You are on page 1of 176

STP 1434

The Use of Glass in Buildings

VaIerie L. Block, editor

ASTM Stock Number: STP1434

ASTM International 100 Barr Harbor Drive PO Box C700 West Conshohocken, PA 19428-2959 Printed in the U.S.A.

Library of Congress Cataloging-in-Publication Data

Symposium on the Use of Glass in Buildings (1st : 2002 : Pittsburgh, Pa.) The use of glass in buildings/[edited by] Valerie L. Block. p. cm.--ASTM special technical publication; 1434 Includes bibliographical references and index. "ASTM stock number: STP1434." ISBN 0-8031-3458-4 1. Glass construction--Congresses. 2. Glazing--Congresses. 3. Safety glass--Congresses. I. Block, Valerie L., 1951- II. Title. TH1560 .S96 2002 691'.6--dc21 2002038238

Copyright 9 2002 ASTM International, West Conshohocken, PA. All rights reserved. This material may not be reproduced or copied, in whole or in part, in any printed, mechanical, electronic, film, or other distribution and storage media, without the written consent of the publisher.

Photocopy Rights Authorization to PhOtocopy items for internal, personal, or educational classroom use, or the internal, personal, or educational classroom use of specific clients, is granted by ASTM International (/L~TM) provided that the appropriate fee is paid to the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923; Tel: 978-750-8400; online: http://

Peer Review Policy
Each paper published in this volume was evaluated by two peer reviewers and at least one editor. The authors addressed all of the reviewers' comments to the satisfaction of both the technical editor(s) and the ASTM International Committee on Publications. To make technical information available as quickly as possible, the peer-reviewed papers in this publication were prepared "camera-ready" as submitted by the authors. The quality of the papers in this publication reflects not only the obvious efforts of the authors and the technical editor(s), but also the work of the peer reviewers. In keeping with long-standing publication practices, ASTM International maintains the anonymity of the peer reviewers. The ASTM International Committee on Publications acknowledges with appreciation their dedication and contribution of time and effort on behalf of ASTM International.

Printed in Bridgeport, NJ December 2002

The symposium chair of this publication was Valerie L. Pennsylvania on 14 April. . ASTM International Committee E06 on Performance of Buildings served as its sponsor. Block. 2002.Foreword The Symposium on The Use of Glass in Buildings was held in Pittsburgh.

C.Based Stress Measuring System for On-line Quality Control of Tempered and Heat. WISE AND BIPIN V.Contents Overview vii SESSION I: QUALITY ISSUES ASTM C 1036: Does It Work for Field Inspections of Surface Blemishes?-TED W. AND ANAND JAGOTA 57 Development of Design Methodology for Rectangular Glass Supported on Three Sides to Resist Lateral Uniformity Distributed Loads-MOSTAFA M.Strengthened GlasS---ALEXS. EL-SHAMI.CleaningGlass---CHRISTOPHER J. EL-SHAM! AND H. RYAN JACKSON. and Beyond-VALERIE L.. SHAH 49 SESSION III: GLASS DESIGN Structural Performance of Laminated Glass Made with a "Stiff" Interlayer-STEPHEN J. SCOTT NORVILLE 66 79 Wind Load Resistance of Large Trapezoidal Glass Lites--H. ALEX VAN DUSER.. MAZULA AND IVAR HENNINGS Codes and Standards Affecting Glass in Buildings: The U. 20 PC. BLOCK The 8 BARRY AND THOMAS O'DAY Impact of Serf. AND GEORGE JOHNSON . MOSTAFA M. REDNER 26 SESSION II" PERFORMANCE ASSESSMENTS In-Situ Dew-Point Measurement to Assess Life Span of Insulating Glass U n i t s . WERNER LICHTENBERGER. ANTHONY SMITH. BENNISGN. AND ALLAN MAJOR 35 Evaluation of the Condensation Resistance Rating as Determined Using the NFRC 500 Progedure----DAN1EL J.S. scoyr NORVILLE.G E O R G E R. TOROK.

AND JEFFREY C. PILCHER. SCIAUDONE 121 Testing of Annealed Glass With Anchored-Film Glass Retention Systems for Fallout Protection after Thermal Stress Cracking--BRUCE S.Resistant Laminated Glass--H. AND ROGER E. LYNN BEASON AND A.. SCHMIDT. BEERS. PEARSON. KASKEL. MARK K. WILLIAM LINGNELL 90 105 SESSION IV: GLASS IN HURRICANES Retrofitting Commercial Structures with Laminated Glass to Withstand Hurricane E f f e c t s . PELLETIER 131 SESSION V: GLASS FOR FIRE SAFETY AND SECURITY The Advantages of Glazing in Overall Security Strategy--MiCHAELBETTEN AND HENRI BERUBE 147 153 The Relationship Between Sprinkler Systems and GlasS--JERRY RAZWICK Design Procedure for Blast. MARK A. JOHN E. MORSE A Thermal Stress Evaluation Procedure for Monolithic Annealed Glass-W. SCOTTNORVILLE AND EDWARD J.P A U L E.Window Glass Design Software--STEPHEN M. CONRATH 159 171 Index .

Another paper examined the interrelationship between building codes and glass standards. A second paper discussed testing of annealed glass with anchored-film glass retention systems. and glass for fire safety and security were presented. one speaker addressed retrofitting commercial structures with laminated glass to withstand hurricane effects. April 14. vii . Architectural glass was the broad focus for this symposium. A second presented an assessment of annual energy consumption of ventilated double glass facades using computer simulation. A fourth paper assessed the impact of self-cleaning glass. Glass in Hurricanes Glass used in hurricane-prone areas requires special design consideration. large trapezoidal window glass lites. performance assessments. A third paper discussed an on-line quality control measuring system for tempered and heatstrengthened glass. Pittsburgh. One paper examined the structural performance of laminated 'glass made with stiff interlayers. One paper focused on the problems associated with the use of ASTM C1036 for field inspections of glass. Papers discussing quality issues. A third paper focused on the evaluation of a condensation resistance rating as determined using the National Fenestration Rating Council (NFRC) 500 procedure. and window glass design software based on ASTM El300. The papers contained in this publication represent the commitment of the ASTM E-06. use. Several papers dealt with design methodologies for glass. One paper discussed in-situ dew point testing to assess life span of insulating glass units. Glass Design A series of papers were presented on glass design. PA. Emphasis on glass design was also a key feature to the symposium. and performance of architectural glass.Overview This book represents the work of numerous authors at the first Symposium on the Use of Glass in Buildings. In this session. Performance Assessments The intent of this section was to present developments around the performance of insulating glass and glass facades. glass design glass in hurricane-prone areas.51 subcommittee to providing timely and comprehensive information on glass used in buildings. Quality Issues Quality issues were addressed from several points of view. Papers and presentations were targeted to deliver information the user may find useful related to the quality. Another paper introduced a new procedure for thermal stress evaluation of monolithic glass. Common themes throughout the tenure of this symposium can be found in this issue. 2002. design. The symposium had a broad focus that incorporated a variety of glass-related topics. including rectangular window glass supported on three sides.

fire rated glass and sprinklers.viii THE USE OF GLASS IN BUILDINGS Glass for Fire Safety and Security This section was developed to cover a broad spectrum of topics. PA . Valerie Block Narberth. and a design procedure for blast resistant laminated glass. including security glazing. Ms.


glass inspection. Glazing Consultants. parts of ASTM C1036 are helpful in defining acceptable scratch criteria. Standard Specification for Flat Glass to inspect the glass. FL. this standard is not intended for use in the field. 329. PA. FL. Keywords: damaged glass. Glazing Consultants. and often turn to ASTM C1036.. 1Associate Consultant. ASTM STP 1434.. glass storage Introduction Inspecting scratched glass in the field is far from an exact science. Inc. . Fort Myers. Block. "ASTM C 1036: Does It Work for Field Inspections of Surface Blemishes?" The Use of Glass in Buildings. Inc. When a project has damaged glass. Standard Specification for Flat Glass. and other select glazing applications. 1325 Rotonda Point. and Hennings. ASTM International. However. glass specification. the parties look for an industry quality standard. Ed. This level of quality is recommended for architectural applications including reflective and low emissivity coated glass products. and in lieu of any other field inspection standards.Ted W.. Under ASTM C1036. It is the most commonly specified quality of glass in the industry I and refers to Table 4 criteria (Figure 1) for the maximum allowable blemishes for 6. Mazula I and Ivar Hennings 2 ASTM C 1036: Does It Work for Field Inspections of Surface Blemishes? References: Mazula.) or less glass thickness. 32746. 33913. Glass quality is addressed in ASTM C1036.. Abstract: Glass can be damaged after installation. It is useful for the proper specification of glass quality. and often the home or building owner is left trying to determine if the resulting surface damage is acceptable.0 mm (1/4 in. I. West Conshohocken.astm. 3 Copyright9 by ASTM International www. scratched glass. It is quite common for the project specifications to overlook the type of scratches that are acceptable. Ste. V. Lake Mary. medium-intensity scratches are allowed for glass quality Glazing Select (Q3). 2002. The owner and contractor are both exposed to risk in this situation. 2Vice President. 11910 Cypress Links Drive.W.

) Q3 Quality 3 Di~tribr Allowed A/lowed Allowed Allowed Allowed with a minimum separation of 600mm (24in.) to 0.4 TFIE USE OF GLASS IN BUILDINGS ASTM C 1036-01 Table 4 Allowable Linear Blemish Size and Distribution for Cut Size and Stock Sheet Qualities Thicknesses 6.) Heavy Medium Light Faint Figure 1 . or other uniform diffused background lighting that simulates daylight.0 mm (1/2 in.3 meters ( 132 in. Allowable blemishes for glass thicker than 12.) to 1.) Heavy < 150 ram (6in.) None Allowed Glass thicker than 6. with a .) 1 meter (39 in.) None Allowed None Allowed Heaw > 150 mm f~iin. Table 4 does not apply to glass thicker than 12. Rubs.3 meters (132 in.5 for detection of linear blemishes.) Less than 0.) or LessA Linear Blemish SizeB Intensity Length Faint < 75ram (3in. and Other Similar Blemishes) as follows: Place samples in a vertical position to the viewer.5 Detection for Linear Blemishes (Scratches.) Light < 75 mm (3in.2 meters (8 in.0 ram (t/2 in.Example of Table 4 Table 4 defaults to allow m e d i u m intensity scratches that are 75 m m (3 in.) from specimen and look through the sample at an angle o f 90 ~ (perpendicular) to the surface using daylight (without direct sunlight).) Medium > 75 mm (3in. Digs. The inspection is to be conducted per item 6.0 mm (l/4in.01 meters (40 in.1. Table 4 Blemisll lnt~p~ty Chart (continued) Deteetigl~Dist~ce Blemish Intensity Over 3.) Light > 75 mm (3in.) apart. n See 6.) Medium _<75 mm (3in.) Faint > 75 mm Oin.) and less than or equal to 12. The viewer shall stand approximately 4 m (160 in.1.) long providing any two scratches are not less than 609 m m (24 in.).0 mm (1/4 in.2 meters (8 in.) may contain proportionally more and longer blemishes.) shall be determined by agreement between the buyer and the seller.) 3.0 mm (1/2 in.

