumb, forefinger and
middle finger of the right hand when positioned as shown in Figure 3. Note that one can choose the positive x-axis to
point in the direction of either the thumb, forefinger or middle finger as shown in the orientations '1, 2, and 3 of Figure 3.
However, once this decision is made then the positive directions of the y- and z-axes must be as indicated by
corresponding orientation shown in Figure 3.
+x +Z +Y
+Y +X +Z
il
/ v I
+Z +Y +X
Orientation 1 Orientation 2 Orientation 3
Positive angular motion and moment directions are determined by the right-handed screw rule. lf a posit¡ve axis is
grasped w¡th the right hand w¡th the thumb extended in the positive direction as shown in Figure 4, then the curl of the
fingers indicate the positive direction for angular motions and moments.
+x
A simple method to determine if a (x, y, z) coordinate system is righþhanded is to rotate the system 90 degrees about the
z-axis using the righthanded screw rule. lf the pos¡tive x-axis rotates to the position previously occup¡ed by the positive
y-axis, then the system is right handed.
For vehicle and laboratory coordinate systems, positive z-axis will be directed downward, positive x-axis will be directed
forward relat¡ve to the vehicle and positive y-axis will be directed away from the vehicle's left to its right (see SAE J670 -
Vehicle Dynamics Term¡nology). For structures within the vehicle that have a principle axis of motion such as the steering
wheel and column, the vehicle coordinate system may be rotated about the y-axis such that the positive x-ax¡s ¡s directed
along the column axis.
Coordinate systems can be afüxed to any point on the dummy. To determine the orientatíon of the coordinate axes, the
dummy will always be considered as standing erect. For this posture, the posltive y-axis will be directed from the
dummy's lefr to itS right side, the positive z-axis will be directed downward from head to toe, and the positive x-axis will be
directed fon¡vard. ln anatomical terminology, the positive x-axis is directed from the posterior to the anterior (P-A), the
positive y-axis is directed form the left to right (L-R), and the positive z-axis is directed from superior to inferior (Sl).
Figure 5 shows examples of th¡s standardized orientation for coordinate systems attached to a few body points. Note that
as the dummy is articulated to sit in a vehicle or if the dummy is arliculated for a test, the coordinate systems rotate with
the¡r respect¡ve dummy parts.
Positive recorded outputs for these lransducers are to be consistent with the positive axes of the coordinate system
defined for the speciflc dummy or vehicle point being measured. For example, a blow to the back of the dummy's head
produces an acceleration in the forward direction (+x) which shall be recorded aç a positive acceleration. A blow to the
top ofthe head produces a +z acceleration. A blowto the left side ofthe head produces a +y acceleration.
For displacements, the coordinate systems of inlerest must be defined. For example, frontal chest compression is the
distance that the sternum moves relative to the thorac¡c spine. ln this case, the coordinate system is fixed to the thoracic
spine. When the sternum moves closer to the spine, its displacement is rean¡rard relative to the spine which is in the
negative x-direct¡on. Hence, the polaíty for frontal chest compression is negative. For lateral chest compression, a blow
to the left side of the chest produces a positive displacement of the impacted ribs relat¡ve to the thoracic spine. However,
a blow to the right side of the chest produces a negative rib displacement. The rearward displacement of the tibia relative
to the femur that is measured by the knee shear transducer is in the negative x-direction. The polarity for this motion shall
be negative.
+Z
For load cells that measure loads applied directly to the dummy or vehicle structure, their recorded output polarities shall
be consistent w¡th the direction ofthe applied external load referenced to the standardized coordinate system at the point
of the load application. For example, load cells that measure shoulder belt loading of the clavicle are designed to
measure Fx and Fz applied to the clavicle. Positive directions for these applied forces would be forward (+x) and
downward (+z) relative to the dummy, respectively. For the BlOSlD, a lateral inward load applied to the crest of the |eft
ilium would be positive (+y), while a lateral inward load applied to the crest of the right ilium would be negative (-y).
Defining recorded output polarities for load cells that measure loads internal to the dummy requires a standardized
dummy sectioning scheme and a definition of what sectioned dummy part is to be loaded in the positive direction since
internal loads occur in pairs of equal magnitudes but opposite directions. The standardized sectioning scheme is
illustrated by the free-body diagram of a cube shown in F¡gure 6. lt is assumed that the load cell of interest is contained
within the cube and responds to loads applied to the surfaces of the cube. Load cell outputs shall be recorded with
positive polarities when normal loads, shear loads, torques or moments are applíed in the positive direction, as defined by
the standardized coordinate system, to the right, front, andior bottom surfaces ofthe cube. These loads are represented
by solid arrows. For static equilibrium, equal magnitude but opposile (negative) direction loads must be applied to the left,
back and/or top surfaces of the cube as indicated by the dashed arrows. For example, upper and lower neck, lumbar
spine, and upper and lower tibia load cells shall have positive recorded outputs when the dummy is sectioned below the
load cell in quesùon and positive loads are applied to the bottom surface of the load cell in question. Free-body diagrams
show¡ng the load systems that produce the required recorded polarities for specific dummy load cells are given in the
appropriate dummy users manual and in SAE lnformation Report J1733 - Sign Convention for Vehicle Crash Testing.
I
ùt
I tl
--Þ Èói¡ S¡ettu of
S, .t etetl¡h!¡ç
t ùqì¡¡S¡¡
I I I
,l Í 't\
I
+l- --þ \
--¡5
¿r
3r 9. 3.
ì't
q/
zÞ
4
\
t'
FIGURE 6 - DIRECÏIONS OF LOADS APPLIED TO FREE-BODY DIAGRAM OF THE SECTIONED
DUMMY PARÏ CONTAINING THE LOAD CELL OF INTEREST (ILLUSTRATED
AS A CUBE) FOR pOStTtVE RECORDED pOLARtTtES
1. Principle axes of load cell are aligned parallel to respective axes of local dummy coordinate system.
2. Bold arrows of normal forces (F), shear forces (S), and moments (M) are shown in positive directions. These Positive
loads are applied to front, right, and bottom surfaces ofthe cube. Dotted arrows indicate d¡rection of loads applied to
the back, left and top sufaces for stat¡c equilibrium.
Table 2 contains descriptions of dummy manipulat¡ons and the corresponding polarities for recorded outputs for some of
the more common load cells used in dummies. Manipulations for checking polarities of load cells not listed infable2
may be found in the users manual for the speclfìc dummy being used or SAE lnformation Report J'1733 - Sign Convention
for Vehicle Crash Testing.
This section establishes guidelines for digital data processing equipment used by crash test¡ng agencies.
8.1 Pre-sampleFiltering
Filtering coresponding to the frequencies of the data channel class may be canied out during processing of data.
Since crash test data may have high-frequency components above the channel class F¡, pre-sample filtering shall be
used to keep these components from causing aliasing errors in the sampling process. Before recording, analog anti-
aliasing filtering, at a level corresponding to CFC 1000 or higher, shall take place. Overall data acquisition system pre-
sample filtering must be such that the total system response shall comply with section 3.7. The maximum enor induced
by aliasing at the
F¡ Hertz).
FH frequency shall not exceed 0.1% of the CAC (more than 60 decibels of attenuation at sample rate -
The minimum acceptable sampling rate is a function of many variables, particularly sophistication of the reconstruction
method used in the processing sofiware. For those installations ut¡liz¡ng only simple reconstruction software, the sample
rate shall be a minimum of ten times the Fn. ln installations with CFC 1000 pre-sample filters, this corresponds to a
m¡ri¡mum sampling rate of no less than10 000 samples per second per channel.
8.3 Resolution
Digital word lengths of at least 12 bits (including sign) shall be used to be assured of reasonable accuracy in processing.
The least significant bit shall conespond to no greater lhan 0.2% of the CAC.
Processing software is typically used to scale and filter data, determine zero levels, perform mathematical operations and
prepare data plot formats.
8.4.'l DigitalFiltering
Phaseless filters are to be used for post-test filtering. Phaseless filters will cause t¡me uncerta¡nty in turn causing
problems in comparing data to film, and comparing data to data if the class filters are different. Filtering shall precede all
nonl¡near operations, such as calculation of resultant vectors and injury indices. Any phaseless filtering algorithm can be
used for CFC filtering as long as the total system response conforms to the data channel performance requirements as
given ¡n Section 4 (The J211-1 March 1995 Appendix C filter algorithm continues to meet this requ¡rement for CFC-60
and CFC-180). An FFT filtering system is available in Appendix C. Compliance to these corridors can be determined by
measuring the attenuation of a known analog input. Alternatively a sum of the data acquisition system, transducer
calibration pass/fail corridors, and the phase-less filter response can be compared to the CFC corridor. The type of digital
filter used shall be reported. The user is cautioned to examine the unfiltered data for signal overloads, since the filtering
process can mask certain overload conditions.
8.4.2 Scaling
A check of the channel gain accuracy shall be applied to each analog input before every test, rather than relying on the
set gain values. Corrective measures shall be required when the results of this check are outs¡de of the requirements of
section 4.6.1.
8.4.3 Zeroing
The zero level of each data channel shall be corrected post-test. This correction shall bring the normal¡zed value of a
stable pre-test section of data to the proper initial value for the transducer.
lnjury index calculations shall use all sampled data points. Details on various injury index calculations can be found in
SAE lnformation Report J'1727 - lnjury Calculation Guidelines.
9. TIMING MARKS
Timing marks are essential in data analysis and correlation of high-speed film to other data channels. Timing frequency
error shall be less than 1% of the chosen or designated frequency. Timing synchronization shall be wilhin *1 ms.
Time of initial contact shall be recorded within 0.4 ms of the closing of a switch actuated by the impact. lt shall also be
recorded ¡n film data using a visual indicator such as strobe lights or timing mark channels. When a sw¡tch is not
applicable, Time of Contact may be established through post-test correction or re-setting the T0 mark to coincide with first
sample to exceed a predetermined value (for example, 0.5 Gs) when processed with an appropriate phaseless CFC filter.
The established Time Zero (if different from the initial Time of Contact) shall be reflected in all digital systems, with no
more than 0.4 ms difference between any two systems, and noted in all analog data. Non-analysis cameras shall be
exempt.
ln reporting results of tests, the following information shall be provided w¡th data tabulations, time history traces, etc.:
For purposes of digital data exchange it is recommended that the ISO or NHTSA Format for Data Exchange be followed.
The ISO multimedia data exchange format is defined by the Technical Specification ISO/TS 13499:2003. This
publication is available through an order to ISO/CS (lnternational Standards Organization) or the National ISO Member
ANSI. lhe documentation for the NHTSA format is available through the Office of Crashworthiness Research (NHTSA).
The title of the documentation is - NHTSA Version 5 Test Reference Guides.
The change bar (l) located in the lefi margln is for the convenience of the user in locatlng areas lvlrere technical revislons
have been made to the prevlous lssue of the report. An (R) symbol to the lefr of the document title lndicates a complete
revlslon of the report.
Resldual crush ls speclfed by one or more single-valued data points, w¡th respect to designated reference points.
Determination of res¡dual crush shall be with an enor of less than 5% of the actual crush.
Maximum dynamic crush is a measurement of the maximum deformation of the test speclmen during the impact. This is
also measured with respect to one or more designated reference points. Contingent on the size of the speclmen and the
magnitude of the expected dynamic crush, the following are possible measurement methods:
Displacements relative to designated reference points on the vehicle can be measured by various techniques. The
coordinate system in which displacement,is measured shall be indicated. Determination of steering column displacement
shall be accurate to t0.5 in. (t1.27 cm).
Transducer selection is a primary concern in the process of selecting a data channel configuration. Furthermore, the
variety of transducers, within a transducer category, i.e., acceleration, pressure, force, displacemenl, etc., is ever
widening, yielding concern regarding equivalence and/or correlation of various lransducer performances for any specific
application. As the transducer forms an integral part of the data channel, the enor contribution of transducer related
factors shall be included when considering overall data channel performance requirements as outlined in Section 4.
The purpose of SAE J211 is to recommend "techniques of measurement used in ¡mpact tests" which '\¡rill provide a basis
for meaningful comparison of test results from different sources." This is the goal of establishing equivalency of
performance for various transducer types. Such a goal, considered relative to a transducer category, requires thorough
knowledge of the measurement object¡ve and ¡ts environment. lt is not necessar¡ly limited by technological constraints of
transducer fabrication or design but is unmistakably linked to both the transient and steady responses under ¡mpact
loading conditions.
Both steady-state and lransient transducer responses depend on the combined performance of the transducer
characteristics or specifcations. lnteraction of transducer characterislics can precipitate signmcant differences in both
steady-state and transient responses between transducers with similar spec¡fications. The user must ascertain combined
performance to establish equivalence.
Potential sources of transducer errors are related to data channel compatibility (i.e., lransducer may adversely interact
with the remainder of the data channel upstream of the anti-aliasing filter), e.9., resonance/damping influences. Other
sources of errors are related to environmental compatibility, i.e., non-measurement influences and interactions, which
encompass temperature drifr, magnetic and electric fields effects, elc., and transducer size, weight, seismic mass
location, transverse sensitivity, gage creep, etc. As such, the user must ascertain data channel and environmental
compatibility when establishÍng equivalence.
To establish equivalence, it is the use/s responsibility to perform the tests necessary to ensure that the transducers under
consideration yield similar results for the application of interest, within the error tolerances recommended by SAE J211 for
the whole data channel.
lf feasible, to verifr equ¡valence, transducers may be used back-to-back and/or side.by-side in actual test condilions for a
given measurement application. A statistically significant number of tests shall be performed to validate the results. This
method also may be used if there are no calibration methods to validate the dynamic response of the lransducer type in
question.
An SAE report, J2570, titled "Performance Specifications for Anthropomorphic Test Device Transducers", addresses
issues of transducer equivalency. The purpose of the report is to establish guidelines in comparison and selection of
various transducer types.
To veri! compl¡ance with SAE J211 'Total system response" requirements four sources of attenuation/distortion must be
summed and compared to an acceptable limit. The limit is provided by the SAE J2'11 CFC system corridors, the sources
of distortion include the data acquisition system, the transducer, the post-test filter algorithm and other sources such as
noise induced into the cabling. ln lhese routines an allowance of 0.1 db is provided as a default for all cables between the
transducer and the data acquisition system (CBL and CBU). The testing agency shall veriff their cables meet this value
or modify CBL and CBU if necessary. The total distortion from the data acquisition system and transducer are bounded
by their individual calibration pass/fail corridors. Although exact values could be used, the pass/fail corridors provide a
worst case scenario for system response. A CFC compliant filter must provide sufücient attenuation to keep the "sum of
all attenuation/distort¡on'within the SAE J211 CFC system conidor.
ln general these routines shall not need to be modified. However, if modifcation to the routines is necessary they shall be
made using good engineering judgment, and aimed at meeting the SAE J211 goal of improving compatib¡l¡ty between test
sites. For example, if a data acquisition system has a low frequenry antÈaliasing filter it may fall below the lower
CFC-1000 boundary. ln such a case the post-test fìlter would be required to add gain at specific frequencies (which the
routines cunently do not allow), to keep the system within the SAE J21 1 CFC corridor. Such a change would require a
careful consideration of the afiects on noise and distortion, and verif¡cation that none of the SAE J211 parameters are
violated.
These routines use the Fast Fourier Transform (FFT) defined by the FFT(v) function from MathCad 8, described as:
where
The length of the TD vectorfrom this function will be 0.5.L+1. Some commercial Fast Fourier Transform functions will
result ¡n TD being the same length as L. Within these, the second half oi TD vector will be a refleclion of the first half.
The use of such a FFT function will require some modifications to the routines supplied herein.
Definitions:
CBL The Cable Boundary Lower, the negative lower limit of the attenuation due to sources other than the data
acquis¡t¡on system, transducer, and filter algorithm, such as cabling (default to a max. of -0.1).
CBU The Cable Boundary Upper, the posit¡ve upper limit of the attenuation due to sources other than the data
acquisition system, transducer, and fìlter algorithm, such as cabling (default to a min. of 0.1).
coR A 7 by 16 matrix which defines the nodal points of the CFC corridors.
DAS The data acquisition system. This describes a set ofparameters used to collect a data channel for the purpose of
designing a custom filterforthat channel.
DFZ The delta value for the reference frequency vector, equal to SR/L.
DLA A vector of the attenuations of the nodes in the data acquisition systems lower pass/fail boundary
DLF A vector of the frequencies of the nodes in the data acquisition systems lower pass/fail boundary.