The parties review the contract documents to see if the subject of glass quality has been addressed. The viewer shall move towards the specimen until a blemish is detected (if any). and expense to resolve the situation. Homeowners do not want to look through scratched glass. Blemish Length is determined by measuring the perpendicular distance between the ends of the blemish. He or she will probably inspect the glass from a much closer distance and in direct sunlight (Figure 2). There are. The owner's first thought is to complain to the developer. it is still utilized to inspect installed glass. The end result is that ASTM C1036 is found to be unsuitable and all parties may be forced to expend considerable time. however. some inherent problems with . As its title suggests.~ Figure 2 . If there are a large number of windows and/or doors with reported scratches.Typical Surface Blemish (Scratch) Drawbacks to Using ASTM C1036 in the Field Despite the difficulties of using ASTM C1036 in the field. It is recognized that the industry has used segments of the procedures outlined in the standard for inspection on glazing systems installed in the field. Blemish intensity is determined by comparing the Detection Distance to the Blemish Intensity Chart at the bottom of Table 4. effort. In any event. The distance from the viewer to glass surface when the blemish is In'st detectable is defined as the Detection Distance. who then calls the contractor for warranty service. When the sun is setting. In extreme cases. the homeowner does not want to inspect the glass at 3. the homeowner may even contact a glass expert to inspect the glass and help solve the problem. the cost of replacement may be substantial. especially if they have paid top dollar for a condominium overlooking the ocean. These documents typically establish glass quality as Q3 from ASTM C 1036 or do not address the issue at all.3 m (132 in.) with uniform light as required by the standard. even a small scratch in a patio door or window can be disturbing.MAZULA AND HENNINGS ON ASTM 1036 5 minimum illuminate of 160 foot-candles. the standard provides more of a guideline for "specifying" glass than it does for "field inspecting" glass.

Third. handling.e. class. tall will view the glass differently than an inspector at 5 ft. type. This process also establishes a post installation time-line.-4 in. Suggested Procedures Quality of glass and the manner in which glass is to be inspected should be specified prior to the construction process." To achieve a higher quality on projects. even on the same piece of glass. Fourth. tall due to the geometry of the viewing angle.6 THE USE OF GLASS IN BUILDINGS these guidelines.3 m (132 in. the procedures to inspect the glass allow significant latitude. form. which can help identify the point at which damage occurred to the glass. Flat) shall not be greater than those listed in Table 4. This requires inspection upon receipt of the product from the manufacturer to the project. cleaning prior to inspection may not be possible.) at the exterior of the specimen can either draw to or detract attention from the scratch in question. style. and finish). For example. These conditions will undoubtedly vary from building to building.) distance required for the inspection. Blemishes for Type I (Transparent Glass. therefore. Specifiers need to avoid simply referencing the ASTM C 1036 "Standard Specification for Flat Glass" in general terms. Flat).) thick clear glass product can be represented in Specifications as follows: "Type I . existing interior conditions perpendicular to the specimen may not provide the mandatory 3. which ot~en results in contradictory conclusions by separate inspectors. quality. and installation. proactive steps should be taken by inspecting the glass at key points in the construction schedule to identify if glass damage is present. First of all.-8 in. weather conditions. This process assists in identifying damage that can occur during delivery. They should scrutinize the ASTM C 1036 Standard to indicate the glass classification (i. Second. Glazing Select Quality (Q3) . The cost to include glass quality and field inspection guidelines in the project specification manual is minimal and is recommended for all . etc. the natural background (trees.(Transparent Glass. Implementation of inspection "sign-off sheets" for the glass and glazing system should be completed and dated immediately after installation. storage.intended for architectural applications including reflective and low emissivity coated glass products. Evaluation of this data can reduce the number of trades that may have been performing work in the immediate area where damage took place. and other select glazing applications. The typical 6 mm (% in. Implementation of a quality control program to inspect the glass during the product's life cycle from manufacturing through installation is beneficial in detecting surface damage. In addition to tightening the specifications. an inspector that is 6 t~. Finally. Inspection without consideration of cleaning the exterior glass could skew the results. adjacent buildings. Class 1 Clear. requiring the inspector to view "through" the glass as defined in the standard and detect a scratch is extremely subjective and creates discord among the concerned parties. specifiers should consider specifying Select Quality (Q3) adding criteria as follows: Glass surfaces with detectable linear blemishes that exceed Light Intensity will not be accepted (refer to Table 4 in the standard). fixed glass specimens located on shear walls may not be accessible from the exterior.

Consideration of the project size. This document should provide a clear outline with fair and consistent inspection procedures and evaluation criteria to represent all parties (developers. contractors and owners). p. 73. References [1] Gana . it does not meet the needs of the industry for field inspections to evaluate damaged glass. A new document is needed that will specifically address the field inspection of glass for damage. Furthermore. frequent inspections from receipt of glass to installation are important in monitoring surface damage. and access for replacing glass are key components in establishing a glass inspection program. 1997 Edition. manufacturers. The costs to setup and implement a daily glass inspection schedule may be substantial and should be considered on an individual project basis. type of glass. In the meantime. this document should address all relevant field conditions and eliminate as much subjectivity as possible.MAZULA AND HENNINGS ON ASTM 1036 7 projects. . GlazingManual.Glass Association of North America. however. Conclusion ASTM C1036 is useful in specifying glass.

"Codes and Standards Affecting Glass in Buildings: The U. and glass strength. . and Beyond. KS 66614-5321 8 Copyright9 by ASTM International www.S. Although the development process is different. V. an energy standard that includes building envelope requirements for commercial and high-rise residential glazing. and Air-conditioning Engineers (ASHRAE).D.B. has produced ASHRAE 90. and practices through ASTM International and the American National Standards Institute (ANSI). this paper will review the work of ISO Technical Committee 160. skylights. the National Fenestration Rating Council (NFRC). In many instances.building codes. ASTM International. and skylights. West Conshohocken.1 standard guide the designer's selection of windows." The Use of Glass in Buildings. doors.S. Suite A. the adopted standards clarify and enhance the use of glass in building construction. To verify performance..standardsarc referencedor included in the buildingcodes. handrails. glass requirements for safety glazing performance and glass quality will be compared to existing Mexican and Canadian requirements. Over the past twenty years. manufacturers and building i Technical Director. Topeka. Primary Glass Manufacturers Council.and glassquality. current activities. Keywords: Glass. Refrigerating. L. The fenestration performance requirements for thermal transmission (Ufactor) and Solar Heat Gain Coefficient in the 90. Eds. American Society for Heating. 2002.Valerie L..internationalstandards. and the International Organization for Standardization (ISO) will be reviewed. A. There arc other cases where standardshave lead to federal regulations. and the interrelationship of national and international standards in the workplace. such as the American Society for Testing and Materials (ASTM). for example. Jones. Smith and C. At the international level.S. Specific U. Issues covered include safety glazing. Introduction Building codes and standardsgo through specific development and adoption processes in the United States. performance and quality specifications. and Air-conditioning Engineers (ASHRAE). 2945 SW Wanamaker Drive.astm. Refrigerating. The American Society for Heating. American National Standards Institute (ANSI).1. Abstract: This paper examines the development and adoption process of building code requirements and standards related to glass in buildings in the United States. Block 1 Codes and Standards Affecting Glass in Buildings: The U. PA. and Beyond Reference: Block. The glass industryhas developed testmethods. ASTM STP 1434. The relationship between the building codes and consensus-based standard organizations. its working groups. other organizations have developed standards that impact glass.

[1] . the three regional code organizations united in 1994 to form the International Code Council (ICC). At the hearings.BLOCK ON CODES AND STANDARDS 9 code officials look to the National Fenestration Rating Council (NFRC). they have no vested interest in any specific building product. They were the Building Officials and Code Administrators (BOCA). In 1998. often times. These members consist of public building and fire officials from local communities across the country. testing. as well as a certification and labeling program to assure compliance. they are no longer being updated and the regional building code organizations actively promote state adoption of the new I-Codes. Standards writing. and certification activities have occurred in other countries and at the international level. Building Codes For many years. Because the code groups themselves recognized the duplication and. This paper will explore the interrelationship between the building codes and industry standards in order to establish their significance in building construction. The ICC is a nonprofit organization dedicated to developing a single set of comprehensive and coordinated national codes that identify minimum health. three regional building code organizations developed and published building codes in the United States. individuals can speak for or against a proposal. a national organization that has developed measurement standards for fenestration. the International Conference of Building Officials (ICBO).S. safety. Public hearings are held to give individuals and organizations access to the code development process. These regional code organizations developed "model" building codes that could be adopted by a state or used as a model for a state-developed building code. the ICC published its first set of construction codes that included: 9 International Building Code 9 International Residential Code 9 International Electrical Code 9 International Mechanical Code 9 International Fire Code 9 International Plumbing Code 9 International Energy Conservation Code While the three model codes still exist. confusion in building code requirements from one model code to another. As part of the code development process. U. Committee recommendations are sent to the ICC code official members for ratification and a final vote. and the Southern Building Code Congress International (SBCCI). As impartial officials. interested individuals may submit new code proposals and code change proposals. and general welfare standards.

and dead loads on glass. screening Vertical and slopedglazing.multiqight . the building official is able to confirm the load resistance of glass. the federal regulation. Voluntary standards that are referenced or excerpted in the code become mandatory once the code has been adopted. and the IECC includes requirements for thermal transmission (U-factor). Federal regulations and other consensus standards are often referenced in building codes. twelve nonfactored load charts are reprinted from ASTM E1300. parts of the standards are included verbatim in the code. Chapter 24 Glass and Glazing. It is important to note that until the building codes are supports. (Table 1). within Section 2406.. Identification(labeling). and air leakage. are referenced. through its code consultants. light-transmittingceramicand light-transmitting plastic panels. and glass in floors and sidewalks. The requirements of the IBC are diverse and focus on wind.10 THE USE OF GLASS IN BUILDINGS Glass requirements are found in Chapter 24 of the International Building Code (IBC). The Glazing Industry Code Committee (GICC) has represented the interests of the U. Both of these standards contain test methods to evaluate the safe performance of glass. glass in handrails and guards. safety glazing. In some cases. sloped glazing and skylights. in Section 2405 of the IBC on Sloped Glazing and Skylights. factors Human impactloads. For instance. Visible Transmittance. ASTM E1886 and E1996 are referenced under Protection of Openings. Standard Practice for Determining Load Resistance of Glass in Buildings. Section R308 of the International Residential Code (IRC). In Section 1609. glazing in athletic facilities. The glass and fenestration industries have a voice in the code process via their trade associations or on an individual company basis. The IRC specifically addresses safety glazing and skylights (Table 2). identificationof safetyglazing. Each chart covers a specific thickness of monolithic annealed glass. Safety Glazing. glazingreplacement DaUeGlass. For example. of the International Building Code.S. they have no basis in law and are much like any other standards that may or may not be referenced by the designer of a building. framing.2000 International Building Code. louveredwindowsor jalousies Vertical glass Allowable glazingmaterialsand limitations.4 of the International Building Code. I. ANSI Z97. snow. etc. General Definitions General Requirements Wind. The final part of the building code process is adoption through the appropriate local or state legislative and administrative processes. Solar Heat Gain Coefficient.interiorglazed areas. TABLE 1 . and in the International Energy Conservation Code (IECC). By using the charts and appropriate factors for single and insulating glasses. and the American National Standard. has actually participated in writing the chapters on glass in the codes.1. snowand dead loads on glass Sloped glazingand skylights Non-factoredload charts SafetyGlazing Coversglass. CPSC 16 CFR 1201. DecorativeGlass. These ASTM standards give testing information and use parameters for fenestration and storm shelters installed in hurricane-prone areas. glass and fenestration industries for many years and.

users.S. wired glass prohibited with wire exposed on longitudinal edges Same as IBC Must complywith 2404 of IBC Definition. The American National Standards Institute was founded in 1918 with the purpose of ensuring that U. laminated glass. screens.S. others can have "16 CFR 1201" No thinner than 4. The approval process verifies that the principles of openness and due process have been followed and that a consensus of all interested parties has been reached.S.2000 International Residential Code. and conflict. curbs for skylights Identificationof multipane assemblies Louveredwindows or jalousies Hazardous locations Site built windows Skylights and sloped glazings U. They solve issues of product compatibility and address consumer safety and health concerns. parking garages Testing Design loads. "Standards also allow for the systemic elimination of non-value-added product differences (thereby increasing a user's ability to compare competing products). permitted materials. loads.76mm(3/16 inch).BLOCK ON CODES AND STANDARDS 11 Glass in handrails and guards Glazing in athletic facilities Glass in Floors and Sidewalks assemblies. and oRen simplify product development. are the American National Standards Institute (ANSI) and the ASTM International. An appeals process and a requirement for balance assure that no one interest can manipulate the process unfairly. support. duplication of efforts. screen characteristics. screens not required."[2] Two of the most important standards organizations in the U. reduce costs. glass in greenhouses. Section R308 Glazing.[3] Standards consider the needs of producers. desi~ formula TABLE 2 .. According to Amy Marasco. and other interest groups. hazardous locations. tempered glass must have permanent label One pane fully labeled. no longerthan 1219 mm (48 inches) . Identification Permanent label required for safetyglazingunless building code official approvesof certificate. fire departmentaccess panels Materials. Standards There are thousands of voluntary standards in the United States that benefit both the manufacturer and user of products. The ANSI process is based on determining whether a standard meets the necessary criteria to be approved as an American National Standard. . tempered spandrel may have removalpaper label. voluntary standards minimize waste.

Dudley issued standard material specifications for the company's suppliers of oil. "ASTM's method of developing standards is based on consensus without borders. product users. (Fig. although they may attend meetings. They are typically other committees within ISO with related interests or other international organizations.S. steel. Liaison members have no power of vote. paint. ISO is a worldwide federation of national standards bodies from some 140 countries. These standards are beneficial because they often reduce time-to-market and lower product development costs. ANSI interfaces at the international level. In 2002. and other materials. and host meetings. but do not actively participate. To alleviate problems. or Liaison members. a chemist with the Pennsylvania Railroad. ANSI offers U. Our process ensures that interested individuals and organizations representing academia. regional level. engineering standards that cover product design.12 TFIE USE OF GLASS IN BUILDINGS ASTM traces its roots back to the 19th century and the driving force of Charles Dudley. In addition. one from each country.. In addition. offer candidates for leadership positions."[5] International Standards In the 1980s and 1990s. The glass industry is actively developing international standards through the International Organization for Standardization (ISO). global standards facilitate the introduction of products to abroad range of countries. They have no power of vote within the committee. including developing countries. and national level with a variety of standards organizations.J7] National bodies are characterized as Participating (P) members. "O" members monitor the technical work. They send delegates to meetings. and government alike all have an equal vote in determining a standard's content.S. Participants are welcome from anywhere on the globe. Observer (O). According to ASTM President Jim Thomas. but are able to attend meetings and receive documents. ANSI is responsible for submitting the U. The United States is a "P" member of many ISO Technical Committees through ANSI. he organized technical committees to discuss the specifications and testing procedures and form consensus.[4] Today there are more than 80 technical committees involved in a wide range of activities. "P" members are responsible for submitting votes and/or comments on all technical matters coming before the committee. 1).S. While there are other national standards organizations in the U.[6] Consumers also benefit from products that are safer and of a higher quality. ASTM standards include material standards that cover quality of a building product. and testing standards covering the product performance. industry. a non-gnvemmental organization established in 1947. global manufacturers began to demand international standards to minimize confusion caused f~om many proprietary and regional standards. participants training and support on international procedures and standards writing. . ASTM changed its name to ASTM International and its focus from that of a national standards organization to one with an international scope. vote on ballots issued by ISO.

. Legend AAMA ANSI ASCE BSI CEN - COPANT ETSI IBN IEC ISO NFPA - UNI American ArchitecturalManufacturers Association American National Standards Institute American Society for Civil Engineering British Standards Institute European Committee for Standardization Pan American Standards Commission European Telecommunications Standards Institute Institut Beige de Normalisation International Electrotechnical Commission International Organization for Standardization National Fire Protection Association Ente Nazionale Italiano di Unificazione .Standards Organizations. Standards Bodies I I I I io--..BLOCK ON CODES AND STANDARDS 13 Public Sector Private Sector .i Standards Bodies I Regional Standards Bodies ISO I AAMA I I COPANT IEC I I NFPA I ASTM ASCE I ron I CEN I UNI I I ETSI I I I Figure 1 .ANSI I I I Intl.

the U. but in many cases they are totally new standards developed through the expertise of working group members. and security glazing tests. has produced seven draft documents on security glazing tests. mirrors. Approval of a final standard is based on acceptance by a two-thirds majority of Pmembers voting and not more than a quarter of the total votes cast being negative. there are working groups on basic glass products. glazing subject to arena airblast load. These drafts include test methods and classifications for destructive windstorm resistant glazing material. bullet resistant glass. for example. Within the subcommittees are smaller working groups consisting of national experts who collaborate on the development of intemational standards. Experts are appointed from the TAG to represent the United States at interuational Working Group. In this way.S. a working draft must be approved within six months. Florida causing $20 billion in property damage. 36 months are allowed to take the draft through the ballot process to final publication. This means that there is general agreement.[8] Three other South Florida counties and areas along the Gulf Coast in Texas enacted similar requirements. laminated glass. one on glass properties and the other on glass uses. The . This catastrophic weather event resulted in a change to the South Florida Building Code requiring all external glazing material to be either capable of resisting windborne debris or to be protected by shutters. The glass industry participates and responds to ISO ballots through a Technical Advisory Group (TAG) that is administered by ASTM. There are fifteen TC 160 working groups. Hurricane Andrew hit Dade County. and Full Committee meetings oflSO TCI60. The ISO standards development process is based on consensus. and individuals can become members of the U. light and energy transmission properties and thermal properties of glazing. airborne sound insulation. Subcommittee. Within each group. organizations. One working group. Interested U. Each of these standards must eventually go through a ballot review process in order to become a recognized ISO standard. glass blocks and glass paver units. position is heard around the world. Under subcommittee one. coated glass.92 under ASTM C14 Glass and Glass Products. A Committee Draft is required 18 months after the working draft has been submitted for review. Once a New Work item has been approved. fire resistant glazed assemblies. Once a Committee Draft has been reproduced. The technical committee (TC) has two subcommittees. work is underway to produce international standards. insulating glass units. toughened glass.S. These standards may contain requirements found in other national standards. Recognizing the Need for Codes and Standards On August 24. assembly rules and structural sealant glazing. and curved/bent glass. but does not imply unanimity.S. Another is working on a test method for safety glazing. 1992. Under subcommittee two. repetitive axe and manual attack). safety glazing tests. companies. working groups are developing standards on the design strength of glazing.14 THE USE OF GLASS IN BUILDINGS Glass interests are organized under ISO's Technical Committee (TC) 160. Glass in Buildings. explosion resistant glazing (shock tube loading). ISO adheres to established target dates. TAG. The TAG is recognized as Task Group C14. and forced-entry resistant glazing products (tests include repetitive ball drop.