DUA A vector of the attenuations of the nodes in the data acquisition systems upper pass/fail boundary.
DUF A vector of the frequencies of the nodes ¡n the data acquisition systems upper pass/fail boundary.
FM A matrix containing the FZ vector and each of the CFC filter vectors (as the ratio of attenuation).
FV A vector which provides the custom filter's ratio of attenuation at each of the FZ frequencies (one column from the
FM matrix).
FZ A vector of equally spaced reference frequencies, with a delta value of DFZ (conesponding to the FFT
transformed data vector).
FZH The greatest FFT reference frequency value, equal to half the DAS Sample Rate (SR).
L Length of the data vector, with padding (a power of two, at least 15% longer than LDO).
Create the COR matr¡x us¡ng the following values (4 columns per filter corridor).
0.1 0.5 0.1 -0.5 0.1 0.5 0.'l -0.5 0.1 0.5 0.1 -0.5 0.1 0.5 0.1 -0.5
60 0.5 60 -1 180 0.5 180 -1 600 0.5 600 -1 1 000 0.5 1 000 -1
75 -0.3 75 -1.8 225 -0.3 225 -1.8 1200 0.5 1 000 -4 2000 0.5 '1650 4
100 -1.8 100 -3.8 300 -1.8 300 -3.8 38ô5 -40 2119 -30 6442 -40 3496 -30
130 -5.2 130 -8.2 390 -5.2 390 -8.2 FZH -40 2120 -9000 FZH -40 3497 -9000
160 -9.2 160 -13.2 480 -9.2 480 -13.2 0 0 0 0 0 0 0 0
452 -40 452 -48.3 1310 -40 131 0 -48.3 0 0 0 0 0 0 0 0
FZH -40 453 -9000 FZH -40 131 | -9000 0 0 0 0 0 0 0 0
FZH is the greatest FFT reference frequency value, equal to the DAS Sample Rate (SR) divided by 2. To add additional
filters there are two expected elements, which must be followed:
2. lf the lower boundary drops to minus infinity, set the attenuation to, at most -8000, the routines wÍll recognize this and
treat ¡t as minus infinity.
Use the CORs routine to expand the SAE J211 corridors, referenced to the FZ frequencies. SCS = CORs(DFZ,LF,COR),
to generate an expanded J21 1 corridor matrix using the CORs matrix. The first column ¡s the FZ vector. The following
eight columns will provide the upper and lower expanded boundaries of the CFC corridors, in the order they appear in the
CORs matrix. Using the same format, create vectors for the data acquisition system and transducer pass/fail nodes.
Example: An accelerometer with a +2o/o ¡o -2o/o lo 2000n2 and +5o/o to -5olo to 4000 hz transducer pass/fail corridor
would be as follows:
Use the FILL routine to expand the four boundary vectors, referencing them to the FZ frequencies.
CBL and CBU are values that represent possible distortion due to cabling. Set these CBU = 0.1 and CBL = -0.1 as a
default. lf other sources of distortion are suspected, increase the CBL and CBU values to account for the distortion.
Repeat for all combinations of CFC, transducer, and DAS pass/fail corridors, which appear in the processed data.
The CENTERM routine can be used to generate filter response curves, which center the data acquisition system pass/fail
con¡dor in the SAE J21 1 system corridor. These shall be the target response for all filter routines.
The CENTERM routine can be used to derive the CFC filters directly, when filtering in the frequency domain. To apply
this method the data must be converted into the frequency domain using a Fourier Transform, multiplied by an attenuat¡on
ratio (the f¡lter vector), and converted back to the time domain. This method allows the CFC filter vectors to be calculated
once and stored. The fìlter does nol need to be re-calculated each time it is used, as time based filters are. For this
method follow lhe procedure below.
Choose a value for L if a longer than the min¡mum value ¡s desired (L is equal to 2n, where 2n this is at least 15% longer
than the raw data length).
Use the PREPARE routine to convert the data to the frequency domain. The FFT(x) and IFFT(x) elements are a Fast
Fourier Iransform, and its inverse (the FFT(x) function will provide a complex vector as a result). Multiple each element
of the frequency domain data vector by the conesponding element of the filter vector (one column from the FM matrix).
Use the REMOVE routine to convert the frequency domain and fìltered vector to lhe time domain.
The TC_COR routine is provided for your convenience. lt calculates a maximum transducer pass/fail corridor,
considering all of the FM filters and SAE J21 1 conidors, when only the DAS pass/fail corridor and CBU and CBL values
are provided.
For more insight into frequency domain data processing see the SAE paper #2002-01-0796 (the routines in SAE J211
represent an upgraded to the routines in this paper).
CENTERM(COR,LF,DU,DL,DFZ): a function to make a matrix of the customized FFT filters that centers the data
acquisition system pass/fail corridor in the SAE J211 CFC corridors. Beyond the lower CFC boundary drop off frequency
(DOl) DB attempts to continue the total system response al -Q4 db per octave. Th¡s function does not allow the filter to
add any gain. To generate a plot showing the total system response see the TSR routine.
SCS = call CORs(DFZ,LF,COR) Call the CORs subroutine to generate the system response conidors.
forr=0toLF-1
if DL(r) >= -7ggg CD = defines the center of the data acquisition system pass/fail corridor.
then cD(r) = 0.5 . (DU(r)+ DL(r)) beyond the CFC corridor drop off it continues at the previous slope.
else CD(r) = CD(r - 2) + (CD(r - 1) - CD(r - 2)) . log(r / (r - 2)) / log((r -1) I (r - 2))
endif
FM(r,O) = SCS(r,O) Set the first column of FM equal to the FZ vector (SCS first column).
next r
lor c= 2 to cols(SCS) - 1 step 2 For loop in c by 2's from 2Îhru the columns in the SCS matrix
CC = int(COR(1 ,c' 2 - 2) I DFZ) CC = the index of FZ vector corresponding to the CFC frequency
CN = int(0.7 .
CC) CN = the index of FZ at 70% of the CC frequency.
DOI =LF-1 lnitializeDOl
while SCS(DOI,c) < {0 find the index of where the lower CFC boundary drops below40 db
DOI = DOI -'t
Wend
forr=0toCN
if CD(r) < 0
then DB(r) = Q Do not allow DB(r) filter to add gain'
else DB(r) = -1 - CD(r) Have the system provide no attenuation below 90% of F¡.
endif
next r
for r = CC + 1 to DOI For loop ¡n r (row) from CC plus one to DOI
DB(r)=¡.5*(SCS(r,c-1)+SCS(r,c))-CD(r) DB=attenuat¡onneededtocenterCDintheSAEJ2ll
corridor between CC+1 and DOI fequency indexes.
¡/F = SCS(r,c) + 0.2 - DL(r) MF = the maximum allowable amount of attenuation at this frequency.
if DB(r) < MF then DB(r) = ¡¡p Do not allow DB(r) filter to be less then MF
¡f DB(r) > 0 then DB(r) = g Do not allow DB(r) filter to add gain
next r
for r = CN+1 to CC DB = altenuation needed to center CD in the SAE J21 1 corridor.
DB(r) = ¡s1ç¡, + (DB(CC + 1)-DB(CN) - log(r/CN ) /log((cC + 1) /cN )
lf DB(r) > 0 then DB(r) = $ Do not allow DB(r) filter to add gain
next r
MTU = SCS(DOI,c- 1)- DB(DOI) - DU(DOI) MTU = the gap between DU and the upper CFC corridor boundary
at the DOI frequency, used as a minimum gap size.
if MTU > 2 then MTU = 2 Limil MTU (the minimum gap size) to a maximum of 2 db.
for r = DOI + 1 to LF - 1 For loop from the DU drop off frequency to FZH.
DB(r) = 991¡tl) + CD(DOI) - CD(r) - 24 - log(r / DOI) I log ( 2 )
Cause the upper data acquisition system
passifail boundary plus DB to continue aI-24 db per octave.
if 1> 2'CN then
MF = SCS(r,c - 1) - MTU - DU(r) MF = the maximum allowable amount of attenuation at this frequency
to meet MTU. This keeps the upper transducer pass/fail corridor open.
¡f DB(r) > MF then DB(r) = ¡¡p Do not allow the transducer upper boundary to be less than MF.
endif
¡f DB(r) > 0 then DB(r) = a Do not allow DB(r) filter to add gain
next r
for r = 0 to LF - 1 Convert the filter vector, DB into the ratio of attenuation and stores it.
FM(r,c " 0.5)= '¡9n1o.ou - DB(r)) as a column in the filter matrix, FM
next r
next c
(return the FM matrix)
GORs(DFZ,LF,COR): a funct¡on to expand the SAE J21'l CFC corridors defined by the COR matrix.
coR A matrix defining the nodal points of the SAE J211 CFC corridors.
DFZ Delta FZ the step size in the FZ vector (pFZ = SR / L).
LF The length of the frequency domain data vector.
CO=cols(COR)-0.5-1 CO = the number of CFC boundaries (upper and lower), minus o¡s = (the
"colsQ" function returns the number of columns in the COR matrix).
ROW=rows(COR)-1 ROW = the number of rows in the COR matrix, minus one
(the "rowsfl" function returns the number of rows in the COR matrix).
forc=0toCO Loop through each of the boundary corridors
F=0 Clear the frequency vector.
A=0 Clear the Amplitude vector.
E=0 Clear the expanded boundary vector.
forr=0toROW
F(t)=6gPt''t-" copy the F vector from the COR matr¡x
A(r)=CAP,t't.'*'t, copy the A vector from the COR matrix
next r
E = call FILL(DFZ,LF,F,A, use the FILL routine to generate the expanded boundary vector
forr=0toLF-1
S(r,c+l)=E(r) copy the E vector into the S matr¡x
ne)d r
next c
forr=0toLF-1
S(r'0)=riDFZ make the first column of S equal to the FZ vector,
next r
(return the S matrix)
FILL(DFZ,LF,F,A): a function Used to expand a table of nodal points into a vector, with a length of LF and referenced to
the FZ frequencies. lf the attenuation of the last nodal point drops to minus infinity, set the value to at most -10000. lf the
boundary becomes horizontal, at least two nodes must be defined.
LA= LA-1
wend
forr=1toLA
if F(r) < F(r -1) + DFZ then F(r) = p1t - 1) + DFZ spread out frequencies which are less than DFZ aparl.
next r
for r = 0 to LA For loop in r to determine the values of lN.
lN(r) = ¡nt(F(r) / DFZ) lN is a vector of the FZ index numbers that correspond to the F vector
frequencies.
if A(r) < -7000 then lN(r) = lN(r - 1) + 'l Def¡ne lN for nodes beyond the point the aüenuation drops to - infinity.
¡f lN(r) > LF- 1 then lN(r) = ¡¡- 1 Do notallow lN to begreaterthan LF- 1.
next r
forr=0tolN(0)
V(r) = A(0) For frequencies below F(0) set V equal to A(0)
next r
ford=0tol-A-1
¡flN(d)+2<LFthen
@ 201 6 by the American Association of SLate Highway and Transportation Olïc ials
All rights reserved Duplication is a violation of applicable law.
198 | Manual forAssessingSafety Hardware
FILTER(FV,DO): A function to apply a FFT filter to the raw data and return time data in the time domain. This tunction
assumes the FV vector is the 'ratio of attenuation", this is one column from the output of the CENTERM routine.
FD = call PREPARE((length(F\4 ' 2 - 2),DO) Call the PREPARE lunction to convert DO to the frequency domain.
forr= 0 to length(FD) - 1 forloop in rfom 0 to the numberof elements in FD, minus one.
FFDG) = FD(r) ' FV(r) Filter the frequenry domain data by multiplying by the ratio of attenuation.
next r
TD = call REMOVE(length(DO),FFD) Call the REMOVE function to convert FFD to the time domain.
(return the TD vector)
PREPARE(L,DO): A function to add padding to each end of the raw data vector, br¡ng its length a power of 2, and convert
it to the fequency domain.
LDO = length(DO) LDO = the number of samples in the raw data vector.
if L > LDO . 1 .15 Veriff the suggested power of two length is large enough.
then B = L - LDO Set the value of B
else B = 2^[1+ int(log(LDO . 1.15) / log(2)] - LDO lf L is too small, determine the minimum value for L and set B
end if
La = LDO- 1
LB = int(B - 0.5) Set LB, the post-data bridge length, at half the total bridge length.
UB = B - LB Set UB, the pre-data bridge length, at the remainder of the bridge length.
For r = 1 to UB -
For loop in c through the post-data bridge section.
D(UB- r) = (2 " DO(O) - DO(r)) (UB- r) / UB
The pre-data bridge section is a diminishing mirror of the data.
next r
for r = 0 to La For loop in r (row) through the raw data vector.
D(r + UB) = DO(r) Copy the raw data ínto the D vector.
next r
for r = 1 to LB For loop in r through the length ofthe post data bridge.
D(r+ La + UB)= (2 - DO(La)-DO(La-r)) - (LB-r) / LB
Add the posttestbridge section (a decaying mirror).
next r
FD = FFT(D) Transform the padded data vector D to the frequency domain.
(return the FD vector)
REMOVE(LDO,FD): a function to convert a frequency domain data vector to the time domain, and remove the bridge
section.
TC_COR(SCS,FM,DU,DL,CBU,CBL): a function to generate the maximal transducer pass/fail corridor for a transducer
that may be used with all CFC corridors listed in the CORs matrix. The user may wish to choose nodal points which fall
within these max¡mums, to define their transducer pass/fail corridor.
CBL An allowance for other sources of signal loss (safety margin) (example: use -0.1 for connector losses).
CBU An allowance for other sources of signal gain (safety margin) (typically equal to -1 ' CBL for symmetry)
DL The lower data acquisition system pass/ fail boundary.
DU The upper data acquisition syslem pass/ fail boundary.
FM A matrix conta¡n¡ng the FZ vector and the individual filter vectors (ratio of attenuation).
SCS A matrix containing the FZ vector and the individual SAE J211 system upper and lower boundaries.
for r = 0 to length(DU) - 1 For loop in r (row) over the length of the filter vector.
for c = 1 to cols(FM) - 1 For loop in c (column) through the SCS filter columns.
A(c - 1) = SCS(r,c ' 2- 1\- 20 - log(FM(r,c)) - OU(r) - CBU Calculate the upper transducer pass/fail
boundary.
E(c - 1) = SCS(r,c ' 2)- 20. log(FM(r,c)) - DL(r) - CBL Calculate the lower transducer pass/fail boundary.
if DL(r)<-60thenE(c-1)=-1'A(c-1) MakeTCsymmetrical if DLdropsoff (outputis<0.1%of input)
if SCS(r,c-2).-60thenE(c-1)=-1"4(c-1) MakeTCsymmetrical aftertheJ2lllowerboundary
next c drops off (allows output to be < 0.1 % of input).
TC(r,O) = ¡¡¡¡14¡ The minimum value in the A vector (upper boundaries)
TC(r,1) = ¡¿¡1E¡ The maximum value in the E vector (lower boundaries)
next r
(return the TC vector)
TSR(FM,DU,DL,TU,TL): a routine for generating the Total System Response vector, of the FM filters, which can be
graphed against FZ. This uses the custom filters, data acqu¡sition system pass/fail conidor, and transducer pass/fail
corridor.
for r = 0 to rows(FM) - 1 for loop in r from zero to the number of rows in the FM matrix, minus one
CD(r) = 9.5 . (DU(r) + DL(r) + TU(r) + TL(r)) CD represents the center of the combined data acquis¡tion and
next r lransducer passifail corridors.
for c = I to cols(FM) - 1 for loop in c from one to the number of columns in the FM matrix, minus one
for r = 0 to rows(FM) - 1 for loop in r from zero to the number of rows in the FM matrix, minus one
0
if FM(r,c) > Verify the FM element is not equal to zero to avoid a log(0) error.
then TR(r,c-1) = 20 - log(FM(r,c)) + CD(r) convert the FM element to db and add the CD attenuation.
else TR(r,c-1) = - 9000 if the FM element = 0 set to default value of -9000.
end if
next r
next c
(return the TR matrix)
REAF.
1. Scope-The purpose of this SAE Recommended Practice is to define criteria of pertormance tor an optical
data channel when numerical time and space data are taken from lhe images to analyze impact test resulls.
2. Reletences
2.1 Appllcable Publicatlons-The following publications form a part of this document to the extent specified
herein. The latest issue of all publications shall apply. ln the event of a conflict between the text of th¡s
document and references cited herein, the lext of this document takes precedence.