Federal Standards In the late 1960s. ISO TC 160 SC 1 Working Group 1 has reviewed the CEN. U. Despite its adoption as an industry standard.S.J9] Though unlike ANSI Z97. Doors and Storm Shutters Impacted by Windborne Debris in Hurricanes (El 996). This resulted in two standards. The standard. like the voluntary ANSI standard. the standards will be consistent and accurate. Curtain Walls. This standard was initiated by the glass industry as a means of reducing glass-related injuries. the Consumer Product Safety Commission (CPSC). in 1981. the three regional model code bodies enacted safety glazing provisions for all hazardous location applications conforming to CPSC standards. ASTM Test Method for Performance of Exterior Windows. When .1. Curtain Walls. with the power to promulgate consumer product standards. Once these standards were adopted as consensus standards. One such standard is the American National Standard Z97. By allowing input from key producers. Initially. the Consumer Products Safety Act was passed. the CPSC standard applied to glazing in doors and other glazed panels in hazardous locations. CPSC 16 CFR 1201 was mandatory for all parts of the United States. subject to human impact. Doors and Storm Shutters Impacted by Missile(s) and Exposed to Cyclic Pressure Differentials (El 886) and ASTM Specification for Performance of Exterior Windows. With the support of industry. the federal commission granted a petition to develop a federal safety standard for architectural glass. labor. such as sidelites and panels adjacent to walkways. these international standards will facilitate international trade and communication by defining clear and unambiguous provisions. Since basic glass products are sold around the world. that were consistent with the regulations of CPSC. the CPSC standard 16 CFR Part 1201 was enacted by the federal government. and will represent the state of the art in float glass production capabilities. and general interest groups.1. proponents moved them into the building code arena where they were eventually adopted into the International Building Code. State and local building code officials were expected to impose criteria for the use of glass. and Japanese quality standards in order to dratt an ISO standard on physical and mechanical properties of soda-lime silicate float glass and stock sized and cut sizes of fiat glass. Many industry standards were developed at this time. CPSC withdrew its glazed panel provisions to permit regulation and enforcement of glazings in those locations by state and local building code authorities. Soon after South Florida building officials adopted requirements for windborne debris. as ASTM Working Group organized to develop a consensus test method and specification that addressed requirements for glazing subjected to the severe effects of wind events. it became clear that it would only be effective if it were adopted as a regulation. establishing a federal commission. there was a pattern of social activism in the United States that gave rise to a grassroots consumer rights movement. was designed to reduce or eliminate the unreasonable risks of injury associated with architectural glazing materials. safety. and in 1977. However.BLOCK ON CODES AND STANDARDS 15 building officials used Australian data supplied by the glass industry to develop these windborne debris requirements. As required by the federal preemption mandate. In 1972. In the area of glass quality.

Productos de vidrio . the North American Experience Safety glazing regulations. 9 Glazing adjacent to stairways. and one or more walking surface(s) are within 36 inches horizontally of the plane of glazing. steam rooms. NOM-146-SCFI2001. but it also covers physical attack.S. 9 Glazing in any portion of a building wall enclosing these compartments.Vidrio de seguridad usado en la construcci6nEspecificaciones y m6todos de prueba. 9 Glazing in unframed swinging doors. bathtubs and showers. Today hazardous locations requiring labeled safety glazing materials in the model building codes are defined to include: [ 1O] * 9 Glazing in swinging doors except jalousies. It does establish a minimum level of safety in order to reduce the threat of injury to people from glass breakage caused by human impact. the exposed top edge is greater than 36 inches above the floor. standard. Safety Glazing and Glass Quality Standards. With the exception of polished wired glass. accidental and natural events. the scope of the Mexican standard is broader than the U. 9 Glazing in storm doors. and acts of aggression and vandalism. but only recently has Mexico adopted a safety glazing/quality standard for glass. (Table 3) . these requirements were incorporated into the new building code.16 THE USE OF GLASS IN BUILDINGS the three regional codes united to produce the Intemational Building Code. including structural baluster panels and nonstructural in-fill panels. The Canadian safety glazing standard is similar to the U. 9 Glazing in an individual fixed or operable panel adjacent to a door. 9 Glazing in individual fixed or operable panels where the exposed area of an individual pane is greater than nine square feet and the exposed bottom edge is less than 18 inches above the floor. saunas. Glazing in fixed and sliding panels of sliding patio door assemblies and panels in other doors. 9 Glazing in doors and enclosures of hot tubs. For one. including walk-in closets and wardrobes. The model codes refer to the Consumer Product Safety Commission standard 16 CFR Part 1201 for impact test parameters. standards and code requirements have been in existence for decades in the United States.S. standard. It is an official Mexican standard. landings and ramps. 9 Glazing in walls and fences enclosing indoor and outdoor swimming pools and spas. there are some notable differences. standards. whirlpools. 9 Glazing in guards and railings. Although many parts of the Mexican standard were modeled after U. all safety glazing products must meet Category I or Category II requirements of the CPSC standard.S.

GlassMagazine.0 m (29.0 m (19. **Norma Oficial Mexicana NOM-146-SCFI-2001.22 m (48 inch) 3-3.930 mm (34 x 76 inch) CAN/CGSB-12.8 ft) 4 = 6. I-M M-I for Cat I only . in.5 t~) 86.S.12. No break No open > 3 inch 10 largest particles < l0 sq.3 cm x 193 cm (34 • 76 inch) Included in safety standard Single impact Multiple impacts for Levels 3-5 Boil. Does not remain in frame and no break No break PERMANENT MARKINGS Legibly and permanently marked: Manufacturer name or logo Classification and/or designation of the glass. level of resistance "Made in Mexico" NOM. Productos de vidrio-Vidrio de segnridad usado r la construeci6n-Especificaciones y m~odos de prueba.3 Single impact ENVIRONMENTAL TEST FRAGMENTATION CENTER PUNCH NUMBER OF SPECIMENS Boil. alternate impacts ACCEPTANCE CRITERIA No opening > 3 inch/4 lb sphere can pass 10 largest particles < 10 sq.BLOCK ON CODES AND STANDARDS 17 TABLE 3 -- Comparisonof Safety GlazingRequirements.6 fl) 5 = 9. in. 60-61.8 square meter (9 SQ FT) Up to 865 x 1. CAN/CGSB 12. Legibly and permanently marked: Manufacturer name or logo CAN/CGSB. weathering None Not specified. in.1" CANADA Tempered or Laminated Glass Type l=Laminated Type 2=Tempered NOM-146-SCFI-2001** MEXICO Tempered or Laminated Glass Class l=Tempered Class 2=Laminated Type A=PVB Type B=Resin 1 = 0.0m (9.September 2001. pp. accelerated weathering Yes 3 COUNTRY PRODUCTS COVERED PRODUCT CLASS CPSC 16 CFR 1201' U.2 and 12. Safety Glazing None IMPACT CLASS Cat 1=18 inch drop < 9 square feet Cat 1I--48 inch drop SIZE CLASS QUALITY REFERENCE IMPACT TEST Largest manufactured up to 34 x 76 inch N/A Single impact Cat I = 460mm (18 inch) drop < .106-SCFI *Safety Glazing Testing Comparison. humidity.46 m (18 inch) 2 = 1. Does not remain in frame and no break No break Permanent label or paper certificate: Standard reference Date of manufacture Place of manufacture Name of manufacturer Boil Yes 4 Asymmetrical. for nonsymmetrical equal number of specimens from each side No opening>3 inch/4 lb sphere can pass 10 largest particles < 10 sq.220 mm (48 inch) drop > .8 square meter Cat II = 1.

Manufacturers test their products to determine thermal and solar performance. The Safety Glazing Certification Council (SGCC) offers manufacturers in the United States an opportunity to test. many North American companies have participated in this certification program. Certification programs are designed to assure high quality and performance. While the NFRC program is broader in scope than SGCC. 4. Participants in the NFRC program are required to follow specific labeling guidelines. and the voluntary ANSI Z97. it essentially provides the same benefits.I. 5. approving and registering the form of a Licensee's label. This ANSI-accredited Certification Program has given manufacturers a way to independently demonstrate product performance. 6. A permanent label usually identifies program compliance. The Hallmark Certification Program developed by the Window and Door Manufacturers Association (WDMA) consists of a series of inspections and tests to determine that products are being manufactured in the same way in which they were tested. and labeling of fenestration products. Part 1201. 2. 3. This legislation calls for all products used in buildings to satisfy six basic criteria: 1. independent testing laboratories. In other countries. 2 standard. The European community has embraced certification of products with its Construction Products Directive (CPD). Once testing information is available. NFRC standards are now referenced in the International Building Codes.1 safety standard. the government regulates the certification process. Health and Environment Safety in Use Protection Against Noise Energy and Heat Retention . The SGCC is a nonprofit corporation established in 1971 by manufacturers.18 THE USE OF GLASS IN BUILDINGS Certification Product certification demonstrates compliance with standards and regulations and is mandatory to place a product on the market in many countries. Mechanical Resistance and Stability Safety in Case of Fire Hygiene. and label products to both the federal safety glazing standard CPSC 16 CFR. building code officials and others interested in public safety. certification. product performance can be determined through simulation. AAMA's Certification Label tells customers that products have been verified as conforming to ANSI/AAMA/NWWDA 101. In some countries. certify. Trade associations have also recognized the value of certification programs in the United States. The Council is responsible for conducting independent routine sampling and the testing program.[11] Over the years.S. The National Fenestration Rating Council (NFRC) provides a framework for testing. manufacturers are able to selfcertify or test and certify through third party. The American Architectural Manufacturers Association (AAMA) has maintained a certification program since 1962. and withdrawing authority to use that label if products do not meet specifications. Products are evaluated by performance requirements in one of the WDMA standards or test methods. as well as in the ASHRAE energy standards.

[5] "A New Name A Longstanding Commitment. p. p." Glass ProcessingDays Conference Proceedings. 2002. [2] Morasco. procedures. [4] "Innovation by Consensus: ASTM's First Century.. "Consensus Codes-Does It Matter?" ICC Newsletter.68. [8] Smith. Tampere (Finland). D.715-718. As more products are sold globally.4. "Standards Development: Are You at Risk?" ASTMStandardization News. measurements of quality and many other important requirements or conditions to which a product or material must conform. "Hurricane Glazing Building Codes Continue to Evolve. September 2001. such as fire resistance. and the quality and performance of manufacturing in the United States. "The Effects of European Standardization on the Smaller Company. p. pp. and practices that define product usage.. A. p. performance and quality specifications. The CPD requires the product to have a CE mark in order to be "placed on the market.. they raise the bar on quality and safety of building construction." http://208. June 2001. Box 9. Safety Glazing Certification Council.22.. classifications. the number of international standards will increase. Together.htm. W. P. January. A. performance tests are required to demonstrate the ability of a product to perform a particular function. [ 10] For complete code requirements see Chapter 24 of the International Building Code. R. J. A. V.. [3]Morasco. [12] Colvin. J. [11] CertifiedProducts Directory." ASTMStandardization News. "Standards Development: Arc You at Risk?" ASTMStandardization News." Glass Magazine." ASTMStandardization News."[ 12] Conclusion Building codes and standards rely on test methods. June 2000..O.80/ANNIVERJconsensus.25. [7] "What is ISO?" International Organization for Standardization. July 2001. Building codes provide regulations for adoption and enforcement. p. [6] Sterling. June 2000. A.iso.22.27. June 2001.BLOCK ON CODES AND STANDARDS 19 In order to comply with the CPD. offering the same benefits and providing a common basis for product usage and understanding around the world.7. September 2001. "Going Global." Glass Magazine.html.. "What the Safety Regulations Are for Glass. as well as factory production control tests to demonstrate that the product continues to pass the required tests. p. http://www. New York 13651. Henderson Harbor.211. [9] Block. p. Industry standards provide definitions. L. p. References [I] Nickson. 2001.

V.. The former is composed of molecules containing a carbon atom that can be broken down by chemical reactions. hydrophobic. Toledo. 2 Sales and Marketing Dept. which admit daylight. OH 43697. PA. and reflective coatings. Inorganic materials are not broken down by chemical means but must be prevented from sticking to glass. photocatalytic lntroduc~on Self-cleaning glass has to deal with organic and inorganic dirt. Some of them can break down deposited organic dirt using a catalytic action powered by the ultraviolet component of daylight. Some can rinse inorganic dust offthe glass with rain or water by their hydrophobic. Toledo. or minerals and inorganic salts from lawn sprinklers. A third form of dirt on windows is seen when salt from sea spray. or not . "The Impact of Serf-Cleaning Glass. Pilkington North America Inc. At first there were the nearly invisible. Keywords: glass. These low-emissivity coatings also enhance the effectiveness of heat absorbing solar control tinted glass. 20 Copyright9 by ASTM International www. Ed. hydrophilic. J. These coatings act in different ways to prevent the deposition and build-up of dirt. West Conshohocken. T. 1Director of Technical Services. Some of them can also admit beneficial passive solar gain.. Use of Glass in Buildings.astm. low-emissivity coatings. Pilkington North America Inc. C. self-cleaning. ASTM International. OH 43697. or hydrophilic properties. Abstract: Today there is yet one more invisible coating available to improve the properties of window glass. 2002." ASTM STP 1434. or must be removed from it by breaking down the adhesive which holds it in place. Now clear. by preventing absorbed solar heat from radiating towards the room side of a window... self-cleaning coatings are available for the outer surface of the window. Barry. create deposits on glass as the water evaporates. Block. Inorganic dirt is found as dust and grit from fine earth or sand particles from road dust.. t and Thomas O'Day: The Impact of Self-Cleaning Glass Reference: Barry. long-wave (10 micrometer wavelength) infrared thermal radiation. These prevent winter heat loss by reflecting. and O'Day.Christopher J.

which help carry off dirt containing water drops. Aesthetic The continuous cleaning action of the self-cleaning window means that under normal weather conditions.BARRY AND O'DAY ON SELF-CLEANING GLASS 21 Clear coatings on glass for self-cleaning or dirt-resisting properties can be temporary. Under normal conditions one could expect the need for manual cleaning to he at least one half to one quarter as frequent as for plain glass. They found that in the United States there were 88 reported falling accidents involving window cleaners. the level of visible dirt can he expected to stabilize. of which 62 were fatalities. besides convenience. This is aided by the higher wind speeds for cars as compared to buildings. either by vacuum deposition or by the pyrolytic chemical vapor deposition (CVD) processes. where rain can occasionally rinse the glass. on installed glass. National Institute for Occupational Safety and Health. are as follows: Economic Benefits For commercial buildings where professional window cleaning services are used and their cost is known. it is very easy to calculate an economic case for self-cleaning glass. . or by attracting it (hydrophilic). in September 2000. Benefits of Self-Cleaning Glass The immediately obvious benefits. hand applied. This can be done either by repelling water (hydrophobic). or permanent. Water Repelling Automotive applications can use hydrophobic coatings which increase the wetting angle of water drops. was undertaken by the Department of Health and Human Services. This is in sharp contrast to ordinary glass where the level of dirt continues to accumulate until someone decides it's time to clean the windows. applied while the glass is being made. Self-Cleaning Methods Self-cleaning is partly effected by controlling the action of rain water. Safety A study of"Worker Deaths by Falls" [1] over a 15-year period. This paper addresses only the permanent dirt resisting and self-cleaning coatings and not those hand-applied solutions which need to he reapplied periodically.

or indirect as reflected from clouds. If it is necessary to rinse glass with hard water then a few drops of liquid dish washing detergent can be added as a surfactant to prevent droplet formation. But it is expected that the hydrophilic action of an activated coating will make it much easier to rinse off salt deposits with a hose. The action of solar ultraviolet light. glass and leave spots behind as the mineral content and inorganic particles are concentrated in one location. or mineral deposits from hard water in lawn sprinklers on a hydrophilic coating will not be different from ordinary glass because of the salts' inorganic nature. either direct. with a wetting angle in the 10 degree range. These are the droplets and rivulets which dry on ordinary. When water droplets coalesce on a vertical or sloped surface they form a sheet of water which slides down under the influence of gravity. garden spray bottle with plain distilled water from a hardware store. without droplets or rivulets.22 THE USE OF GLASS IN BUILDINGS 105~ a wetting ~ I Figure 1 . the action of a water attracting (hydrophilic) coating is very effective as a self-cleaning method. hand-pump pressurized.Hydrophobic Wetting Angle I Water Attracting Perhaps surprisingly. can charge a titanium dioxide (TiO2) coating by raising the outer electrons to a higher band. 1 or 2 gallon. This 'Invisible SqueegeerM' action leaves behind a relatively clean and dry glass surface. In its charged state it becomes hydrophilic. Mineral deposits from hard water should be prevented from building up. non-coated.Hydrophilic Wetting Angle The build-up of salt deposits from sea water spray. 10~ w I e t t i n ~ Figure 2 . The best method is to rinse off dust with a portable. . At the top edge of the water sheet a thin section shows interference colors as it slides down. pulling with it inorganic dirt particles.

reflected UV light from the sun. Applications of Self-Cleaning Glass in Buildings An invisible self-cleaning coating is now added to the list of available glass options to control window appearance. behind insect screens and under roof eaves.BARRY AND O'DAY ON BEEF-CLEANING GLASS 23 Photocatalyfic Breakdown A titanium dioxide (TiO2) coating acts. in a vacuum chamber. One manufacturer of self-cleaning glass calls their product "PhotoActivrM'' to illustrate this action. breaking it down into carbon dioxide (CO2) and water vapor (H20) gases in an accelerated version of naturally occurring decomposition. This new coating must be properly fabricated and installed to achieve the desired results. Figure 3 . But the nature of the vacuum deposition process appears to give these coatings only hydrophilic (or hydrophobic) properties. can also create dirt resistant coatings. present in all outdoor daylight. It should be noted that indirect UV reflected from clouds and buildings is sufficient to activate the'coating. The direct or indirect. in a simplified description. performs the charging action on the electrons. amongst other design characteristics. heat loss and heat gain. by first having its electrons raised to a higher level where they react with water vapor molecules to create OH radicals. The higher temperature of the glass causes a reaction to occur forming a TiO2 coating on the glass. The created OH radicals react with organic dirt on the glass. . without any effective photocataiytic activity. Manufacturing by Vacuum Deposition Process Sputter coating various materials on glass. Fully activated coatings have been seen on north elevations.CVD Pyrolytic Process A reactive gas mixture is presented to the freshly formed hot glass ribbon while it is still in the float bath.

but it also offers reduced accident potential in an area where OSHA has reported many accidents and fatalities from falls. it has been seen that most silicone sealants can release silicone oils which are not organic and do not take part in the photocatalytic self-cleaning reaction. These oils can cover the sell-cleaning coating for a band a few centimeters wide all around the window and permanently inhibit photocatalytic reaction. but there would also be no rinsing action available from rain to remove inorganic dirt. but the organic breakdown of the complete deposit by the coating alone will typically take an unreasonably long time. admit sufficient daylight. The Impact of Self-Cleaning Glass The new self-cleaning coating offers yet one more option to the window designer. have a color and appearance selected from a wide range of tints and reflectivities. provide acoustic insulation. Fabrication and Glazing Tools It is obviously vitally important that the coating is correctly installed in a window. it can reduce winter nighttime heat loss. such as bird droppings. By intelligent selection of products. from silicone sealants. The coatings work with daylight and rain. In dry areas a light hosing will be needed to remove dust. These work by emitting and detecting the reflection of UV light from the TiO2 coating. a window can now be designed and manufactured that is strong enough to resist hurricane winds and the associated debris. . Where large deposits of organic dirt.24 THE USE OF GLASS IN BUILDINGS Glazing Details Self-cleaning glass must be protected from contaminants which could smother or otherwise inhibit the catalytic and hydrophilic actions. Handheld tools are also available to detect low-emissivity coatings by measuring their electrical conductivity but unfortunately they do not detect the self-cleaning coatings. Handheld portable detectors are available to correctly identify the photocatalytic coating. are involved the coating is overwhelmed by the amount of material to be broken down. and finally have an exterior self-cleaning surface. Finally. Such deposits are quickly loosened by the coating and can normally be removed by hosing. it can keep out unwanted solar heat gain. Such contaminants include metal oxide run down from copper or lead roofing and silicone oils which can be leached out. Self-cleaning glass significantly reduces the labor needed to maintain the exterior surface of windows and represents a considerable cost saving to a building owner over its useful life. Placing the coating incorrectly on the room side surface would not only deny it most of the UV radiation needed to activate the TiO2. by water. Such light would not be reflected as strongly from clear glass. The self-cleaning coatings available today do not mean that windows will never have to be cleaned again.