2.1.1 ISO PuBLtcArtoNs-Available from ANSI, 11 West 42nd Street, New York, NY 10036-8002.
3. Deîinitlons
3.1 Analysfs System-A system for measuring and collecting the coordinates ol image points as a function of
time and converting lhem into coordinates in object space.
3.2 Average Distortion Error-A quality parameter of the optical data channel that measures symmetrícal image
distortion independent of any noise and misalignment present.
3.3 Data Collectlon Error-A quality parameter of the optical data channel that measures the precision with
which the analysis system can locate coordinates of image points. This is often referred to as the "noise" in
the data collection process.
3.4 Horizontal Field of Vlew (HFOV)-The longer of the two sides of a camera's image,
3.5 lmaging Rate-The frequency of renewal of information for a given point expressed in renewals per second.
3,6 Misalignment Error-A quality parameter of the optical data channel that measures the image distortion due
to any camera-to-target pattern misalignment that may have existed when the target pattern was imaged, and
any lens de-centering.
3.7 Optical Axis-A straight line which passes through the centers of curvature of lhe camera lens surfaces. ln a
peÍectly aligned camera system, this line would interàect the image plane at the center of the ¡mage.
SAE Technical Standards Board Rules provide that'Th¡s r€port is published by SAE to êdvance lhe stale oltechnical and eng¡neering sc¡€nces. The use of this report ls enllrely
voluntary, and lts applicab¡llty and su¡lab¡llly lorany partlcular use, lncludlng any patenl ¡nlringemenl añsing lherefrcm, ls the sole responslbil¡ty ot lho user.'
SAE reviews each technical report at leasl every f¡vê years at wh¡ch lim€ it may be reaffm€d, rev¡sed, or canælled. SAE ¡nv¡les your wrinen comments and suggestions.
Copyrighl @ 2008 SAE lnlernat¡onal
All rights reswed. No part of lh¡s publ¡cal¡on may be r€produced, slored in a reùioval system or lfansm¡tted, in any fom or by iy means, electronic, mechanlcal, photocopying,
recording, or othem¡so, w¡thout lhe priorwrinen pem¡ss¡on olSAE.
Reprinted with permission from SAE J21112@I2OOB SAE lnternational. Additional copies
may be purchased directly from SAE at http://sae.org or call (877) 606-7323.
3.8 Optical Data Channel-A system composed of an image taking device (lor example, camera and lenses), a
record¡ng medium for these images (film, disc, magnetic tape), an optical path (for example, fiber-optic cable)
and a system for analyzing the images, including any analys¡s procedure.
3,9 Overall Error-An overall quality parameter of the optical data channel that quantil¡es image distortion, noise
and misalignment present in the system.
3.10 Tlme Base System-A system to enable the determination of the time interval between any two recorded
events.
3.11 TimeOrlg¡nldentif¡catlopA method for identifying and recording the instant chosen as the time origin,
usually the beginning ol impacl or other defined event.
3.12 Valldation Target Palr-A pair ot targels placed in the field of view such that the d¡stance separat¡ng them
remains constant, and that they are both visible throughout the impact lest.
4.1 Optlca! Performanc+The optical performance is determined by analyzíng the variation of apparent
distances between targeté on the Target Pattern Drawing and the aclual distances according to the procedure
described in the Appendix A, or in an equivalent manner. These performance parameters provide the position
errors in terms of a percentage of the HFOV. The error expressed in engineering units may then be calculated
if the width of the area to be imaged during the test is known.
4.1.1 OVEBALL ERRoR-This parameler is uselul in determining lhe maximum error present, regardless ol the
source. Each standard or specification that references this document should establish an upper limit for this
parameter that is appropriate for the intended application. For reference, the maximum error produced by a
typical high quality system is less than 0.25%.
4.1.2 DATA CoLLEcTtoN EFBoF-This parameter quantifies the random errors that occur when determining the
centers of the targets. This parameter is also useful for comparing the accuracy potential of various lens/
camera"/analysis system combinations. Since the procedure used to derive this parameter is also sensitive
to any camera to target misalignment, careful setup is required in order to make an accurate assessment.
4.1.3 AVERAGE DtSToFTtoN EBRoB-Th¡s parameter quantifies the error due to symmetr¡c distortion, independent
of other sources. This can be very useful in evaluating those systems that have a high Data Collection or
Misalignment Enor.
4.1.4 M|SAL|GNMENT EBBoR-This parameler is useful in determining the source of a larger than expected Overall
Error. A large misalignment error indicates lhat e¡ther the camera axis was nol perpendicular to the Target
plane, or there ¡s a misalignmenl between the lens elements and the image plane (caused by a
manufacturing defect in one or both). Since this parameter is also affected to some degree by the Data
Collection Error, it should be compared with Ít to determine the dominant factor.
4.2 Time Bas+A time base is required. lt shall perm¡t the determination of the time between recorded events
with an error of less than 10% of the reciprocal ol the imaging øle ot O.2"/o of the actual time, whichever is
greater,
4.3 Tlme Origin ldentlficatlon-The accuracy of the method shall be at least equal to the value expressed in
seconds of the reciprocal of the imaging rate.
4.4 lmaglng Rate-The imag¡ng rate shall be left to the user's init¡ative, taking into account the intent of the
imaging coverage and the following main four factors:
a. Aliasing
b. The limitations due to the equipment (blur, etc.)
c. The need to combine data from several image taking devices and from electronic recordings of the
impact test.
d. Expected object motion frequencies of interest
4.5.1 A lenglh calibration shall be performed which perm¡ts the determination of lengths within any requirements
set for their accuracy. Each standard or specification that references th¡s document should establish an
accuracy requirement that is appropriate for the intended application.
4.5.2 The accuracy of length calculations may be assessed for each impact test by comparing the determ¡ned
distance between two points with the actual distance between them. This validation target pair should be
separated by at least 1/6 of the horizontal field of view, be fixed with respect to one another, and not be lhe
only points used for length cal¡brat¡on. When viewed by cameras that do not travel with the test sublect,
target pa¡rs should be placed so lhat they are oriented perpendicular to the d¡rection of motion, travêrse the
field of view during th€ impact test, and rEmain visible in all frames of interest, When tracked throughout the
test, the determined distance between these targets should remain constant and be equal to the actual
distance.
5. Image File Formaf-To permit universal access to high-speed images for analysis, archiving, and
interchange, image files should be producible that conform to the requirements of ISO/TC 22lSC1?/.NGs
Multimedia Data Exchange Format for lmpact Têsts.
6. wotes
6.1 Marglnal lndlclæThe (R) is for the convenience of the user in localing areas where technical revisions have
been made to the prev¡ous issue of the report. ll the symbol is nelrt to the report litle, it indicates a complete
rev¡sion of the report.
6.2 Key Words-lnstrumentation, photographic, impact test, distortion, error, accuracy, f¡le format.
APPEND¡X A
A.1 Test Equipmenl-A Target Paltern Drawing shall be used that is of lhe appropriate shape for the imager
being tested and has dimensions in conformity with the spec¡f¡cat¡ons:
The drawing consists of a target at the center of the FOV, four groups of 3 targets located along the outermost
portion of each diagonal, and a centered c¡rcle of eight equally spaced targets. The narrowest dimension of
the imaged area (Target Height) shall be 750 mm. The other dimension (TaÌget Width) shall be as requi¡ed to
fit the aspect ratio of the camera image. Frame edge markers (25 mm tall, filled triangles) are placed at the
mid' point ol each imaged area s¡des. The circle radius and th6 spacing of the target marks along each
'diagonal shall be calculated using Equat¡on 41, then rounded to the nearest mill¡meter.
The heavy border shall be 5 mm wide with its centerlíne 20 mm beyond the imaged area. The targets used
shall be 25 mm in diameter.
Each of the circle targets shall be rotated such that its primary ax¡s is tangent to the circle.
Drawings meeting these specifications for 4/3 or square aspect ratios are shown in Figures A1 and A2. CAD
files of the two drawings are available from SAE. Other targets, as required for specific analysis sottware or
procedures, may be substituted for those shown in the Figures. This drawing must be pr¡nted on a stable
based material and bonded to a flat surface to avoid dimensional errors. lf desired, a sub-scale size of the
drawing may be used.
4.2 Tesl Procedure-The drawing shall be exposed "full lrame" by the camera forming one of the elements of the
data channel to be tested. Framing shall be such that the image shows the frame edge markers and all of the
targets but not the heavy border. Also, the camera must be aligned such that the optical axis is perpendicular
to the drawing plane.
lf film or video lape is used to store images for evaluation, it shall be of the same type and quality as that used
lor impact tèsting.
The coordinates shall be measured with the same analysis equipment and procedures that are used for impact
testing.
o
\t
o
i
II
.t
+q
soît
o
c ,D
ü
t,'
(o
I (f,
o
ô¡
o" o1
ro
i* 180
c 270
+.1
I o-
tl
- 395 - NOTE: All dimensons i 0.5 mm
790
-)
790
395 NOTE: All dlmensons t 0.5 mm
353.5s
o"
o
o)
¡- e,
a
8., \)e¿
) o
o
Q,, ct
\ tt,
ll-
e^' e,,
750.00 -
o_
.o^, ,S",
\7
-6-
_)
4.3 Analysisloreach diagonal, i= 1to 4, label the 3targets j= 1to 3,from the centeroutward. Forthe circle, i
4,3.1 Data Collection Error-The measured diameters (d¡¡) of the circle shall be calculated using Equation A2:
where:
i=5
i= 1,2,3,4
The mean diameter (D) and the standard deviation (S) shall be calculated taking the 4 d¡ameter values (d¡) into
account per Equations Ag and A4:
(Eq. A3)
(Eq. Aa)
'= {[,=-;,.,,-'"]"]
Calculate the scale factor (F) per Equation A5:
where:
The Data Collection Error, expressed as a percentage of the Horizontal Field of View, is (2S)/(F"Target Width)
multiplied by 100%.
A.3,2 Overall Error-The actual location ol each target (a¡, b¡) with respect to the center target, and the measured
location of each target (x¡¡, y¡¡) corresponding to the Target Pattern Drawing imaged shall be used in Equations
A6 and 47. Calculate the measured radial distances (r¡) and the actual scaled radial distances (R¡¡) per
Equations A6 and A7:
where:
i = 1,2,3,4
j=1,2,3
7-
The maximum enor (E), calculated per Equation A8 is the maximum difference between the measured radial
distance (r) and actual scaled radial distance (R) from the center target to each target.
where:
i = 1 ,2,3,4
j- 1,2,3
The Overall Eror, expressed as a percenlage of the Horizontal Field of View, is U(F'Target Width) multiplied
by'100%.
A.3,3 Average Dlstortlon Error-The average radial error (A¡) for each group of symmetrically dispersed targets j =
1 to 3, is calculated using Equation A9:
4= | {r',- n',) 4
l,
(Eq. Ae)
1
where:
i = 1,2,3,4
A = MAXTABS(Ai)l (Eq.410)
where:
i= 1'2'3
The Average Dlstortion Error, expressed as a percentage of the Horizontal Field of Víew, is A/(F'Target Width)
multiplied by 100%.
A.3.4MisalignmentError-The maximum misalignment enor (M), calculated per Equation A1 1, is the maximum
enor (E) less the maximum average error (A):
M=E-A (Eq. Al l)
The Misalignment Error, expressed as a percentage of the Horizontal Field of View, is M/(F'Target Width)
multiplied by 100%.
4,4 Alternate Procedure4ther target pattern drawings w¡th ditferent sizes or patterns of targets may be used, in
which case the user has to delermine the accuracy in an indirect way,
There is then also a need to show that the indirect method gives a result equ¡valent to the use of the pattern
specified in this appendix.
Development of an alteinate procedure may be advantageous if a specific pattern of largets is employed for
the delermination ol lens corrections, which is needed lor some wide-angle lenses.
Rational*The committee reviewed this Recommended Practice to determine what changes were necessary
to encompass the high-speed digital imagers that were replacing film cameras in ¡mpact tests. Several
improvements were developed, many of which are applicable to both video and film cameras. The major
changes are outlined as follows:
Optical Performanc+A new target pattern was developed that assesses the peÉormance of an
analysis system over lhe entire field of view. The previous pattern did not address the comers of the
image where lhe maximum distortion is usually present. This target pattern can also be adapted for
imagers with other than the 4/3 aspecl ratio of 16mm film.
A new procedure lor determining optical performance was developed that directly compares the
dimensions measured from the image with the actual dimensions of the target pattern drawing. Th¡s
procedure derives positional errors that are then expressed as a percentage of the hodzontal lield of
view to allow users to easily determine the maximum error that would be present in the¡r expeÍments.
The index number produced by the previous procedure d¡d not d¡rectly correlate to a positional error.
This new procedure only considers the radial component of distortion since both optical theory and
practical experience indicates that the tangential component is negligible for the types of lenses used in
impact tests. The new parameters calculated are as follows:
The Data Collection error is lwice the standard deviation of the 4 center circle diameters. This parameter
quantifies the analysis syslem's level ot precision in locating the centers of the targets. lt can be thought
of as the effective resolution of the system.
The Overall Error is the maximum diffe¡ence between lhe measured radial location of each of the
diagonal targets and their actual location. This parameter provides the maximum amount displacement
seen anywhere in the FOV, regardless of ¡ts source, i.e., distort¡on, de-centering, or data collection
errors. For wide-angle lenses, this point is always at one of the corner targêts. For telephoto lenses, this
poinl could be at any of the diagonal targets.
The Average Distortion Error is the maximum d¡fference between the average measured radial distance
for each of the three symmetrical groups of diagonal targets and the actual distance for each group. This
averaging eliminates any non-symmetrical d¡stortion effects, and reduces the data collection errors to
allow accurate measurement of only the radial distortion. The points that exhibit the maximum value will
still be the same as for the Overall Error.
The Misalignment Error is the Overall Error less the Average Distortion Error. This difference represents
the amount of error that is due to lens de-centering (a manulacturing defect) or improper alignment of the
target pattern drawing during the evalualion test.
Length Determination-The general guidance on this subject has been supplemented with a
procedure for determining the accuracy of each impact test's analysis. The procedure compares the
determined distance between two points w¡th the actual distance between them throughout the test.
Pass/Fail CriteriæThe limits on the distortion and length determination errors have been replaced w¡th
a recommendation that each standard or specification thal references this document set a limit that is
appropriate for the intended application.
lmage File Format-A requirement that image files be producible that meet the latest ISO standard was
added to permit universal access to high-speed images for analysis, archiving, and interchange.
Applicatlon-The purpose of this SAE Recommended Practice is to define criteria of performance for an
optical data channel when numerical time and space data are taken hom the images to analyze impact
test results. The requirements are lo facilitate comparisons between results obtained by diflerent
laboratories
Reference Sectlon
@ 2016 by the Anrerican Association olState Highway and Transportat ion Olïcials.
All rights reserved, Duplication is a violation ofapplicable law
210 | Manual for Assessing Safety Hardware
SAE J2l1 should be used as the primary reference for instrumentation specifications. Part I relates to
electronic instrumentation and Part 2 relates to photographic instrumentation. Part I of the specifica-
tions applies to the collection and analysis of vehicle acceleration and rotation data; determination of
impact speed; and measurement, recording, and reduction of surrogate occupant data, including speci-
fications for data channel performance requirements, data channel selection, transducer mounting, sign
convention, digital data processing, timing marks, time of contact, and presentation of results. Part 2 of
the standard addresses the use of photographic instrumentation to collect and analyze the same data.
SAE J2l l-l IUL2007 and J21l-2 NOV2008 are reprinted in Appendix C with the permission of the
Society ofAutomotive Engineers. Additional copies may be purchased directly from SAE at
http://www.sae.org or call (877) 606-7323.
Analytical and
Experimental Tools
h" design, synthesis, and development of a new safety feature is not a straightforward procedure
T
I but is an iterative process requiring tradeoffs among sometimes conflicting impact performance
requirements, environmental considerations, and costs. This appendix summarizes some common ana-
lytical and experimental tools that can be used to support the development and evaluation of new safety
features. Applications and limitations of these techniques are also discussed.
Structural loading and design procedures are contained in numerous civil engineering textbooks,
AASHTO design manuals, and research publications. References containing recommendations for de-
sign loadings and analysis procedures for each type ofsafety feature are listed in Table D-1. Designers/
developels should consult these references to estimate desìgn loads and proportion a new design for'
subsequent evaluation steps. Static and/or dynamic testing and computer simulation should be imple-
mented whenevel necessary to assure that features have sufficient structural capacity.