References [1] "Worker Deaths by Falls. the others are all wrong." Department of Health and Human Services. its occurrence has already been seen in the field. there are up to 8 or 12 different assembly combinations for positioning the coatings: only one is correct. While this problem may appear to be trivial. The benefits of such coatings are now universally recognized and they can be found on commercial and residential windows in the cold north and the warm south where they give improved comfort and energy savings by day and by night. September 2000. When the typical double-glazed window includes a clear low-emissivity and a clear self-cleaning coating. When the performance and benefits of the clear and colorless self-cleaning coatings are recognized. There is now a requirement upon the window fabrication and glazing industries to recognize the availability of these significant glass improvements and to promote.BARRY AND O'DAY ON SELF-CLEANING GLASS 25 The window fabricators and installers now have an added complication to their work. in both summer and winter. it is suggested by the authors that in less than 20 years these coatings too will become similarly ubiquitous and will be used in windows as readily as the low-emissivity coatings. Conclusion It is less than 20 years since clear and essentially colorless low-emissivity glass coatings have been readily available. National Institute for Occupational Safety and Health. . fabricate and install them correctly.

Keywords: Stress. has the desired strength and that test documentation supporting the test data is available. depending upon aspect ratio and other factors such as the actual residual surface compression present [1]." for measuring surface and edge compression stresses. modified to permit the very high-speed data acquisition needed for measuring stress in a moving item. Redner l PC-Based Stress-Measuring System for On-Line Quality Control of Tempered and Heat-Strengthened Glass Reference: Redner. the testing should help the manufacturer maintain uniformity and production economy. ASTM International. quality control Introduction Heat-strengthened and fully tempered glass can be as much as 3 and 7 times as strong as annealed glass. In addition. PA 2002. "PC-Based Stress-Measuring System for On-Line Quality Control of Tempered and Heat-Strengthened Glass.astm. V. Heat-Strengthened. impact and service loads. Because the increased strength is the product's main characteristic. Quality Control procedures are designed to assure the producer and the user that the product meets material specifications. "Test Method for Non-Destructive Photoelastic Measurement of Edge and Surface Stresses in Annealed. ASTMSTP 1434... The additional strength is due to the presence of residual compressive stress in the surface and edge layers. An example of this influence can be seen in the specification C1048-97b where the residual stress range allowed in Heat-strengthened glass was narrowed as a result of the availability of more accurate test methods described in C1279-94. there is an obvious need to implement systematic Quality Control stress testing programs to assure conformance to US and foreign specifications [2]." Use of Glass in Buildings. S. West Conshohocken. In addition. tempered glass.Alex S. "Specification for Heat Treated Flat Glass . ASTM C1048-97b. within a region where the stress gradient is high. offsetting tension due to handling. the C1048-9To specification includes a reference to ASTM C 1279-94. Abstract: A new PC-based stress measuring system was developed for measuring edge-stress in tempered glass. . A. Ed. Kind FT Coated and Uncoated Glass. The system is based on the Spectral-Contents Analysis method.Kind HS. and Fully Tempered Flat Glass. the specification changes have been closely related to development of new stress measuring instruments." clearly defines the surface and edge pre-stress levels that must be met to satisfy "Fully Tempered" (FT) or "Heat-Strengthened" (HS) glass specifications. Block. 26 Copyright9 by ASTM International www.

the edge of tempered glass is. automotive and "IV glass industries [4]. and prEN 12120. however. Testing Procedures The fragmentation test is suitable only for "safety" glass. part 1 and part 2. While the GASP is extremely valuable for Quality Control. that the surface compression increases the strength due to service loads. the European standards are nearly identical to ASTM C1048-97b. just another free surface (E). conformance verification requires an accurate and economical stress measuring test procedure. In both architectural and automotive applications. Figure 1 . The new method described below permits "On-Line" measurements. [2] where only the strength is specified. Assuming. part 1 and part 2. assuring 100% quality control. In automotive glass. This is a destructive test seldom practiced in strict accordance with the test method. Edge Stresses in Heat-Strengthened and Tempered Glass As shown in Figure 1. in reality. edge stress testing has higher emphasis. Measuring surface compression using surface polarimetry [3] [4] and edge stress using transmitted light are described in the ASTM C1279 test method.REDNER ON HEAT-STRENGTHENED GLASS 27 These surface and edge stress levels are not explicitly stated in the European Norms prEN1863. The glass edge is exposed to air quenching in the tempering process [6] and develops edge-surface stress related to the temperature gradient developed during the quenching process in the layers adjacent to the edge-surface (E). The Grazing Angle Surface Polarimeter (GASP*) is a non-destructive test instrument extensively used in architectural. The European standard [2] also includes the required frequency of fragmentation testing and documentation. Product specifications published by automotive industries define an additional test parameter: the maximum average tensile stress occurring near the edges. Measurement of edge stress offers a viable solution for this purpose. justified by the frequency of installation and service failures. the method requires glass contact and is not adaptable to an "On-Line" process control.Development of Residual Stresses on Edge-Surfaces . [5]. and correlation between the measured surface compression and the strength is the user's burden.

A minimum of 10 points are needed to fit the polynomial with a suitable confidence level within the critical region 2 to 5 mm from the edge. The tensile stresses balancing the surface-compression add to the edge-balancing tension. The edge stress also reveals the strength and service performance of the product. the edge surface compression (-) is balanced by tensile (+) stress in the region adjacent to the edge. The average stress. creating a region where the average or integrated stress is tensile (+).00 .00 9 l mm Figure 2 . decreases rapidly as the distance from the edge increases. The Figure 2 shows experimental results acquired using several samples [6]. linear extrapolation technique using results of measurements at 2 points. a typical parabolic distribution develops. These calculations make the procedure cumbersome. Combining the linear speed and the length of the measured re- .< I I M.n 2.00 ' 2.00 I 1. (but not necessarily most accurate). 3. In addition. A simple. As a result. Edge stresses are routinely measured in transmitted light.28 THE USE OF GLASS IN BUILDINGS Surface compression (-) is balanced by tensile (+) stresses in the mid-layers. In regions distant from the edges. The speed of data acquisition of the SCA method makes it possible to implement On-Line edge stress measurements. glass exiting the tempering furnace moves with a linear velocity ranging between 100 and 500 ram/see. xl and x2 (Figure 2) is included in the ASTM C1279-94 test method.00 Compression MEASURED RETARDATION VI DISTANCE FROM THE EDGE t. The PC-based SCA method [8] automates the test and eliminates the difficulties of manual extrapolation.Average Stress Measured in Transmitted Light Similarly. tempered and heat-strengthened glass is substantially weakened near the edge [7].00 Tension 4. This weakened region is of concern to the automotive industry and maximum tensile average stress is stated in automotive glass specifications. measured in transmission.~IMUM AVERAGE TENSILE STRESS 0. As result of edge finish geometry. Automated On-Line Production Control of Prestress The edge stress gradient is very high in a very narrow region.00 6. edges are not transparent and an extrapolation method is needed to obtain the "real edge" stress. It should be noted that a 4 th degree polynomial provides an excellent fit to experimentally acquired data points.

For each wavelength 2.. installed at the exit of a tempering furnace. The system yields an over-abundant number of data points to accurately establish the leading and trailing edge-stress. Using a spectrophotometer. The transmitted light acquires a retardation 8. A light-weight sensor is used to scan the edge of automotive glass. The system schematic (Figure 3) illustrates the components used to implement this method On-Line. as well as the maximum average tensile stress in the near-edge region. This SCA method has been used extensively during the last 10 years for On-Line stress measurement in float glass. An SCA based edge-stress scanning system used to measure edge stress in automotive glass is shown in Figure 4. The basic principles and applications of this method are extensively documented in several publications.1 mm. the transmitted light intensity is modified. 8=txCxS where: fi t C S Retardation related to the Stress Thickness of material Material Constant Stress (1) The method is using white light containing a broad spectrum of wavelength g. measures stress at a speed in excess of 2000 points/see. yielding 1000 data points/sec and a spacial resolution of 0. related to the stress S. the light intensity Ii is measured at several wavelengths providing a set of simultaneous equations (2) sufficient to retrieve the retardation 8 and compute stress. SCA Measuring Method [8] When polarized light crosses a sheet of glass exiting a tempering furnace. one finds that data acquisition near an edge must be completed in less than 0. The SCA sensor described below.REDNER ON HEAT-STRENGTHENEDGLASS 29 gion. The light source (Figure 3) projects polarized light of intensity Io on the glass. demonstrating the method capability.. requiring a measuring speed of 1000 data points/second. The selection the spectral range determines the speed of .01 second. the transmitted light intensity I becomes: Ii =Iosin2~rfl/~i where: (2) Ii /o 2i Intensity of light measured at ~ wavelengths Source intensity Wavelength The light transmittance Ii/Io becomes a function of stress and of the wavelength L.

30 THE USE OF GLASS IN BUILDINGS S P A G Io Io sin2.Portable Stress Scanner using PC-Based SCA Method .~ PC L P S A G Light Source Iii ~ l J Polarizer Analyzer i MeasuredSample ] SpectralAnalyzer I Io sm2~rS 9 Figure 3 .Schematic of the SCA Measuring System Figure 4 .

On the other hand. resolution and maximum retardation measuring range. the average stress measured in transmission decreases to zero. . The measuring sensitivity of the SCA sensor shown in Figure 4 was 1 nm (0. The system is PC based allowing 100% inspection and documentation. At a small distance from the edge. the retardation to be measured is less than 200 nm. eurve fitting.Stress Scan Graph Conclusions A new stress measuring system was developed. The speed and spatial resolution of the method permits On-Line monitoring of tempered and heat-strengthened glass.02 MPa in 2. The SCA systern designed for measuring edge stress was evaluated up to 8.5 mm thick glass). verification of"zero" and performs a scan based on selected scan length and scan speed.000 nm. A typical edge scan result is shown in Figure 5. The software permits automated calibration. Figure 5 . in a 2 mm thick heat-strengthened glass (stress is 30-40 MPa).REDNER ON HEAT-STRENGTHENEDGLASS 31 data-acquisition. not possible to obtain using present Quality Control methods. requiring a resolution of I nm. capable of measuring 150 MPa stresses in glass thickness up to 20 mm thick.

S. and Voloshin. "Delayed Cracking in Automotive Windshields. Redner. Redner. "Stress Measurement in TV Production.S. A. pp. A. ffune 1985. B-1050 Brussels. 1990.32 THE USE OF GLASS IN BUILDINGS References [1] [2] [3] [4] [5] [6] [7] [8] Minor. and Acloque. 1962. A." GLASS. 415-424. CEN. 74 (6).. June 1997. Guillemet. 218-219." Material Science Forum 210-213. and Bhat. part 2) and prEN12120 (part 1... 8-90." Experimental Mechanics 25(2) pp. "Basic Glass Strength Factors. . "New Optical Method for Determination of Stresses Near the Surface.. C. "Surface and Edge Stress in Tempered Glass. 157-163. part 2) draft." Proceedings. pp. 1996. 2000. G.T. 169-171. rue de Stassart 36.E. P.S. 9th International Conference on Experimental Mechanics.K. Paris." 2nd GAMAC Conference. 148-153. 9-2000. pp." Glass Digest. pp. European Standards prEN1863 (part 1. "Precision of Surface Stress Measurement Test Methods and Their Correlation to Properties.S.. Gulati.S. "Photoelastic Measurements by Means of Computer-Assisted Spectral Contents Analysis. 52-57. GPD. A. J.. et al. Redner. pp.. June 1999. Copenhagen. S. Redner." Proceedings. A.


2395 Speakman Drive. W. 260 Jackson Street West. Ontario. 3 Senior Technologist. "In-Situ Dew-point Measurement to Assess Life Span of Insulating Glass Units. R.. LgP IM5. longevity.. The life span of insulating glass units in service is not well known.. ASTM STP 1434. ASTM International. This technique is examined in this paper and modifications are suggested to improve accuracy. ToroL ~Wemer Lichtenberger. 35 Copyright9 by ASTMInternational www.. could be used to estimate time to failure oflG units in service. Keywords: Insulating glass units. Replacement costs can be estimated fairly accurately. However. Block.. Spetz proposed that one component of the laboratory test method. PA. dew-point measurement of cavity gas fill. service life. were intended to assess the likelihood of successful performance through the IG unit manufacturer's warranty period.2 and Allan Majorz In-Situ Dew-point Measurement to Assess Life Span of Insulating Glass Units Reference: Torok. Ontario. West Conshohocken. Mississauga. In the 1980's.George P.. Gerald g. Ontario. not to determine service life . Insulating glass units have been made in North America since the late 1950s.Tmseal Technologies Ltd. timing is less certain.astm. Lichtenberger.. time estimates for replacement are o/~en based on poor understanding of the causes of IG unit "failure" (water vapour condensation on glass surfaces facing the IG unit cavity) and previous negative experience. dew-point i Project Manager. and thus are reactive rather than predictive. G. In the author's experience. Insulating Glass Laboratory. V. Canada. Canada. 2002. Hamilton. Laboratory test methods developed in Canada in the late 1950s and early 1960s. L3Y 3Z6. Canada. 2 Special Projects Manager. Bodycote Materials Testing Canada Inc. Ed. L5K 1B3. 27 Main Street North. A. Newmarket. subsequently used as the basis of most IG unit test methods worldwide.. Abstract: Replacement of insulating glass (IG) units in buildings is expensive.." The Use of Glass in Buildings. G-engeBuilding Consultants Inc. based on in-situ testing for the "Field Correlation Study" by the Sealed Insulating Glass Manufacturers Association (SIGMA) in the USA. and Major.

or "fogging. and units with a dew-point warmer than 30~ were considered to have "failed" (water vapour had condensed on surfaces bounding the air space. although ASTM E773 / E774 does include UV exposure in the lab during weather cycling). In their 1962 paper. on average. 2. to provide some correlation to "real" life including exposure to UV radiation (currently. drying. in a double-lite IG unit. on surface 2 or 3). The number of cycles of heating. The DBR test protocol consisted of the following components: 9 An initial seal test to test the integrity of the hermetic seal." to be objectionable [2]. the dew-point temperature of the IG unit cavity gas fill was measured. the Sealed Insulating Glass Unit Manufacturers Association (SIGMA) in the USA embarked on a "Field Correlation Study" to confirm the apparent correlation between the lab test protocol and field service life. it was . neither the Canadian CGSB-12.8-97 Insulating Glass Units standard nor the American ASTM Test Methods for Seal Durability of Sealed Insulating Glass Units (E 773) and ASTM Specification for Sealed Insulating Glass Units (E774) laboratory test standards include outdoor exposure or UV testing. At the time. The DBR test protocol was developed at the request of what is now the Canada Mortgage and Housing Corporation (CMHC) as a tool to evaluate the suitability oflG units promoted by manufacturers for installation in new housing funded under the Canadian National Housing Act (NHA) and administered by CMHC. However. and indeed many of the IG unit laboratory test methods world wide. the maximum (warmes0 dew-point was set at -40~ (-40~ Correlation of the DBR test protocol and the subsequent CGSB and ASTM protocols with service life is limited. and 9 Outdoor exposure to "natural" weather cycling. This temperature was arbitrarily selected by the DBR researchers as a likely temperature at which building occupants might consider such condensation. 9 Repeated cycles of exposure to high humidity with pressure change to test the water vapour resistance of the seal (other standard test methods do not include cycling of pressure and humidity). This was also (and still is) the length of the industry standard IG unit warranty against seal failure. Wilson and Nowak at the Division of Building Research (DBR).36 THE USE OF GLASS IN BUILDINGS North American Laboratory Test Methods to Assess IG Unit Performance Existing North American IG unit laboratory test methods.8-97 Insulating Glass Units standard. Thus was established the supposed correlation of laboratory testing to 5 years of"real" service life and the industry standard 5 year warranty period. water spray. water spray. given that winter temperatures in much of Canada often fall below 30~ the value was not considered sufficiently severe and in the CGSB-12. In the late 1970s. Throughout testing. are based on research by Solvason. 9 Repeated cycles of heating. 3]. and cooling primarily to test mechanical strength of the perimeter seal. Wilson and Solvason noted a "rough correspondence" between failures of units subject to varying lengths of laboratory testing (fi'om 0 to 880 cycles) and of units of the same type subject to one year of outdoor exposure testing [2]. drainage and cooling were eventually standardized to 320. across Canada in a 5 year period. National Research Council Canada in the late 1950s and early 1960s [1. about the same number of thermal cycles experienced.

and the units studied were made with available sealant products. and on learned dissertations in technical papers. was about 2. installed so that the perimeter seals were not subject to prolonged wetting. it was possible to estimate desiccant moisture content. after about 25 years many of the original IG units are still performing without even transient evidence of failure (fogging on cold nights). The authors have encountered many IG units still in service. Thus the results are. By relating dew-point measurements to desiccant manufacturer's isostere charts (plots of desiccant saturation as a function of desiccant temperature and dew-point of the air exposed to the desiccant). to estimate remaining life span [8. to units of that vintage. This result was anticipated by Wilson and Solvason.8-97. fog free. For example. and installed in accordance with practices of the day. This was based on analysis of dew-point measurements made during the first 10 years of the study.400 IG units e r a population of 40. etc. unique to the USA. to units with those orientations [4. We are therefore left with estimates based on personal experience. under limited circumstances. Dew-point Testing to Assess IG Unit Performance Existing Technique An outcome of the SIGMA "Field Correlation Study" was a proposal by Spetz (the auditor of the units) that in-situ dew-point measurement could be used to assess performance and. 7]. most of which faced south or southwest. Field studies began in 1980 and terminated 15 years later. 9]. after 46 years of service. In their 1962 paper to the Canadian Ceramic Society. Within these limitations. renovation or subsequent denial of access. test protocol and thus somewhat different than other laboratory test standards. But what is the anticipated service life of a unit? We cannot rely on the results of the SIGMA study because it was terminated after 15 years.TOROK ET AL. a modified version of the DBR. some of the units were "lost" during the test period because of demolition. .9% after 15 years [5]. ON INSULATING GLASS UNITS 37 generally understood that IG unit constructions tested successfully in the laboratory were capable o f much longer service lives than 5 years. up to about 26 years in age in service. Others have made similar estimates [I. perhaps. The study has its limitations: it is a comparison of field exposure to the American ASTM laboratory test protocol ASTM E773 and accompanying specification E774. Francis [6] reported that IG units manufactured in accordance with current industry standards and properly glazed should achieve life spans of 20 years or more. 2. desiccants.. the SIGMA "Field Correlation Study" revealed that failure oflG units made to the highest performance level of the ASTM E774 specification ("CBA"). such as CGSB-12. In the building to be discussed later. on reports in the glazing media. they remarked that diffusion of water into an IG unit cavity and saturation of the desiccant was unlikely to lead to failure within the industry standard 5-year warranty period but it might within the anticipated service life of a unit [2].000 in 40 buildings in 14 cities in the continental USA were studied. when most of the units had not yet failed. and in one exceptional case (given the technology available at the time of manufacture). Spetz found that units with desiccant close to saturation were likely to fail within a short time.5].