Longitudinal Barriers
B. Sign Supports 3, 6, 16, 51, 57, 83, 91, '102,',106, 121,122, 123, 138
V. Roadside Geometric Features 4, 38, 40, 69, 101 , 127, 130,152, 160
in a controlled manner. As a result, specialized static tests are often necessary that do not conform to
standard tests suggested by ASTM.
Static testing is often used to compare the performance of competing design details. When such tests
are used to evaluate safety feature components that must perform under dynamic loading, developers
should be aware of the many problems that can arise from material load rate sensitivity. A primary con-
cem in the design of many roadside safety features is the energy absorbed as a component fails. Static
testing generally allows a component to fail at the lowest possible load. However, the lowest failure
load may not coffespond to the lowest energy failure mode. For example, wood posts embedded in soil
seldom fracture under static loading and energy dissipation is usually high as the post rotates in the soil.
Under dynamic loading, soils can generate much higher loadings and a wood post can fracture prerna-
turely with little energy dissipation.
In general, it is anticipated that there are significant variations in the mechanical properties of most
materials. Further, the mechanical characteristics of many materials are specified only in terms of
minimum values, and actual material strengths can be almost double the rated minimum. If a safety
feature relies on the controlled bending or fracture of a material, excessive material strength may be just
as dangerous as a strength that is below the minimum. For example, research has shown that the ener-
gy required to fragment a frangible transformer base can vary more than 100 percent with even minor
changes in heat treatment of the aluminum alloy. Furlher, soil conditions can exhibit even wider season-
al variations as a soil goes through saturated, dry and frozen situations. Designers should utilize static
and/or dynamic testing to evaluate safety feature performance over the expected range of variation in
material properties.
Even at this stage, the developer should be aware of value engineering by avoiding over speciffing
materials, especially components that are not critical to system performance. V/here possible, the devel-
oper should use standard hardware elements for initial economy and to minimize costs associated with
inventory maintenance (9).
Most of the available simulation programs have been correlated to sotne degree with crash tests. For the
validated cases, simulation results can be very helpful to the safety feature designer by providing unique
insight into the collision event. 'Where the program has been validated for multiple impact conditions,
it can sontetimes accurately predict behavior for impact conditions that are bracketed by the validated
conditions. Although conrputer simulations have proven to be invaluable in the development of new
roadside safety features, the accuracy ofthese programs has not yet reached a point that required com-
pliance tests can be replaced by computer modeling. The most important simulation programs are given
in Table D-2 and are described below.
HVOSM-The Highway Vehicle Object Simulation Program (HVOSM) (136) is a very sophisticat-
ed and widely used vehicle handling model. HVOSM incorporates an I l-degree of freedom (DOF)
vehicle model with relatively sophisticated suspension and tire models. This program has been exten-
sively validated against a wide variety of crash tests involving many different terain configurations.
The program has demonstrated validity for modeling vehicle traversals of ditches, driveways, and a
variety of roadside slopes (127). HVOSM is especially suited for evaluation of roadway and roadside
geometrics where vehicle stability is a primary concern. Although this program has also been used for
simulating rigid barrier impacts (104), it has now been superseded by LS-DYNA.
Barrier VII-The Barrier VII program (l I l) is a widely used model for simulating impacts with
flexible barriers. This program incorporates a beam and column finite element model (FEM) of the
barrier and a two-dimensional vehicle model. The FEM code incorporates both geometric and ma-
terial nonlinearities as well as a number of specialized barrier elements including nonlinear springs
and dashpots. Although the vehicle model incorporates relatively simple bilinear spring elements and
is limited to three DOF, this program has been successfully validated for a wide variety of flexible
barriers and a number of different vehicles. The primary limitation of this program is that it cannot be
used to predict vehicle stability. Howeve¡ the program is especially suited for use as a tool for barrier
design in predicting maximum loads on and strains in barrier components. Further, the program has
proven useful for identification of critical impact locations as well as predicting vehicle snagging and
pocketing (18, 137).
LS-DYNA-This program is a highly sophistìcated, nonlinear finite element computer program (82).
The program is commonly used to nlodel vehicular impacts with vimually any roadside safety feature.
Although this program has proven to be capable of accurately predicting the behavior of vehicles and
roadside safety features duling high-speed impacts, some significant limitations remain. LS-DYNA's
primary limitations are related to failure of barrier and vehicle components. Finite element modeling
of metal rupture has not progressed to the point that it can be reliably predicted. Dynamic rupture of
metal components has proven to be very mesh dependent. A developer can sometimes tune a par-
ticular model to fail at the appropriate time in a dynamic test, but even rrinor changes in the load
orientation or model mesh configuration can destroy the program's accuracy. As a result, LS-DYNA's
ability to accurately predict vehicle suspension failure or barrier component rupture is very limited.
Vi = Jrsh
\ilhere:
vi velocity at impact
q
ò acceleration due to gravity
h drop height of mass
The above formula ignores speed losses due to friction and aerodynamic drag and therefore should be
corrected to more accurately assess the actual pendulum impact speeds. As an example, for an impact
speed of 32 mph (9.7 m/s), a drop height of 15.7 ft (4.8 m) is required. The swing radius is usually
considerably larger than the drop height. Gravitational pendulums are commonly used to evaluate
petformance at impact speeds of approximately 25 mph (40 km/h) or less. A gravitational pendulum
capable of high-speed impacts would require very high drop heights and is impractical.
A primary problem associated with this type of testing is the type of impact surface or crushable nose
used on the pendulum. A rigid nose greatly increases the impact forces applied to the pendulum while
reducing the energy dissipation during the test. An excessively soft nose will minimize impact forces
and maximize energy dissipation associated with the tested feature. Although simulated soft noses have
been developed for subcompact and minisize vehicles (19, 85), these devices may be out of date. Nose
assembly systems should be evaluated to ensure that they accurately replicate modern vehicles.
Pendulum testing is frequently used to evaluate the performance of breakaway structures such as lumi-
naire and sign supports. Such systems often absorb more impact energy during low-speed crashes than
during high-speed impacts. As a result, pendulums are an inexpensive method for evaluating the low-
speed perfomrance of prototype design alternatives. Some breakaway systems have been placed into
service based solely on pendulum testing. The acceptance ofsafety features based on such testing is left
to the discretion ofthe user agency.
Pendulums can also be used for dynamic testing of various safety feature components. For example,
pendulums are often used for dynamic testing of barrier posts embedded in soil and crash cushion atten-
'O 2016 by the American Association of State Highway and Transportation Oflìcials
All rights reserved Duplícation is a violation ofapplicable law
216 | Manual for Assessing Safety Hardware
uator elements. This type of testing is not sensitive to the design of the pendulum's crushable nose and
can yield valuable information with a rigid impact surface.
Bogie vehicles may be used to simulate impacts with breakaway structures, work-zone traffic control
devices, longitndinal barriers, or components of such systems. As discussed in Section 4.2.2, bogie ve-
hicles must be revalidated periodically to ensure that the devices are representative of modern vehicles
D2 COMPARISON OF TECHNTQUES
Applications and limitations of safety feature development techniques are given in Table D-3
Development
Princlpal Areas of Appllcation Possible Limitations
Technlque
Preliminary and f¡nal design of Dynamics and kinematics of feature
feature for environment and non-colli- and collision vehicle are not ad-
s¡on performance dressed
Development
Principal Areas of Appllcation Possible Limitations
Technique
4 Dynamic Tests
Compliance test for luminaire lmpact speed 25 mph (40 km/h) or less
and single-leg sign breakaway
supports For dual-leg supports, upper-hinge
mechanisms are not examined
Evaluation of breakaway mecha-
nisms Does not simulate off-center impacts
Measurement of
Vehicle Deformation ]E
n order to determine the amount of interior crush deformation, both pre-impact and post-impact
J
I measurements of the vehicle interior must be taken. Accurate determination of the position of the
floorboard is often critical to this determination. Four sets of measurements are recommended: two
pre-impact data sets and two post-impact data sets. These two sets of data allow for the accurate deter-
mination of interior crush defomation.
E1 PRE.IMPACT MEASUREMENTS
First, a reference line along the floorboard is established at the longitudinal centerline of the vehicle,
extending from the firewall to the back of the interior of the vehicle. This reference line serves as the
X-axis for the measurement of lateral displacement. A second reference line is drawn along the back of
the vehicle's interior, serving as the Y-axis for measurements along the vehicle's length. Accordingly,
if the impact points ale on the driver side of the vehicle, the Y-component values are negative. If the
impact points are on the passenger side, the Y-component values are positive. All X points are positive.
The coordinate system is shown in Figure E-1.
Dashboard 2s 24 2z
3
I 2
I 10
7
13 14
15
11
Door Door
't6 I
'19 20
Ten to fifteen points should be plotted uniformly on the interior side of the vehicle's roof and five to ten
points should also be plotted across both the dash and the interior wall adjacent to the vehicle impact
zone. The X and Y values for each point are then measured in the X-Y coordinate system, using a level
to ensure in-plane measurements. These nìeasurements are to have a minimum accuracy of +l/s in.
(3 mm). Some consideration should be given to the expected deformations when selecting the number
altl locaLiotrs of points to ploL. More points shuuld bc plol.Lecl when a narrow inl.rusion is anticipat-
ed, such as from a luminaire pole falling on a vehicle's roof. When there is a signiûcant risk of direct
impact on an unusual part of the occupant compartment, such as the underside, additional points should
be plotted across the associated area ofthe vehicle as shown in Figure E-1.
In order to measure vertical locations, a swivel laser bracket (or similar leveling device) is aligned
with the back interior wall of the vehicle above the X-Y axis created above. The vertical position of
the swivel laser bracket is documented in two ways. First, the outline of the bracket is marked on the
rear wall of the vehicle. Second, at least two reference points elsewhere on the vehicle are also marked
and documented. Because these reference points are required to maintain the same vertical orientation
for post-impact measurements, these reference points must be placed in locations that have a very low
probability of deformation. The vertical location of the swivel laser level and its corresponding refer-
ence points are designated as the origin of the vertical axis,Z¡An example of a swivel laser bracket is
shown in Figure E-2.
Having established the origin for the coordinate system, the vertical locations of each of the reference
points are then measured using a tape measure and the laser swivel bracket. It is paramount that the
vertical location be measured accurately; accordingly, the tape measure should be plumbed with a level.
Positive Z-values are above the horizontal plane created by the swivel laser bracket; negative Z-values
correspond to points below the horizontal plane. The measurement of vertical positions is shown in
Figure E-3.
A second origin, X'-Y'-Z', is then chosen to provide a reference frame for shifts due to distortions in the
original reference frame. In order to create this second origin, string lines parallel (but not coincident)
with the original reference X and Y axes are selected. This new X'-Y'-Z'coordinate system should be
selected in a manner that minimizes deformations, preferably away from the impact side of the vehicle.
The steps above are repeated in order to acquire X',Y', andZ' locations for each of the points.
After impact, post-impact measurements are taken in the same manner as described above, using the
pleviously defined X-Y-Z andX'-Y'-Z'coordinate systems. A typical plotting diagram for the crush data
points and recolding spreadsheet are shown in Figure E-4. It may be useful to develop a spreadsheet
that automatically plots the crush points on the flool pan drawing.
E3 PHOTOGRAPHIC DOCUMENTATION
Detailed photographs should be included in the crash test report to allow readers to identify both the
locations of the pre-impact measurement grid and document the extent of deformation during the test.
Particular attention should be provided to any area of the vehicle interior where significant deformation
is observed.
Dashboard
Door Door
z
Figure E-4. Vehicle Deformation Spreadsheet
The procedure for lneasuring external crush is described in depth in "Vehicle Measurement
Teclmiques," published by the United States Department of Transportation (I 56). The procedure sum-
marized below is based on this publication.
The purpose of obtaining crush measurements is to develop a mechanism for linking vehicular defor-
mations from full-scale crash tests to real world crash data collected in the NASS-CDS. Further, the
measurements provide a basis for determining vehicle crush energy.
The basis for field data collection is the point-to-point vehicle measurement technique. The measure-
ments are obtained by establishing a reference line, rneasuring the residual crush, and subtracting the
bumper/body taper to obtain the resultant crush profile. An example of a configured reference line is
shown in Figure E-5.
Reference Line
.\'r
\ \,...
I
¡¡-Tape Measure
I
Overall
Length
Tape +
Measure
Equal Distances
The Field Length, or Field L, is the length of contact and induced damage measured parallel to the
undeformed end or side plane of the vehicle. For impacts where the induced damage extends across the
entire width of the end plane, the undeformed end width (UEW) is used in reconstruction programs as
the Iength. For narrow impacts, such as pole impacts, the Field L value is used. Examples of field length
measurements are shown in Figure E-6.
Field L L = Recon.
RF
Bumper
LF
Bumper
Corner
Bumper Corners
Corner
\ \
Crush depth measurements, or C dimensions, are taken at six equally spaced locations across the
Field L. A contour gauge is used to measure crush depths, as shown in Figure E-7 . lt is critical to ensure
that the contour gauge is parallel to the origínal end plane of the vehicle. Note that a vehicle damage
profile is not always uniform. Vehicle crush should be measured at the vertical location of the most rigid
structure in the damaged plane (i.e., frame or side door beam).
c6
cs
tt ¡
c4 I
.tt
c3 i
/
c2
cr
c6 cs c4 ca c2 c1
C"*
Figure E-7. Crush Depth Measurements
The final requisite values to be measured identif, the location of the damage with respect to the center
of gravity of the vehicle, as shown in Figure E-8. The contact damage dimension, D", is the distance
from the center of the contact damage width to either the vehicle's damaged end-plane center or the
damaged vehicle's wheelbase center, for end-plane and side-plane impacts, respectively.
+ + + Original End
Plane Center
Dc
Dc
¡, I
+ +
Center of
-- Field L
I \
Center of
Contact
A Damage
B c
The Field L dimension, Dp¡, is the distance from the center of the Field L (including both contact and
induced damage) to the vehicle's damaged endplane center or the damaged wheelbase center, measured
along the genel'al slope of the damaged plane.
For end impacts, the vehicle's damaged centerline should be used (i.e., even a shifted or damaged center
point on the endplane) to measure Ds¡. For side impacts, the center of the vehicle's damaged wheelbase
should be used (i.e., even if the vehicle's end is shifted).
Determination
of THIV, PHD, and ASI
Fl INTRODUCTION
h. European Cornmittee for Normalization (CEN) has adopted the Theoretical Head Impact
TI
Velocity (THIV) and associated Post-Impact Head Deceleration (PHD), and the Acceleration
Severity Index (ASI) as measures of occupant risks for purposes of evaluating results of a crash test
(140-142). They are presented herein with the hope and expectation that U.S. testers will determine and
repoft these indices. The goal of this effort is (a) to develop a database from which comparisons can be
made between the THIV, ASI, the flail space indices recommended herein, and other measures of occu-
pant risk, and (b) to provide a basis from which future test and evaluation procedures can be formulated
by and harmonized between the United States, CEN, and other countries.
F2.l. GENERAL
The Theoretical Head Impact Velocity (THIV) concept has been developed for assessing occupant im-
pact severity for vehicles involved in collisions with road vehicle restraint systems (78). The occupant
is considered to be a freely moving object (head) that, as the vehicle changes its speed during contact
with the safety feature, continues moving until it strikes a surface within the interior of the vehicle. The
magnitude of the velocity of the theoretical head impact is considered to be a measure of the impact
severity.
The head is presumed to remain in contact with the surface during the remainder of the impact period.
In so doing, it experiences the same levels of acceleration as the vehicle during the remaining contact
period (Post-lmpact Head Deceleration-PHD) (78).
During itnpact, the vehicle is assumed to move only in a horizontal plane, because high levels of pitch,
roll, or veltical llotion are not of prime impoftance unless the vehicle overturns. This extrenre event
does not need to be considered, as in this case the decision to reject the candidate system will be taken
on the basis ofvisual observation or photographic recording.
Two reference frames are used, as indicated in Figure F-1. The first of these is a vehicular reference
C,r, x being longitudinal and y transversal; the origin C is a point at or near the vehicle's center of mass,
where two accelerometers and a rate gyroscope are typically installed (see Section 4.3.2for recom-
mended procedures to determine accelerations and yaw rate at C if the instrumentation cannot be placed
at or near the center of mass). Let x" and¿ be the accelerations of point Cin ftJs21m/s2¡, respectively,
along the x and vehicle axes, recordèd from the two accelerometers, and Y the yaw rate (in radians per
y
second), recorded from the gyroscope ( ï positive forward, ¡i positive to right hand side, and Y positive
clockwise looking from above).
The second reference frame is a ground reference OXY,with the x axis aligned with the initial vehicular
velocity Vs, and the origin O coinciding with the initial position of the vehicular datum point C.
X"(t), Y"(t) are the gound coo¡dinates of the vehicle reference C, while X6(1), YbQ) are the ground coor-
dinates ofthe theoretical head (see Figure F-2).
With these definitions and simpliffing hypotheses, vehicle and theoretical head motion can be computed
as follows.
x
v,
,l
x
t I
vo
.,
x0 t Theoretical Head
t ,
,
) vEHrcuLAR MOTTON
ir=i"cosY-j;"sinY
1=i"sinY+f"cosY (Eq. F2-3)
LX"dt +Vot
(Eq. F2-s)
LYcdt
Coordinates of the theoretical head with respect to the vehicle's frame can be computed by the formula:
TIME OF FLIGHT
Notional impact surfaces inside the vehicle are assumed to be flat and perpendicular to lhe x and y
vehicular axes (see Figure F-2). The distances ofsuch surfaces from the original head position (flail
distances) are D* forward and Dr laterally on both sides.
VD
xb
I
I
xb I
Dx
Dy Dy
The time of flight of the theoretical head is the time of impact on one of the three notional surfaces in
Figure F-2,i.e., the shoftest time Iwhen one of the three following equalities is satisfied:
THIV
Finally, the Theoretical Head Impact Velocity (THIV) is the relative velocity at time T, i.e.,
\i/ + t+l \P
,tv +2v
k
rY Yo; 2v . *+l\lJ _ t\y+
= = 'Y+
2' 2
k
+. ¿+l
yoþ)- -k a,xrsin ÈY
+ o LYocos oY
n
ou,
=-r L*rcos oY
- o
L:i"sin rY + yoo* o
ou, kY o otP o
= -o LX"sin - LY,cos - yooY
6. Find the minimum value ofj for which one of the three equalities is satisfied
rA
=l-c( i: * );î)/'
o
9. Filter the sequence t'A with a digital Butterworth low-pass filter, having a cut-offfrequency
of l0Hz,
a roll-offof 48 dB/octave, and apply a lO-ms moving average;PHD is the maximum of this filtered
sequence.
F3.l PROCEDURE
The Acceleration Severity Index (ASl), developed by TTI (100), is a function of time, computed with
the following fornrula:
where , â, , and à- are limit values for the components of the acceleration along the body axes x, y,
â.,
andz; a,, ar,andã" arethe componentsof theaccelerationof aselectedpointPofthevehicle, aver-
aged over a moving time interval ô : 50 ms, so that:
I I
a, =!1,'tu o,a,; o, =E ardt; d, clz dt
I'.u ò I'.u (Eq. F3-2)
The index ASI is intended to give a measure of the severity of the vehicular motion during an impact for
a person seated in the proximity of point P.
Averages computed in Equation F3-2 are equivalent to what would be obtained by a low-pass filter, and
take into account the fact that vehiculal accelerations can be transmitted to the occupant body through
relatively soft contacts which cannot pass the highest frequencies. Direct use of vehicular accelerations,
even ifaveraged, implies that the parts ofoccupant body that can be injured are continuously in contact
with some part of the vehicle.
Note that Equation F3- I is a basic interaction formula of three variables. If any two components of ve-
hicular acceleration are null, ASI reaches its limit value of I when the third component reaches its limit
acceleration. When two or three components are non null, ASI may be I with the single components
well below the relevant limits. Limit accelerations are interpreted as the values below which occupant
risk is very srnall (light injuries, if any).
In Europe (France, Germany, and the Netherlands), for occupants wearing safety belts, the generally
used limit accelerations are:
where G:32 ft/s2 (9.81 m/s2) is the acceleration of Eafth gravity at sea level.
With the above definition, ASI is a nondimensional quantity, i.e., a scalar function of time and, in
general, of the selected vehicular point, having only positive values. Occupant risk is assumed to be
propoftional to ASI. Therefore, the maximum value attained by ASI in a collision is assumed as a single
measure of the severity, or:
ASI: max[ASI(r)]
(Eq. F3-a)
Vehicular accelerations in the x, y, and z directions are measured at or near the center of mass of the
vehicle (see Section 4.3.2 for recomlnended procedures to determine accelerations in the x and y
directions at the center of mass if the accelerometers cannot be placed at or near the center of mass).
F3.2 SUMMARY
In summary the following steps are used to compute the ASI
1. Record vehicular accelerations in the x, y, and z directions at or near the vehicle's center of mass (see
Section 4.3.2 if accelerometers can not be placed at or near the center of mass). In general,
accelerations are stored on a magnetic tape as three series ofNnumbers, sampled at a certain
sampling rate S (samples/s).
For these three series of measures where acceleration of gravity (G) is the unit of measurement,
compute:
t2
oP cl¡r.-., Ê-t o, o o*, k+l ap at
t2
clyt ay;... k-l Qy, k dy, k+l dv, N
4v
>
I
o"r2or, .,., k-la_ k
a
k+l
a, az
2. Find the number lz of samples in the averaging window õ 0.05 s; thus, m : INT(ô*S) : :
INT(0.05*S), where INT(R) is the integer nearest to R. For example, if ^S: 500 samples per second,
m:25.
k-
Qr=
t( k
-Qrt k+
Qrt
k+2
a"+. * o*^o")=*ätr-
m'
¡ k+¡t
+k ct
) -lr
mfi
,o
Functions of time
kt:h(k+ml2)
oASr
= +(oa,fs)' *(ra,lrc)'f%
lfu,P¡'
G1. INTRODUCTION
n the design of staged attenuation systems, it is necessary to evaluate the performance of attenuation
J
I devices with a wide range of vehicles in terms of size and mass. The current evaluation criterion
focuses on the 2ndand 90th percentile vehicles and requires testing with a 2,425-lb (1,100-kg) small
car vehicle ( 1I 00C) and a 5,004-lb (2,270-kg) pickup Truck (2270P). The smaller I I 00C vehicle tests
the lower end of the device's performance and poses a severe test of the occupant risk measures due to
the small mass of the vehicle; while the 2210P vehicle evaluates the maximum structural and energy
dissipation capacities of the device.
There is concern that design of crash cushions, terminals, and truck-mounted attenuators (TMAs) to
meet the upper and lower bounds of the vehicle fleet does not adequately address the perfotmance of
attenuation devices with mid-size sedan vehicles, especially with respect to devices with staged ener-
gy-absorption systems. In order to address this issue, a test with a3,307-lb (1,500-kg) sedan vehicle
(15004) has been added to the test matrices for tenninals and crash cushions, Tests 38 and 45, respec-
tively. An optional test, No. 54, has also been added to the test matrix for TMAs. Due to the cost of
conducting an additional crash test with the 15004 vehicle, a procedure has been developed to determine
whether or not Tests 38,45, and 54 are necessary for a given teminal or crash cushion design.
The procedure consists of estimating the occupant risk values for the I 500A test based on the accelera-
tion trace obtained from Tests 31 , 41, or 5 I . Tests 31 , 41, and 5l consist of the 2270P vehicle impacting
with its centerline aligned with the centerline of the test article. These tests involve heavier vehicles im-
pacting under the same conditions as Tests 38,45, and 54. Thus, the acceleration traces from these tests
can be used to identif, the need for the 15004 tests. The acceleration traces can be integrated to obtain
the force-deflection characteristics of the test article. The force-deflection data can then be applied
to the smaller vehicle in order to obtain reasonable estimates of the Occupant Impact Velocity (OIV)
and Occupant Ridedown Deceleration (ORD). Note that this analysis will be conservatiye because the
heavier mass of the 2270P vehicle and its higher crush stiffness will produce higher impact forces than
will if the force versus deflection analysis
be experienced during a mid-sized vehicle impact. Therefore,
predicts that the terminal, crash cushion, or TMA will meet evaluation guidelines for OIV and ORD for
3,307 lb (1,500 kg), Tests 38,45, or 54 are not recommended.
The procedure for estimation of the occupant risk values for 3,307-lb (1,500-kg) vehicles is based on the
longitudinal acceleration trace from Tests 3 1,41, or 51. The analysis described below is based exclusively
on the CFC 180 filtered longitudinal acceleration trace. Lateral and vertical accelerations have virtually no
effect on the results of the occupant risk values in head-on crashes with attenuator systems.
2. Integrate the CFC 180 acceleration trace to obtain the longitudinal change in velocity at each time,
tLzzop .
Lvzzzop = orrror-rrrtlodt22iop
!
(Eq. G2-2)
Lr'r,o, = (oL, o, - r * t so
+ oD\ o p
-
"
*, ù(ry) + n;tl, o,
3. Calculate the actual longitudinal velocity of the 2270P vehicle at each time, t'rrro, .
4. Integrate the longitudinal velocity of the2270P vehicle to obtain the displacement of the vehicle.
Although the displacement of the impacting vehicle is not the same as the deformation of the test
article, in order to simplif, the analysis, they will be assumed to be the same. This step identifies the
estimated displacement of the test article at every data point in the longitudinal acceleration file.
5. Calculate the longitudinal force on the 2270P vehicle, which is equivalent to the force exerted by the
test article on the impacting vehicle. This step produces an estimated deceleration force at every data
point in the longiludinal acceleration file that corresponds with the test article deformation calculated in
Step 4.
FLrot = mzz.tot,a'zztop-ctctto
E't E, (Eq. G2-s)
-
'testarticle-'2210P
6. The calculated force versus displacement curve generated in Steps 4 and 5 is then applied to a 3,307-
lb (1500-kg) sedan impacting the system at 62 mph,9l fls (100 km/h, 27 .7 mls). The work done by the
attenuation system at every time step can be calculated as follows:
.t (Eq. G2-6)
Wrl,"r*,".= (4lu*"" + 4l-,J*"" +w,il^,",,
7. The longitudinal velocity of the 3,307-lb (1,500-kg) sedan at every time step can then be calculated
based upon an energy balance using the work-energy derived for the test article.
+ (Zl,J*,r" -
f,**o^(ujr'- )' 41,, **," ) =
).*o^(uj"o* )'
(Eq.G2-7)
,,"r- (uj"o'- )' * (\lJ*,, . -w,Iu ^n )
"sedm -
2m"ed.n
8. Calculate the time interval for the sedan at each load step based on the avelage velocity between
subsequent steps and the displacement of the test article.
change in displacement
v _
^n" change in time
r-l
vl"¿on + V".dnn
vavg
2
2(ôi",,"ni"r. - õi*|"".,"
,,
,.",,un = ___I:;;;;;lI;;;_ ) +,s",,an
,,-,
(Eq. G2-8)
9. Calculate the change in velocity for the sedan vehicle at each time, rj"oun .
10. Integrate the change in longitudinal velocity ofthe sedan vehicle to obtain the occupant displace-
ment for the sedan vehicle at each time, /j"0,, . The longitudinal occupant impact velocity is then
identified as the Avj"o- at the step in which ô1..,"¿* reaches 24 in. (600 mm).
I l. Calculate the estimated acceleration of the sedan vehicle based on the force of the terminal and the
mass of the sedan at each time, lj"o- .
alra* =
(Eq. G2-l l)
12. Calculate the 10 ms average acceleration versus time for the sedan vehicle. Note that the time interval
varies throughout the analysis. Initially, the time interval will be very nearly the data collection interval asso-
ciated with Íhe 2270P test. Because the velocity of the 15004 vehicle will decay more rapidly, the calculated
time intervals will increase. The changing time interval complicates the average acceleration calculation
procedure. Thus, calculation of the l0 ms average values involves searching forward l0 ms in time from the
current time and calculating the con'esponding change in velocity over the actual time period. The actual time
interval will seldom be exactly l0 ms. In recognition of the conservative natu¡e of this analysis technique, it
is recommended that the selected intervals be no less than l0 ms. In other words, it is recommended that the
forward search be conditional on the time interval being greater than or equal to l0 ms.
Search 5 ms backward
^ 5msbackwud
avsedan ^ i . i-5rns
=avs"dar-al'."d^n
Search 5 ms forward:
The procedures described above allow the calculation of estimated occupant impact velocity and l0 ms
average accelerations for a 15004 vehicle striking a staged attenuation system. If the estimated OIV and
RA values are found to comply with the evaluation criteria set forth in Chapter 5, the attenuation system
can be considered to comply with the requirements of Tests 38,45, or 54. Note that this analysis procedure
has proven to be conservative in that it tends to overestimate occupant risk values for the I 5004 vehicle.
This ñnding is attributable to the fact that the foregoing procedure ignores the reduced stiffness and mass
of the 15004 vehicle relative ïoLhe2270P vehicle. The increased stiffiless of the2270P vehicle reduces
the total vehicle crush that occurs during an impact with an attenuator. Vehicle crush contributes to the
energy dissipation during a head-on crash and reduces occupant risk values. Further, the much higher mass
of the2270P vehicle generates higher impact forces during a momentum transfer event. Actual impact
forces applied to a lighter 15004 vehicle would be significantly lower.
@ 2016 by the American Association of State Highway antl Transportation Olllic ials.
All rights reserved Duplication is aviolation of applicable law.
| 241,
Test Vehicle
Selection Procedures lHt
his appendix provides a description of the procedures used to select the small car and light truck test
f
I vehicles recommended herein. As automobile manufacturers revise vehicle designs more frequently,
it is possible that vehicle sizes and/or major characteristics may change significantly over a very short
time. In that event, it is recommended that the procedures described below be utilized to select ne\4/ test
vehicles on an interim basis, without a complete updating of the impact performance evaluation guide-
lines. Some limited full-scale testing of any new test vehicles may be needed to accurately identifl the
relevance ofchanges in test vehicle sizes.
The process ofselecting the appropriate vehicle types and sizes for use as a surrogate for the entire
vehicle fleet begins with an analysis of the curb weight distributions of all vehicles sold in the nation
for a given year. Note that for some vehicles, especially SUVs, curb weight can vary significantly
depending upon variations in the basic vehicle design, such as four-wheel drive or extended cabs on
pickups. Tabulated curb weight and sales data should correctly identify the proporlion of each vehicle
model that incorporated these impoftant optional features. Also note that other optional featut'es, such as
engine size and wheel diameter can modestly affect vehicle weight. When possible, the mass associated
with the most popular variations of these less important vehicle options should be used in data tabula-
tion. Detailed vehicle sales data can be obtained from the Automotive Yearbook, published by Wards
Communications, Inc., (159) or the Market Data Book, published by the Automotive News magazine
(39). Tabulations of vehicle weights from the 2002 Automotiye Yearbook are shown in Tables H-l and
H-2. Tables such as these can be used to determine the tails of the vehicle weight distribution. Table
H-1 f ists the lightest vehicles sold in 2002 and Table H-2 lists the heaviest vehicles sold. Vehicles in
Table H- I are jn mass ordel beginning with the lightest vehicle sold and extending upward. Table H-2
lists the upper end ofthe vehicle spectrum and vehicles are presented in order ofdescending mass. The
last column in both tables represents the percentage of vehicles sold with a mass equal to or less than
the listed model. Note that these tables exclude specialty vehicles with sales volulnes of less than 4,000
units. In 2002,there were a total of 16.8 million vehicles sold in the United States, and a vehicle model
with only 4,000 units sold would represent only 0.02 percent of total vehicle sales.
It is recommended that, under normal circumstances, the 5th and 95tlr percentile heaviest weights be
selected as the tatget weight for the small and large test vehicles, respectively. In recognition of the fact
that vehicle weights increased dramatically prior to the development of the MASH guidelines, ¡he 2nd
and 90th percentile vehicle weights were chosen for the test vehicles in this document. As shown in
Table H-1, the 4-door Honda Civic, weighing2,42l lb (1,099 kg), represents the 2nd percentile lightest
vehicle sold in 2002. Vehicles within +55 lb (25 kg) of the Civic fall within this weight category and
include more than 400,000 cars, or 2.4 percent of all vehicles sold. Test agencies have reported that an
adequate supply of used test vehicles can normally be obtained whenever more than 50,000 vehicles are
sold nationally for any given model. Nevertheless, it is recommended that no test vehicle be selected
unless there are at least 100,000 units sold each year in the target weight range.
As shown in Table H-2, the 95th percentile heaviest passenger vehicle sold in 2002 weighed approxi-
mately 5,420lb (2,460 kg) while a 5,000-lb (2,270-kg) vehicle was near the 90th percentile. Note that
these vehicle weights represent a dramatic increase since the early 1990s when the 95th percentile
heaviest vehicle was approximately 4,400lb (2,000 kg). Due to the rapid increase in vehicle weights
prior to the development of the guidelines in 2009, the 90th percentile vehicle weight was selected as
the appropriate size for the light truck test vehicle.
Cumulative
Gurb Weight, 2002 Sales, Market Share,
Make/Model Market Share,
lb (ks) No. of Units o/o
ot
to
a sports car
b includes hatchback model
c includes equivalent sister brand
TABLE H-2. L¡ght Truck and SUV Weights and Sales Volumes
Vehicles with curb weights near the two selected target weight categories were then compared to
identifl their basic characteristics such as body style, wheelbase, track width, c. g. height, front
overhang, rear overhang, weight distribution, overall height, overall length, and overall width. Other
characteristics, such as engine location, suspension type, and drive wheels were also identified.
Specialty vehicles, such as hybrids or spofts cars, are generally excluded from use as test vehicles due
to the fact that these vehicles are not representative of most vehicles in the fleet. Table H-4 presents a
summary of characteristics of small sedans weighing near2,209lb (1,100 kg).As shown in this table,
all of the vehicles have similar dimensional and mechanical characteristics. Therefore, there was
no reason to exclude any of these vehicles. A similar approach should be taken when selecting the
light truck test vehicle. Note that vehicle c. g. height becomes a very important factor when select-
ing a light truck test vehicle. Provided large SUVs continue to maintain a significant market share
in the future, the light truck test vehicle c. g. height should be selected to be representâtive of this
vehicle class.
A midsize test vehicle, designated 15004, has been added to the test matrix in order to evaluate stag-
ing of energy absorbing terminals, crash cushions, and truck-mounted attenuators. This vehicle will
be used to determine if staging in an attenuation system is designed properly to safely accommodate
high-speed, head-on impaits with mid-sized vehicles. In this situation, the mass of the mid-sized ve-
hicle will carry it beyond the point where the I l00C vehicle is brought to a stop and likely enter into.
the high energy dissipation ranges of an attenuator where deceleration forces may become excessive
for mid-sized cars. Hence, the primary concern is that this test will cause excessive ridedown acceler-
ations. Because activation of attenuation systems is primarily related to vehicle mass and the test is a
head-on impact, where spin-out and rollover are not a factor, total mass is the only important vehicle
parameter for the 15004. The 3,307-lb ( 1,500-kg) vehicle mass was chosen after evaluating the po-
tential for excessive occupant ridedown acceleration in recently tested energy absorbing terminals and
crash cushions. This analysis showed that test vehicles weighing between 2,872 and 3 ,7 55 lb ( I ,3 00
and 1,700 kg) would be most likely to cause excessive ride-down accelerations. The 15004 test vehi-
cle was therefore set to the middle of this range.
Mass, c. g. Height,
Year Make Model
lb (ks) in. (mm)
2001 Ford Expedition (2WD) 5,267 (2,391) 2s.2 (742)
2002 Chevrolet Silverado 1500 HD Crew Cab 5,544 (2,517) 28.2 (716)
2002 GMC Sierra C2500 Regular Cab 5,024 (2,28',t) 27.0 (686)
Front Bumper Ht. (Bottom) 8.5 (216) 8.5 (216) 8.0 (203)
Front Bumper Height (Top) 20.5 (521) 20.0 (508) 20.0 (508)
Rear Bumper Height (Bottom) 10.1 (257) 13.3 (337) 17.5 (445)
Rear Bumper Height [Iop) 20.5 (521\ 23.0 (584) 27.5 (6ee)
Wheel Center Height (Front) 10.6 (270) 10.6 (270) 11.8 (2ee)
Glossary t
ballast-Mass added to vehicle, other than simulated occupant(s) and instrumentation, to simulate car-
go and/or to achieve desired test inefiial mass.
barrier height-The height of a longitudinal bamier measured from the surface of the ground at its face
to the top of the highest longitudinal element.
bogie-A device used as a surrogate for a production model test vehicle. Existing bogies are four-
wheeled devices that are towed into the test afticle. They are typically designed to replicate the dynamic
response of a vehicle for specific tests, e.g., tests of breakaway features. Bogies typically can be used
for both low- and high-speed tests.
center of gravity (c. g.)-Point within test vehicle at which its total mass can be assumed to be
concentrated.
clear zone-The roadside border area, starting at the edge of the traveled way, available for safe use by
errant vehicles. This area may consist ofa shoulder, a recoverable slope, a nonrecoverable slope, and/
or a clear run-out area. Although it is desirable to maximize the available clear zone, minimum width
requirements are dependent on the traffic volumes and speeds and on the roadside geometry.
crash cushion-A device designed primarily to safely stop a vehicle within a relatively short distance.
A redirective crash cushion is designed to contain and redirect a vehicle impacting downstream from
the nose ofthe cushion. A non-redirective crash cushion is designed to contain and capture a vehicle
impacting downstream from the nose of the cushion.
crash test-A test in which a production model test vehicle or a surrogate test vehicle impacts or tra-
verses a highway feature.
critical impact angle (CIA)-For a given test and the attendant range of vehicular impact angles, the
CIA is the angle within this range judged to have the greatest potential for causing a failure when the
test is assessed by the recommended evaluation criteria.
critical impact point (CIP)-For a given test, the CIP is the initial point(s) of vehicular contact with a
feature judged to have the greatest potential for causing a failure when the test is assessed by the recom-
mended evaluation criteria.
curb mass-Mass of test vehicle with standard equipment; maximum capacity of engine fuel, oil and
coolant; and, if so equipped, air conditioning and additional optional mass engine. It does not include
occupants or cargo.
device-Refers to a design or a specific part thereof such as a breakaway device. Note that the terms
"device" and "feature" afe often synonymous.
evaluation criteria-Criteria used to assess the results of a crash test or to assess the in-service
performance of a feature.
exit box-As defined in Chapter 5, the exit box is a rectangular region placed at the point where a vehi-
cle exits from a longitudinal barrier impact. The exit box is utilized to evaluate the vehicle's trajectory
upon exiting a longitudinal barrier installation. It is desirable that an impacting vehicle exits the end of
the box rather than the side of the box.
feature-Refers to a specific element of a highway. It may be a hardware item and its associated foun-
dation, such as a sign or barrier installation, or it may be a geometric element, such as a sideslope or a
ditch cross section.
fill material-Soil placed around roadside safety device during the embedment process.
flail space-Hypothetical space in which a hypothetical occupant is permitted to move during impact.
gating device--A device designed to allow controlled penetration by a vehicle when impacted up-
stieam of the beginning of the length of need (LON). Note there is some distance between the end of a
gating device and the beginning of the LON of the device.
geometric feature-A roadside cl'oss-section element such as a ditch section, an embankment, a drive-
way, a median crossovel or a curb. It also includes drainage structures, such as inlets and culvefi ends,
and devices such as grates used to improve the safety ofthese features.
gross static mass-Sum of test inertial mass and mass of surrogate occupant(s).
Hybrid III dummy-An anthropomorphic dummy, representing the 5Oth percentile male, the spec-
ifications of which are contained in part 572, Subpart E, Title 49 of the Code of Federal Regulations,
Chapter V (10-l -88 Edition).
impact angle (0þ-Angle between normal direction of traffic and approach path of test vehicle into the
test article. The impact angle is the actual vector angle of the vehicle's c. g. with respect to the normal
direction of travel at the time of contact with the test article. Note that the reported impact angle is not
necessarily the same as the vehicle heading angle at the time of impact. The test article should be orient-
ed as it would typically be in service with respect to the normal direction of traffic.
impact point-The initial point on a test afticle contacted by the impacting test vehicle.
impact severity GSFA measul'e of the inrpact severity of a vehicle of lnass M, inrpacting at a speed 4
at an impact angle 0. It is defined as follows: lS: tlzM(Vsin})2.
installation length-The entire length of a safety feature installation. The installation length is mea-
sured parallel to the roadway and includes end treatments.
length of need (LONFThatpart of a longitudinal barrier or terminal designed to contain and redirect
an errant vehicle.
longitudinal barrier-A device whose primary function is to safely redirect an errant vehicle away
from a roadside or median hazard. The three types of longitudinal barriers are roadside barriers, median
barriers, and bridge rails.
longitudinal channelizers-A line of longitudinal elements that are connected together to provide de-
lineation of the edge of the travelway. These systems are normally used to provide a clear indication of
the appropriate route through a work zone or an area with temporary lane deviations. Most longitudinal
channelizers utilize plastic water-filled barrier elements with only a small amount of water to assure that
they remain in placed under environmental loadings. Longitudinal channelizers are not positive barriers
and should never be utilized where a positive barrier is warranted.
non-gating device-A device with redirectional capabilities along its entire length. Note that the end of
a non-gating device is the beginning ofthe length of need for the device.
occupant impact velocity (OlVFVelocity at which a hypothetical "point mass" occLlpant impacts a
surface of a hypothetical occupant compartment.
pendulum-A device used as a surrogate for a production model test vehicle. A mass is attached to
cables, which are in turn suspended from a fixed point. The mass is raised to a selected height and
released, allowing gravity to accelerate the mass as it swings into the test article. The structure of the
mass can be designed to replicate the dynamic crush properties of a production model test vehicle. It is
basically a low-speed test device.
permanent feature-A feature with an anticipated long duration of service, as opposed to one used in a
work or construction zone and having a relatively short duration of service.
pocketing-An undesirable behavior of a redirective device involving relatively large lateral displace-
ments within a relatively short longitudinal distance. The behavior tends to generate large longitudinal
decelerations as the front of the vehicle contacts a poftion of a barrier deformed at a sharp angle relative
to the vehicle's path.
portable work-zone traffic control trailer-A trailer used to carry and/or display a work-zone traffic
control device. Examples include arrow board trailers, changeable message sign trailers, and portable
traffic control signals.
production model test vehicle--A commercially available vehicle with properties matching those
required in a given test.
sprung mass-All mass that is supported by a vehicle's suspension system, including portions of the
mass of the suspension members.
snagging-Contact between a portion of a vehicle, such as a wheel or frame element, and a barrier sys-
tem component that is approximately perpendicular to the normal direction of vehicle travel. The most
common type of snagging is when a wheel engages the side of a post. The degree of snagging depends
on the degree ofengagement. Snagging can cause large and unacceptable vehicular decelerations.
soil strength-A measure of the suppoft of the surrounding soil provided to ground mounted safety de-
vices. As described in Appendix B, soil strength is measured in terms of the lateral resistance provided
by soil for a standardized guardrail post subjected to static and dynamic loading.
surrogate occupant-A dummy, set of sand bags, or other artifact used to simulate the mass effects, to
study the dynamic response of an occupant in a vehicle, or both.
surrogate test vehicle-A bogie, pendulum, or other substitute device designed to replicate the dynam-
ic response of a production model vehicle when in collision with a roadside feature.
temporary feature-A feature nsed in a wolk, construction, or maintenance zone. Its duration of use is
nonnally relatively short, usually one year or less.
terminal-A device designed to treat the end of a rigid hazard or longitudinal banier. A tenninal may
function by (l) deceleraling a vehicle to a safe stop within a relatively short distance, (2) permitting
controlled vehicle penetration behind the device, (3) containing and redirecting the vehicle, or (4) a
combination of 1,2, and 3.
test article (test featureFAll components of a system, including the foundation, being evaluated in a
crash test. Note that the system may be a geometric feature such as a ditch or driveway slope.
test inertial ¡¡1¿55-þl¿ss of test vehicle and all ítems rigidly attached to vehicle's structure, including
ballast and instrumentation. Mass of sunogate occupant(s), if used, is not included in test inertial mass.
Test Level (TLÞ-A set of conditions, defined in terms of vehiculal type and mass, vehicular impact
speed, and vehicular impact angle, that quantifies the impact severity of a matrix of tests.
test vehicle--A commercially available production model vehicle or an approved surrogate vehicle
used in a crash test to evaluate the impact performance of a test a¡ticle.
Theoretical flead Inrpact Velocity (THIVFThe resultant velocity at which a hypothetical "point
mass" occupant intpacts the surface of a hypothetical occupant compartment. Differs from occupant
impact velocity due to the occupant contact being calculated based on the combination of vehicle
deceleration and yaw motion and the impact velocity being a result of both the lateral and longitudinal
occupant velocity values at the time of occupant impact.
transition-That part of a longitudinal barrier system between and connecting sections of differing
lateral stiffness and/or sections of differing design or geometry.
unsprung mass-All mass which is not carried by the suspension system, but is supported directly by
the tire or wheel and considered to move with it.
utility pole-A support structure used to support power transmission or comlnunication lines.
vehicle rebound-The distance that a vehicle rebounds from an impact with a crash cushion or end
treatment. Vehicle rebound ìs intended to provide a measul'e of the risk that a vehicle will bounce off of
an attenuator and roll backwards into the travelway.
working width-The distance between the traffic face of the test article before the impact and the max-
imum lateral position of any major part of the system or vehicle after the impact.
work-zone traffic control device-A device used in a work zone to regulate, warn, and guide road
users and advise them how to tl'averse a section of highway or stl'eet in the proper manner. Work-zone
traffic control devices of interest herein include signs, plastic drums, and lights that may be used there-
on; cones, barricades, chevron panels, and their support system; and any other such device(s) commonly
exposed to traffic that may pose a hazard to occupants of a vehicle, to work-zone
personnel, or to both.
References and
B¡bliography ]Rì
L AA. Aluminum Design Manual.ADM2Ol5. Aluminum Association, Inc., Washington, DC, 2015.
2. AASHTO. Construction Manualfor Highway Construction AmericanAssociation of State
Highway and Transportation Officials, Washington, DC, 1990.
3. AASHTO. LRFD Specifications for Structural Supports for Highway Signs, Luminaires, and
Troffi" Signals. American Association of State Highway and Transportation Officials, Washington,
DC,2015.
4. AASHTO. Roadside Design Guide. American Association of State Highway and Transportation
Officials, Washington, DC, 2011.
5. AASHTO. Standard Specifications for Highway Bridges. American Association of State Highway
and Transportation Offi cials, Washington, DC, 2002.
6. AASHTO. Standard Specffications for Struclural Supports for Highway Signs, Luminaires, and
Trffic Signals.American Association of State Highway and Transportation Officials, Washington,
DC,2013.
7. AASHTO and AWS. AASHTO/AWS DI.5M/D1.5:2015 Bridge llelding Code. American
Association of State Highway and Transpoftation Officials, Vy'ashington, DC, and American
Welding Society, Miami, FL,2015.
8. AASHTO-AGC-ARTBA. A Guide to Standardized Highway Lighting Pole Hqrdware. Joint
Report from Task Force 13, American Association of State Highway and Transpofiation Officials,
Associated General Contractors ofAmerica, American Road and Transpoftation Builders
Association. AASHTO, Washington, DC, April 1980. [Maintained as an online guide.]
9. AASHTO-AGC-ARTBA. A Guide to Standardized Híghway Baruier Hardware. Joint Report
from Task Force 13, American Association of State Highway and Transportation Officials,
Associated General Contractors ofAmerica, American Road and Transportation Builders
Association. AASHTO, Washington, DC, 1995. [Maintained as an online guide.]
10. AASHTO-AGC-ARTBA. A Guide to Small Sign Support Hardware. Joint Report from Task
Force 13, American Association of State Highway and Transportation Officials, Associated
General Contractors ofAmerica, American Road and Transpofiation Builders Association.
AASHTO, Washington, DC, 1998.
I l. AISC. Steel Construction Manual, l4th Edition. American Institute of Steel Construction, Inc.,
Chicago, lL,2011.
12. Alberson, D. C. and D. L. Ivey. Improved Breakaway Utility Pole, AD-IV. ln Transportation
Research Record I46S.Transpoftation Research Board, Vy'ashington, DC, December 1994.
13. Beason, W. L. and T. J. Hirsch. Measurentenl of Heavy Vehicle Impact Forces and Inerlial
Properties, FHWA-RD-89-120. Texas Transportation Institute, Texas A&M University, College
Station, Texas, May 1989.
14. Beedle, L. S. Plastic Design of Steel Frames. John Wiley & Sons, Inc., New York, NY, 1958.
15. Bligh, R. P., D. C. Albertson, W. L. Menges, and R. R. Haug. Evaluation of Dual Support,
Triangular Slip-Base Sign Installations. Inlccident Investigation Quarterly and Journal, No. 41.
Accident Reconstruction Network (ARC Network), Waldorf, MD, 2006.
16. Bligh, R.P., and W.L. Menges. MASH TL-3 Testing and Evaluation of the Midwest Cable Median
Barrier. Texas Transportation Institute Proving Ground, TexasA&M University, College Station,
TX, December20ll.
17. Bligh, R. P., H. E. Ross, and D. L. Bullard. ofArizona Slip-Away Base
Test and Evaluation
Luminaire Supports: Final Reporl. Report No. 7236-lF. Final Report to the Arizona Department of
Transportation. Federal HighwayAdministration, U.S. Department of Transportation, Washington,
DC, September 1994.
18. Bligh, R. P. and D. L. Sicking. Applications of Barier VII in the Design of Flexible Barriers. In
Transportation Research Record I233.Transportation Research Board, Washington, DC, 1990.
19. Brauer, S. K., J. B. Mayer, and R. E. Kirksey. I800-lb Pendulum Mass with I}-Stage Crushable
Nose Element Calibration Tests PCAL-I & PCAL-2, Southwest Research Inslilute Test Report.
Southwest Research Institute, San Antonio, TX, October 1987.
20. Bronstad, M. E. Guardrail Ends. Proceedings
for the American Associationfor Automotive
Medicine, SanAntonio, TX, 1983, pp.389407.
21. Bronstad, M. E., L. R. Calcote, and C. E. Kimball. Concrete Median Barrier Research.Final
Report, Contract DOTFH-l l-8130, Volumes I and II. Southwest Research Institute, San Antonio,
TX, June 1976.
22. Bronstad, M. E. and J. D. Michie. Nationql Cooperative Highway Research Program Report 239:
Multiple-Service-Level Highwøy Bridge Railing Selection Procedures. NCHRP, Transportation
Research Board, Washington, DC, November 198L
23. Bronstad, M. E., J. D. Michie, and J. D. Mayer, Jr. National Cooperative Highway Research
Program Report 289: Perþrmance of Longítudinal Trffic Barriers. NCHRP, Transportation
Research Board, Washington, DC, June 1987.
24. Bronstad, M.E., eTal. Guardrail-Bridge Rail Transition Designs. Research Report FHWA/RD-
861178, Volume I. Southwest Research Institute, San Antonio, TX, April 1988.
25. Brown, C.M. Validation of the ENSCO Surrogate Bogie Vehicle. Research Report, U.S.
Depattment of Transportation Contract No. DTFH6I-91-2-00002, Federal Highway
Administration Report No. FHWA-Rd-93-074. Advanced Technology & Research Corp., Laurel,
MD, for Federal Highway Administration, U.S. Depaftment of Transportation, Washington, DC,
1994.
26. Buth, C. E., W. L. Menges, K. K. Mak, and R. P. Bligh. Transitions from Guardrail to Bridge
Rail That Meet Safety Performance Requirements.ln Transporlalion Research Record 1720.
Transportation Research Board, Washington, DC, 2000.
27. Buth, E., eIal. Developntent of Safer Bridge Railing Designs. Final Report Draft, FHWA Contract
FH-l-9181. Federal HighwayAdministration, U.S. Department of Transportation, Washington,
DC, February 1981.
28. Buth, E., etal. Perþrmance Limits of Longitudinal Banier Systems. Final Report on Contract
DTFH6I-82-C-00051. Texas Transportation Institute, Texas A&M Research Foundation, College
Station, TX, May 1985.
29. Calcote, L. R. Development of a Cost-Effectiveness Model for Guardrail Selection. Final Report
on Contract DOT:FH I l-8827 . Southwest Research Institute, San Antonio, TX, November 1977.
30. Campise, W.L. Comparative Crash Tests Conducted on Seven Different Makes and Models of
Truck Mounted Attenuators (TMAs): Final Repor¡. Final Report to the Texas Department of
Transportation, Texas Transportation Institute, Texas A&M University, College Station, TX,
August 1991.
31. Camey, J. F., III. Development of a Metal Tube Crash Cushionfor Narrow Hazard Highway
^Si/es.
Report FHWA-CTRD-HPR-1080. Federal Highway Administration, U.S. Department of
Transportation, Washington, DC, I 986.
32. Carney, J. F., III, Albin. Development of a 100-km/h Reusable High-
S. Chatterjee, and R. B.
Molecular Weight/High-Density Polyethylene Truck-Mounted Attenuator.In Transportation
Reseorch Record l64T.Transportation Research Board, Washington, DC, November 1998.
33. Carney, J. F., III, C. E. Dougan, and E. C. Lohrey. NCHRP Report 350 Crash Test Results for
Connecticut Truck-Mounted Attenuator.ln Transportatìon Research Record l528.Transpoftation
Research Board, Washington, DC, 1996.
34. Chisholm,.D. B. and J. G. Viner. Dynamic Testing of Luminaire Supports. ReportNo. FH'WA-
RD-73-55. Federal Highway Administration, U.S. Department of Transportation, Washington, DC,
1912.
35. Chou, C. C., K. Hancock, and S. Basu. NARD: Numerical Analysis of Roadside Design, Version
2.0. Final Report on Contract DTFH6I -87 -2-00116. Federal Highway Administration, U.S.
Department of Transportation, Washington, DC, July 1989.
36. Chou, C. C., et al. Guard Versíon 3.I Users and Prograntmers Manual.Final Repoft on Contract
DTFH6I-87-2-00116. Federal Highway Administration, U.S. Deparlment of Transportation,
'Washington,
DC, July 1989.
37. Consolazio, G., K. Gurley, R. Ellis, and J. V/ilkes. Temporary Low Profile Barrierfor Roadside
Sofety, Phase 1L Phase II Reporl to the Florida Depaftment of Transportation, University of
Florida, Gainesville, FL, January 2003.
38. Council, F. M., et al.Safe Geometric Designfor Minicars. ReportNo. FHV/A/RD-871047. Federal
Highway Administration, U.S. Department of Transportation, Washington, DC, March 1987.
39. Crain Communications, Inc. rket Data Book. Published by Automotive News magazine. Crain
Communications, Inc., Detroit, MI, 2008.
40. Deleys, N. J. Safety Aspects of Roadside Cross Section Design. Final Report,
FHWA Contract
No. DOT:FH- I I -8 I 89, Report No. FHWA-RD-75-4 I . Federal Highway Administration, U.S.
Department of Transportation, Washington, DC, February I975.
41. Deleys, N. J. and L. O. Parada. Rollover Potenüql of Vehicles on Embankrnents, Side Slopes,
and Other Roodside Features. Final Reporl, Contract No. DTFH6l-83-C-00060. Calspan Corp.,
Buffalo, Nl August 1986.
42. Dewey, J. F., et al. A Study of the Soil-Sftuclure Interaction Behavior of Highway Guardrail Posts.
Research Report 3431. Texas Transportation Institute, Texas A&M University System, College
Station, TX, July 1983.
43. Edwards, T. C., et al. National Cooperative Highway Research Program Report 77: Development
of Design Criteriafor Safer Luminaire Supports. NCHRP, Transportation Research Board,
Washington, DC,1969.
44. Eggers, D.W, and T. J. Hirsch. The Effects of Embedment Depth, Soil Properties, and Post þpe
on the Performance of Highway Guardrail Posf. Research Report 405-1. Texas Transportation
Institute, Texas A&M University System, College Station, TX, August 1986.
45. European Committee for Standardizalion. Road Restraint Systems. BS EN 1317.
46. Faller, R. K., K. A. Polivka, Beau D. Kuipers, R. W Bielenberg, J. D. Reid, J. R. Rohde, and D. L.
Sicking. Midwest Guardrail System for Standard and Special Applications .ln Transportation
Re s e arc h Re c ord I I 9 0. Tr ansportation Research B oard, Washin gton, D C, 200 4.
47. Faller; R. K., J. D. Reid, and J. R. Rohde. Approach Guardrail Transition for Concrete Safety
Shape Banierc.ln Transportalion Research Record l64T.Transportation Research Board,
Washington, DC, November 1998.
48. Faller, R. K., M. A. Ritter, S. R. Duwadi, and Barry T. Rosson. Railing Systems for Use on
Timber Deck Bridges.ln Transportation Reseqrch Record l656.Transportation Research Board,
Vy'ashington, DC,1999.
49- Faller, R. K., B. T. Rosson, M. A. Ritte¡ E. A. Keller, and S. R. Duwadi. Development of Two Test
Level2 Bridge Railings and Transitions for Use on Transverse Glue-Laminated Deck Bridges. In
Transportation Research Record l743.Transpoftation Research Board, Washington, DC, 2001.
50. Faller, R. K., D. L. Sicking, K. A. Polivka, J. R. Rohde,and B. V/. Bielenberg. Long-Span
Guardrail System for Culveft Applications. ln Transportalion Reseqrch Record 1720.
Transpoftation Research Board, Vy'ashington, DC, 2000.
51. FHV/A. Cost-Effectiveness of Small Highway Sign Supporfs. FHWA Contract FH-l l-8821, Report
No. HV/A/RDI80/502. FederalHighwayAdministration, U.S. Department of Transportation,
'Washington,
DC, 1980.
52. FHV/A. Specifcations for the Collection and Storage of Crash Test Data, Volume II. Report
No. FHWA-RD-91-039. Federal Highway Administration, U.S. Department of Transportation,
Washington, DC, 1991.
53. FHV/A. Manual on Uniþrm Trttffi" Conrrol Devices (MUTCD). Federal Highway Administration,
U.S. Department of Transportation, Vy'ashington, DC, 2009.
54. Fleck, J. T. Validqtion of the Crash hctim Simulalor. Report No. DOT:HS-806 279, Volumes I
through IV. U.S. Department of Transportation, Washington, DC, December l98l .
56. Ford Motor Company. 2005 Body Builder Layout Book, Tnrck Body Builder Advisory Service,
Appendix-Design Recommendations, Second Unit Body Mounting. Ford Motor Company, 2005,
pp. 186-194. [https//www.fleet.ford.com/truckbbas/topics/2005/subm.html]
57. Ghanoudi, M. K., C. M. Brown. Testing of a Modified Oregon Muhidirectional Slip-Base Sign
Support, Foil Test Numbers. Report to MiTech Incorporated for Federal Highway Administration,
U.S. Departrnent of Transportation, rWashington, DC, 1997.
58. Griffin, L. I., III, etal. An Evaluation of SelectedTruck MountedAttenuators (TMAs) with
Recommended Perþrmance Specifications. Texas Transportation Institute, Texas A&M University,
College Station, TX, December 1990.
59. Gurfinkel, G. Wood Engineering. Southern Forest Products Association, New Orleans, LA, 1973.
60. Hansen, A. G., M. W Hargrave, and C. R. Horr. Validation of a Surrogate Vehicle for Luminaire
Support Certification Testing. lnTransportation Research Record l233.Transportation Research
Board,'Washington, DC, 1989.
61. Haque, F., and N.M. Sheikh. Vehicle Dynamics Analysis of MASH Pickup and Small Car Trqversal
through Symmetric V-Ditch. Texas A&M Transportation Institute, Texas A&M University, College
Station, TX, September 2013.
62. Hargrave, M. W., A. G. Hansen, and J. A. Hinch. A Summary of Recent Side Impact Research
Conducted by the Federal Highway Administratio¡n. American Society of Civil Engineers, Reston,
vA, lggg.
63. Hascall, J. A. Investigating the Use of Small-Diameter Softwood for Guardrail Posts. Master's
Thesis, Submitted to the Graduate Çollege of the University of Nebraska-Lincoln, Lincoln, NE,
December 2005.
64. Henson,4., et al. Development of Additional Federal Outdoor Impact Laboratory (FOIL)
Facilìties, Volume II: Valídatton oJ the FOIL Pendtilum Upgrade. Research Repoft, U.S.
Department of Transporlation Contract No. DTFH6 I -87-X-00044, Federal Highway
Administration Report No. FHWA-RD-9O-O85. Scientex Corporation for Federal Highway
Administration, U.S. Department of Transportation, Washington, DC, 1990.
65. Heydinger, G. J. and R. A. Bixel. Rollover Stability fuIeqsurements for 2002 New Car Assessment
Program (NCAP). Final Report on Contract DTNH22-01-C-02004. National Highway Traffic
Safety Administration, S.E.A. Inc., Columbus, OH, December 2002.
66. Hinch, J., et al. Foil Construtclion, Laboratory Proced'ures to Determine the Breakawøy Behavior
of Luminaire Supports in Mini-Sized Vehicle Collisions, Volume I, II, lII, Report Numbers
FHWA-RD-86- 105- 107. Federal Highway Administration, U.S. Department of Transportation,
'Washington,
DC, 1987.
67 . Hinch, J., et al. Impact Attenuators: A Current Engineering Evaluation. In Transportation Research
Record II9S.Transportation Research Board, Washington, DC, 1988.
68. Hirsch, T. J. Longitudinal Baniers for Busses and Trucks. ln Transportation Research Record
I 0 5 2. Transpoftation Research Board, Washington, DC, 1 986.
69. Holloway, J. C., D. L. Sicking, and B. T. Rosson. Perþrmance Evaluation of NDOR Mountable
Curbs. Report No. TRP-03-37-93. Final Report to the Nebraska Depanment of Roads, Midwest
Roadside Safety Facility, Civil Engineering Deparlment, University of Nebraska-Lincoln, Lincoln,
NE, June 1994.
70. Humphries, J. and T. D. Sullivan. Guidelines for the Use of Truck-Mounted Attenuators in Work
Zones. ln Transportation Research Record I304.Transportation Research Board, National
Research Council, Washington, DC, 1991.
71. lSO. Road Vehicles with Two Axles-Determination of Centre of Gravity. Reference Number ISO
10392:20ll.International Organization for Standardization, Geneva, Switzerland, March 15,2011.
72. Ivey, D. L. and Morgan, J. R. Timber Pole Safety by Design. lnTransportation Research Record
I 0 65 . Tr ansportati on Research B oard, Washington, DC, I 9 8 6.
73. Ivey, D. L. and D. L. Sicking. "The Influence of Pavement Edge and Shoulder Characteristics
on Vehicle Handling and Stability ." In Transportation Research Record I 084. Transportation
Research Board, Washington, DC, 1986.
74. James, M. E. and H. E. Ross, Jr. HVOSM User's Manual. Research Report 140-9. Texas
Transportation Institute, Texas A&M University, College Station, TX, August 1974.
75. Karcher, J. Vehicle Dynamics Modeling and Simulation for the Safety Evaluation, Selection,
and Placement of Cable Barrier Systems. The George'Washington University, Master's Thesis,
Washington, DC,2009.
76. Klien, R. M., W.A. Johnson, and H. T. Szostak. Influence of Roadway Disturbances on Vehicle
Handling. Final Report, Contract DOT-HS-5-01223. Systems Technology, Inc., Joliet, IL, October
1976.
77 . Labra, J. J. Development of Safer Utility Poles . Contract DOIFH- I I -8909, Final Report.
Southwest Research Institute, San Antonio, TX, February 1980.
78. Laker, L B. and A. R. Payne. Transportation Research Circulqr Number 396: Theoretical Head
Impacî Velocity Concept. Transportation Research Board, National Research Council, Washington,
DC,May 1992.
79. Lampela, A. A. and A. H. Yang. Analysis of Guardrail Accidents in Michigan. Report TSD-243-74.
Michigan Department of State Highways and Transportation, JuJy 1974.
80. Lawrence, L. R. and J. H. Hatton, Jt. Crash Cushions Seleclion Criteria and Design. Federal
Highway Administration, U.S. Department of Transportation, Washington, DC, September 1975.
8l . Leonin, C. and R. Powers. In-Service Evalualion of Experimental Trcffic Barriers: An Interim
Report. Reporl No.FHWA-DP64lEP7-1. Federal Highway Administration, U.S. Department of
Transportation, Washington, DC, April 1986.
82. Livennore Software Technology Corporation. LS-DYNA software. Livermore Software
Technology Corporation, 7374Las Positas Rd., Livermore, CA.
83. Mak, K. K., R. P. Bligh, and W. B. Wilson. V/yoming Road Closure Gate. ln Transportation
Research Record l52S.Transpoftation Research Board, Washington, DC, September 1996.
84. Mak, K. K. and D. L. Sicking. Rollover Caused by Concrete Safety Shaped Barrier. In
Transportation Research Record l25S.Transpoftation Research Board. Vy'ashington, DC, 1992.
85. Mak, K. K., D. L. Sicking, and H. E. Ross, Jr. Real World Impact Conditions for Ran-Off-the-Road
Accidents. ln Transporlation Research Record 1065.Transportation Research Board, Vy'ashington,
DC, 1986.
86. Marzougui, D., C. D. Kan, and K. S. Opiela. Evaluation of the Influences of Cable Barrier Design
and Placement on Vehicle to Barrier Interface. NCAC Document 2008-W-001. National Crash
Analysis Center, George Washington University, Washington, D.C., 2008.
87. Marzougui, D., C. D. Kan, and K. S. Opiela. Furlher Considerations for Effective Median Barrier
Lateral Placement for Varying Highway Cross Sections. Journal of the Transportation Research
Board,Issue 2437, Washington DC,2014, pp 63-77.
88. Marzougui, D., C. D. Kan, and K. S. Opiela. Vehicle Dynamics Investigations to Develdp
Guidelines for Crash Testing Cable Baniers on Sloped Surfaces. 'Working Paper NCAC 2010-W-
009. National Crash Analysis Center, George Washington University, Ashburn, Virginia, August
2010.
89. Marzougui, D., U. Mahadevaiah, F. Tahan, C. D. Kan, R. McGinnis, and R. Powers. Development
of Guidance for the Selection, Use, and Maintenance of Cable Barrier Systems. National
Cooperative Highway Research Program (NCHRP) Report No. Tll.Transportation Research
Board of the National Academies, Washington,DC,2072.
90. Marzougui, D., K.S. Opiela, C. C. Story C. D. Kan, and E. Arispe. Testing and Analyses of Sloped
Terrain Effects on Vehicle Trajectories and Kinematics. Internal Report. Center for Collision
Safety and Analysis, George Mason University, Yirginia,2014.
91. Mauer, F., D. L. Bullard, D. C. Alberson, and \M. L. Menges. Development and Testing of Steel
U-Channel Slip Safe Sign Support.In Transportation Research Record I5gg.Transpofiation
Research Board, Washington, DC, 1997.
92. Michie, J.D. Nalional Cooperative Highway Research Program Reporr 230: Recommended
Procedures for the SafeQ Performance Evaluation of Highway Appurtenances. NCHRP,
Transportatìon Research Board, Washington, DC, March 1981 .
93. Michie, J.D. Perþrntance qnd Operational Experience of Truck-Mounted Attenuølors, NCHRP
Project 20-5, Topic 22-01. National Academy Press, Transportation Research Board, Vy'ashington,
DC, tgg2.
94. Michie, J. D., L. R. Calcote, and M. E. Bronstad. National Cooperative Highway Research
Program Report lI5: Guardrail Performance and Design NCHRP, Transportation Research
Board, Washington, DC, 197 l.
95. Mongiardini, M., R. K. Faller, S. K. Rosenbaugh, and J. D. Reid. Test Matrices for Evaluating
Cable Median Baruiers Placed in V-Ditches. Research Report TRP-03-265-12. Midwest Road
Safety Facility, University ofNebraska, Lincoln, NE, 2012.
96. NCHRP. Determination of Safe/Cost Effeclive Roadside Slopes qnd Associated Clear Distances.
National Cooperative Highway Research Progranr Project l1-11(02), Texas Transportation
Institute, Texas A&M University, College Station, Texas. (in progress).
97. NCHRP. National Cooperative Research Report 665: Identification of Vehicular Impact
Conditions Associatedwith Serious Ran-Off-Road Crashes. NCHRP, Transportation Research
Board, Washington, DC, 20 I 0.
98. NHTSA. National Accident Sampling Sysrem, Vehicle Measurernent Tecltniques. National
Highway Traffic Safety Administration, U.S. Department of Transportation, Oklahoma City,
Oklahoma, 1998.
99. NSC. for T"ffic Accident Investigators. National Safety Council, 444
Vehicle Damage Scale
Michigan Avenue, Chicago, Illinois, 6061 l, I 984.
100. Olson, R. M., etal. National Cooperative Highway Research Program Reporl 149: Bridge Rail
Design: Factors, Trends, and Guidelines. NCHRP, Transportation Research Board, Washington,
DC, 1974,49 pp.
l0l. Olson, R. M., etal. National Cooperative Highwøy Research Program Report 150: Effect of
Curb Geometry and Location on Vehicle Behavior. NCHRP, Transportation Research Board,
Washington, DC,1974.
102. Paulsen, G. Vy'. and J. D. Reid. Nonlinear Finite-Element Analysis of Dual Support Breakaway
Sign. In Transportation Research Record I52S.Transportation Research Board, lilashington, DC,
September 1996.
103. Perchonok, K. et al. Hazardous Efièct of Highway Features qnd Roadside Objects,Yol.2,
FHWA Report No. FHÌ/A-RD- 7 8-202. F ederal Hi ghway Admini stration, U. S. Department of
Transportation, Washington, DC, September I 978.
104. Perera, H. S. Development of an Improved Highway-Vehicle-Object-Simulation Model for Multi-
Faced Rigid Barriers. lnTransportation Research Record I2 j3-Transportation Research Board,
'Washington,
DC, 1989.
105. Pfiefer, B. G., J. C. Holloway,R.K.Faller, E. R. Post, and D. L. Christiansen. Full-Scale Crash
Tests on a Luminaire Support 4-Bolt Slipbase Design. ln Transportation Research Record I367.
Transpoftation Research Board, Washington, DC, 1992.
106. Pinelli, J. P., C. S. Subramanian, and J. Tabora. Experimental Study of Breakaway Highway Sign
Connections.InJournalofTransportationEngineering,Yol. l28,No. l.AmericanSocietyof
Civil Engineers, Reston, YA, January 2002.
107. Polivka, K.4., R. K. Faller, J. C. Holloway, J. R. Rohde, and D. L. Sicking. Safety Perþrmance
Evaluation of lulissouri's Self-Driving Temporary Sign Stands . Transportation Research Report
No. TRP-03-97-00. Final Report to the Midwest States'Regional Pooled Fund Program, Midwest
Roadside Safety Facility, University ofNebraska-Lincoln, Lincoln, NE, December 13, 2000.
108. Polivka, K. 4., R. K. Faller, and J. R. Rohde. Guardrail Connection for Low-Fill Culverts. In
Transporlation Research Record /85/. Transportation Research Board, Washington, DC, 2003.
109. Polivka, K.4., R. K. Faller, D. L. Sicking, J. D. Reid, J. R. Rohde, and J. C. Holloway. Crash
Testing of MissouriS lIl-Beam to Thrie Beam Trqnsiîion ElemenL Final Report to the Midwest
State's Regional Pooled Fund Program, Transportation Research Report No. TRP-03-94-00,
Project No. SPR-3(017)-Year 9. Midwest Roadside Safety Facility, University of Nebraska,
Lincoln, NE, September 12,2000.
I10. Polivka, K.4., et al. Development of the Midwest Guardrail System (MGS)'W-Beam to Thrie-
Beam Transition. In Transportation Research Record 2025. Transportation Research Board,
National Research Council, Washington, DC, 2007.
II l. Powell, G. H. A Computer Program for Evaluation of Automobile Barrier Systems. Report DOT:
RD-73-51. Federal Highway Administration, U.S. Department of Transportation, Washington,
DC,1973.
I12. Ray, M. H. Preliminaty Recommendations for Perfoming Side Impact Crash Tests of Roadside
Safety Features. In Issues þr Improving Roadside Safety, Transportation Research Circular.
Transportation Research Board, Washington, DC, 1999.
I13. Ray, M. H. and J. F. Camey, lll. Side Impact Test and Evaluation Procedures for Roadside
Structure Crash fusls. Final Report, Contract DTFH6I-88-R-00092, Department of Civil and
Environmental Engineering, Vanderbilt University, Nashville, TN, March 1992.
114. Ray, M. H. and J. F. Camey, III. Test and Evaluation Criteria for Side Impact Crash Tests. In
ASCE Journal of Transportation, Vol. 120 No. 4. American Society of Civil Engineers, Reston,
VA, JulyiAugust 1993.
I15. Ray, M. H., M. V/. Hargrave, J. F. Carney, III, and K. Hiranmayee. Side Impact Test and
Evaluation Criteria for Roadside Safety Hardware. ln General Design and Roadside Sqfety
Features, Transportation Research Record I64T.Transportation Research Board, Washington,
DC, lggg.
I 16. Ray, M. H. and K. Hiranmayee. Evaluating Human Risk ín Side Impact Collisions with
Roadside Objects. ln Roadside Safety Features and Hydraulic, Hydrologt and Water Quality
Issues, Transportation Research Record lT20.Transportation Research Board, Washington, DC,
2000.
I 17. Ray, M. H. and J. D. Michie . Evaluation of Design Analysis Procedures and Acceptance
Criteriafor Roadside Hardware Vol. IV: The Intportance of lhe Occupant Risk Criteria. Report
No. FHWA/RD-87/099. Federal Highway Administration, U.S. Department of Transpoftation,
Washington, DC, July 1987.
I18. Ray, M. H., J. A. Weir, and J. A. Hopp. National Cooperative Highway Research Program
Report No. 490: In-Service Perþrntance of Traffic Barriers. NCHRP, Transpoftation Research
Board, Washington, DC, 2003.
I19. Ray, M.H., etal. Evaluation of DesignAnalysis Procedures andAcceptance Criteriafor
Roadsíde Hardware Vol. V: Hazards of the Redirected Car. Report No. FHWA/RD-87/100,
Federal Highway Administration, U.S. Department of Transportation, Washington, DC, July
1987.
120. Reid, J. D. Designing for the Critical Impact Point on a New Bullnose System. ln International
Journal of Crashworlhiness, Vol. 5, No. 2. Taylor and Francis Group, Oxfordshire, UK, 2000,
pp.141-152.
121. Reid, J. D. Dual-Support Breakaway Sign with Modified Fuse Plate and Multidirectional Slip
Base. In Transportation Research Record l52S.Transportation Research Board, Washington,
DC, September 1996.
122. Reid, J. D. New Breakaway Mailbox Designed Using Nonlinear Finite Element Analysis. ln Finile
Elements in Analysis and Design,Yol.32, No. l. Elsevier Ltd., Kiplington Oxford, UK, March
1999, pp.3749.
123. Reid, J. D. and G. W. Paulsen. Design and Simulation of Large Breakaryay Signs. ln Journal of
Tt'ansportation Engineering,Yol.l24, No. l. American Society of Civil Engineers, Reston, VA,
l 998.
I 30. Ross, H. E., Jr., et al. Safety Treatment of Roadside Drainage Strucflrres. ln Transportation
Research Record 8ó8. Transportation Research Board, Vy'ashington, DC, 1982-
l31 . Rowhani, P., D. Glaluz, and R. L. S. Stoughton. Vehicle Crash Tests of Concrete Median Barrier
Retrofitted With Slipformed Concrete Glare Screen.lnTransportation Research Record 1419.
Transportation Research Board, Washington, DC, 1993.
132. SAE International. Collision Deformation Classification, Standard J224 201105. Society of
Automotive Engineers, New York, NY 201 l.
133. SAE International. On-Highway Vehicles and Off-Highway Machinery Volume 4 of 1986 SAE
Handbook. Society ofAutomotive Engineers, Warrendale, PA, 1986. No longer available.
134. Salmon, C. G. and J. E. John. Steel Structures. Harper & Row Publishers, New York, NY 1990.
137. Sicking, D. L., S. Hemsorach, and R. P. Bligh. Critical Impact Locations for Longitudinal Ban'iers.
ln Journal of Transportafion Engineering,Yol 124, No. 1 . American Society of Civil Engineers,
Reston, VA, January-February 1998, pp. 65-12.
138. Sicking, D.L., J. R. Rohde, and J. D. Reid. Design and Development of Steel Breakaway Posts. In
'Washington,
Transportation Research Record I T20.Transportation Research Board, DC, 2000.
139. Sicking, D. L. and H. E. Ross, Jr. Structulal Optimization of Strong Post W-beam Guardì'ail. In
Transportation Research Record 1133. Transportation Research Board, Washington, DC, 1987.
140. SIS. European Standard SS-EN l3l7-1, Road restraint systems-Part l: Terminology and general
criteria for test methods. European Committee for Standardization, Ref. No. SS-EN 1317-l:2010
E. Swedish Standards Institution, Stockholm, Sweden, July 2010.
l4l. SIS. European Standard SS-EN l3l7-2, Road restraint systems-Paft 2: Performance classes,
impact test acceptance criteria, and test methods for safety barriers including vehicle parapets.
European Committee for Standardizalion, Ref. No. SS-EN l3l7-2:2010 E. Swedísh Standards
Institution, Stockholm, Sweden, July 2010.
142. SIS. European Standard SS-EN l3l7-3, Road restraint systems-Part 3: Performance classes,
impact test acceptance criteria, and test methods for crash cushions. European Committee for
Standardization, Ref. No. SS-EN l3l7-3:2010 E. Swedish Standards Institution, Stockholm,
Sweden, July 2010.
143. Snyder, R. G. State-of-the-Art-Human Impact Tolerances. SAE 700398 (rev. August 1970); re-
printed from 1970 International Automobile Safety Conference Compendium,May 1970.
144. Solomon, D. and H. Boyd. Model Procedure for Jn-Service Evqluation of Roadside Sof"ty
Hardwqre Devices. Report No. FHV/A-IP-86-8. Federal Highway Administration, U.S.
Department of Transportation, Washington, DC, April 1986.
145. Solomon, P. L. The Simulation Model of Autotnobile Collisions (SùAC) Operator's Manual.
National Highway Traffic Safety Administration, Washington, DC, October 1974.
146. Stolle, C. S. Cable Median Barrier Failure Analysis and Remediation. University of Nebraska-
Lincoln, PhD Disserlation, Lincoln, Nebraska, December 2012.
147. Stolle, C. S. and D. L. Sicking. Cable Median Bqruier Failure Analysis and Prevention.Midwes|
Roadside Safety Facility, University of Nebraska, Lincoln, NE, 2012.
148 StoLrghton, R. L., J. R. Stoker, and E. F. Nordlin. Vehicular Impact Tests of a Truck Mounted
Attenuator Containing Vermiculite Concrete Cel/s. Report 33001-609936. California Department
of Transportation, Sacramento, CA, June 1980
149. Stout, D., J. Hinch, and D. Sawyer. Guardrail Testing Progrant. Final Report on Contract
DTFHTI-87-C-00002, Ensco Inc., Springfield, VA, June 1990.
150. Stout, D., J. Hinch, and T. L. Yang. Force-Defiection Characleristics of Guardrqíl Posts.Final
Report on Contract DTFH6I-85-C-00099. Ensco Inc., Springfield, VA, Septernber 1988.
l5l . TASS. MADYMO. Computer simulation program developed by TASS, 38701 Seven Mile Road,
Suite 260, Livonia, MI48152.
152. Thomson, R. and J. Valtonen. Vehicle Impacts in V-Shaped Ditches. ln Transportation Reseqrch
Record I 797. Transpoftation Research Board, Washington, DC, 2002.
153. TRB. Proposed Full-Scale Testing Procedures for Guardrails.ln Highway Research Board
Circulqr 4S2.Transportation Research Board, Washington, DC, September 1962.
154. TRB. Recommended Procedures for Vehicle Crash Testing of Highway Appuftances. In
Transportation Research Circular 191. Transpoftation Research Board, Washington, DC, 1978.
I 55. TTI. Test Risk Assessment Program (TRAP) Version 2.2, User's Manual. Developed by CAPSHER
Technology, Inc. for Texas Transportation Institute, Texas A&M Univelsity, College Station,
Texas, July 2002.
156. U.S. DOT.Yehicle Measurement Techniques. National Automotive Sampling System (NASS),
Transportation Safety Institute, U.S. Depaftment of Transpoftation, Oklahoma City, OK,
September 29,1998.
157. UMTRI . Delphi V-Forecast and Analysis of the U.S. Automotiye Industry through the Year 2000.
University of Michigan Transportation Research Institute, Ann Arbor, MI, July 1989.
158. Viano, D. C., and I. V. Lau. Biomechanics of Impact Injury. Research Publication GMR-6894.
General Motols Research Laboratories,'Warren, Michigan, December 1 989.
159. V/ards. Wards Automotive Yearbook 2002. Wards Communications, Southfield, MI,2002.
160. Weaver; G. D., E. L. Marquis, and R. M. Olson. National Cooperative Highway Research Program
Report 158: The Relation of Side Slope Design to Highwøy Safety. NCHRR Transpoftation
Research Board, Washington, DC, 197 5.
161. Zimmer R. R. and D. L. Ivey. Influence of Roodway Surface Holes' on the Potentictl for Vehicle
Loss of Control. Research Report 328-2F. Texas Transportation Institute, Texas A&M University,
College Station, TX, August 1983.