Critique of the Existing Technique Moisture adsorption capacity of desiccants varies with temperature of the desiccant.38 THE USE OF GLASS IN BUILDINGS Spetz proposed the following evaluation scheme for IG units in service [8]: 9 Dew-point < -80~ there is almost no moisture in the IG unit cavity. The method used by Spetz to measure in-situ dew-point measurement was formalized as ASTM E 576. The next lowest dewpoint temperature range given. 9 Sealant Moisture Vapour Transmission Rate (MVTR. 9 Dew-point between -80~ and 0~ there is some moisture in the cavity. 9 Workmanship. 9 Construction techniques. 15. 9 Water vapour pressure differential across the perimeter seal. and different desiccants adsorb different quantities of moisture at a given temperature [12. Apart from the two articles by Spetz in the 1980s previously referenced. because the suggested time period to permanent fogging. and 9 Service environment . formal development of the technique of in-situ dew-point measurement to evaluate remaining life span of IG units. or water vapour permeance). 14. 9 Dew-point > 32~ permanent fogging of glass surfaces bounding the IG unit cavity can be expected to develop within two years. including [7. 2 years. 17]. 16]: 9 Water vapour adsorption capacity of the desiccant. there does not appear to have been any further. 0~ (-18~ may be sufficient. a dew-point temperature of32~ (0~ is probably too warm to be useful as an indicator of an impending failure problem. thus the IG units will have a relatively short future life. then as for the original DBR test protocol. Test Method for Frost Point of Sealed Insulating Glass Units in the Vertical Position. is too short for building owners to begin a reasonable savings plan to fund replacement. If we assume that the scale given by Spetz is based on a desiccant temperature of 72~ (22~ the standard temperature for IG unit laboratory testing). 15. 14. 16. Estimation of remaining life span requires knowledge of the construction of the units. but that depends on the rate of water vapour gain of the IG unit cavity gas fill. I0. 9 Sealant Moisture Vapour Transmission Path (MVTP) length (from outside the I(3 unit to the cavity) and area (perimeter length x width from spacer bar shoulder to adjacent glass face). I1. 13. This standard addresses only the method of measurement of dew-point temperatures for IG units. thus the IG unit can be expected to have a future clear life less than units with a dew-point < -80~ 9 Dew-point between 0~ and +32~ there is "considerable" moisture in the air space. The rate of increase is a function of several factors. thus the IG units can be expected to have a "very long expected future clear life". It does not include Spetz's proposed assessment scale or other methodology for evaluating the performance of an IG unit and its remaining service life.

For example. and general atmosphere pressure changes). are colder than in Toronto. wintertime temperatures in Edmonton. the average of mean daily minimum temperatures for the winter months. However. 9 Water vapour permeance (MVTR) varies with changes as the water vapour pressure gradient differential across the sealant changes (increasing as relative humidity of the IG unit cavity gas fill increases). a function of several factors. ON INSULATING GLASS UNITS 39 The water vapour adsorption capacity of a desiccant is. in turn. solar gain. temperature change. we propose that it is then possible to make such an assessment. or perhaps for a sensitive client. Therefore. the mean daily minimum temperature for the coldest winter month. for example. 9 MVTP length and area changes as the cavity volume increases and decreases (in response to ambient air. and if care is taken to reduce the effects of the factors listed. Proposed Technique When "failure" of an IG unit due to fogging is considered to occur is relative. However. from a single measurement of dew-point temperature it is not likely that an accurate assessment of remaining fog-free life span can be made. 9 Size of openings of water molecules versus desiccant crystal pore size. 9 Strength of the attractive force between water molecules and the surface of the desiccant. most occupants would likely complain. if measurements are repeated over time. Alberta. Thus the rate of increase in the water vapour content of the IG unit cavity gas fill may not be constant over time. including [12. distinctly more so for silica gels than for molecular sieves). fogging is likely to occur first in Edmonton.TOROK ET AL. One could choose. so for the same IG unit construction. depending on location and sensitivity of the building occupants. or from one occupancy type to another). Thus the first step in assessment of remaining service life of an IG unit is to determine the outdoor ambient air temperature range in which fogging is likely to be considered objectionable. These temperatures could be selected from historical climate records available from government meteorological agencies. from occupant to occupant. Ontario. The effect of many of these factors is dynamic. Many building occupants probably wouldn't be too concerned if fog formed in an IG unit during a very cold night then evaporated shortly after the sun went up. and 9 Temperature of the desiccant. causing the perimeter sealants to extend or compress. although it would raise some questions. 13]: 9 Rate of diffusion through the binder (for bead forms) or matrix (for extruded forms) that holds the desiccant crystals. and 9 The service environment may change (from summer to winter. if fog lingers well into daylight hours when unobstructed vision is generally desired. capacity decreases at low and high temperatures. not static. For example: 9 The water vapour adsorption capacity of both silica gel and molecular sieve desiccants varies with temperature (generally. .

cavitygas fill f / mo~. and Dew-point Measurement . The two could be the same. and thus the dew-point temperature of the gas fill.-%4 Dewpoint measurements made from timetotime Time. and from that an estimate of time to failure is to be made. the moisture content of the IG unit cavity gas fill to which the desiccant is exposed. Thus if dew-point temperatures are to be measured over time and compared to determine a rate of increase.Relationship Over Time Between Cavity Moisture Content. 1.40 THE USE OF GLASS IN BUILDINGS As noted.. is dependent upon the desiccant temperature. When plotted together on a psychrometric chart. years ~/~/ % ~ / Maximum IG y unit. perhaps. Although some error would be introduced by measuring dew-point over a range of temperatures. fogging is unacceptable. the outdoor ambient air temperature. Therefore. as shown in Fig.--. not as clear because the time scale is distorted to fit the temperature scale. i / l . Timeperiods between dew point measurements are of equal length Timeperiodfrom last measurement of dewpoint temperature to cold weather temperature is / estimated. 1. The relationship of time to dew-point temperature rise is. it would be practical for field measurement. Outdoor Ambient Air Temperature. and the measurement of dew-point temperature over time is shown in Fig.tu~ content. the dew-point measurements from year to year should be made when outdoor ambient air temperatures arc about the same./ Outdoorcold weather temperaturerange in whichfogging is unacceptable Temperature. location. / . The relationship between moisture content of the IG unit cavity gas fill. The outdoor ambient air temperature at which we measure dew-point temperature for time to time comparison should relate to the range of outdoor ambient air temperatures in which fogging is considered unacceptable.previous time period ] \ lengths and dewpoint / \ temperatureris. becomes clear. ~ Figure 1 . The desiccant temperature of an IG unit installed in a window in a building will lie somewhere between the outdoor ambient air temperature and the indoor ambient air temperature. basedon [ r. the rate of dew-point temperature rise is decreasing with time. the relationship between cavity gas fill dew-point temperature and the range of outdoor temperatures in which. For example. the water vapour adsorption capacity of a desiccant in an IG unit is dependent upon its temperature. for a particular building. and client.'! -I "~ um 1(7 // / t. ~ / . ON INSULATING GLASS UNITS 41 From Fig. type. l l .~tu.1 t l '. the length of time that may be required for the dew-point temperature to rise from the last recorded measurement into the outdoor cold weather temperature range in which fogging is unacceptable. and at the temperatures at which fogging is not acceptable. and dew-point measurement frequency distributions.. based on previous time period lengths and dewpoint temperature rise.. Since measurements are made at similar temperatures (the smaller the range. .. time to time ~ / ~ 1 ~' I .o. The intent is to make several measurements and yet leave sufficient time from the last measurement to the lower end of the cold weather temperature range for the owner to make reasonable financial arrangements to fund IG unit replacement. 1. outdoor ambient air temperature. and manufacturer of the desiccant and to obtain the manufacturer's isostere charts for estimating moisture content.~ Figure2 . We propose that a measurement program begin as soon as possible after occupancy. With respect to those factors.4.... Outdoor cold weather temperature range in whichfogging is unacceptable ~ Temperature. it is not necessary to determine the quantity. moisture content of the desiccant becomes irrelevant. the symptom of increasing moisture Timeperiods between dew point measurements are of equal length ~...Relationship over time between cavity moisture content.... ~'- Timeperiodfrom last measurement of dewpoint temperature to cold weather temperature is estimated. the better).TOROK ET AL. In essence.o. one of the advantages of measuring dew-point temperatures consistently at the same outdoor ambient air temperature range is the practical elimination of desiccant temperature variation as a potential source of error in estimating time to failure...osl 'cu-. that is. years Frequency distributions I I ] [ ofdewpoint mue er n ces ~_ m t made ] A \[ \ / l d-r so. it can be seen that in order to estimate time to failure. A .-.~ Time. dew-point measurements must be made when those temperatures are likely to be well colder than the outdoor temperature range. .

Even with the practical elimination of dew-point temperature as a factor in dewpoint temperaatre rise. the cost of testing. it can be expected that measurements of a random sample of IG units would yield a distribution of dew-point temperatures. . Over time. Therefore. it is likely that there will be some variation in measured dew-point temperature from one unit to the next. tracked by repeated measurements. Figure 2 is a blend of frequency histograms of temperatures measured at several different times. is of greater practical importance. with the psychrometric chart from Fig. as the IG units age. These concepts will be demonstrated in the following case study. This greatly reduces the complexity and therefore.Municipal Building. This is illustrated in the Figure 2.42 THE USE OF GLASS IN BUILDINGS content. Summer 2002. 1. over time. at the same desiccant temperature. it can be expected that all of the units will not fail at once. the temperatures of this distribution would increase. Case Study: Municipal Building Figure 3 . South Elevation (North Elevation Similar) Curtain Wall Repairs and IG Unit Replacement in Progress. eventually into the outdoor temperature range in which fogging is unacceptable. From this diagram is can be seen that. the rise in dew-point temperature of the cavity gas fill. there remain sufficient other factors such that within a given population of IG units. thus funding for replacement can be arranged to be available over some time period.

If testing revealed that failure was likely to occur within three years then the repairs could take place immediately. presumably. . the future date would be established by repeated dew-pint measurements). Ontario. Figure 4 -Dew-Point Testing with ASTM E547 Test Unit. correctly. Because the majority of the IG units were original (fabricated in 1977) and now aged (23 years at the time of the preliminary investigation). then the repairs could be delayed until some future date (although not stated in the consultant's report. Dew-point testing was recommended to assess the condition of units and determine timing for repair. possibly hastening failure (fogging).TOROK ET AL. the consultant suggested that removal and reinstallation could result in stress to the IG unit perimeter sealants. This was a feasible approach because the water leakage being experienced had virtually no impact on day-today building operations. ON INSULATING GLASS UNITS 43 Dew-point testing to estimate remaining life span of insulating glass units was required for a municipal building in Toronto. A previous consultant had made a preliminary investigation of water leakage problems and had determined. that removal of IG units in curtain wall cladding in the north and south elevations of the building was necessary to correct defects in the original installation (missing framing joint seals) that were responsible for chronic water leakage. if testing revealed that failure was likely to occur beyond three years. Canada (Figure 3).

5~ to -11. for a total of 36. . from -2. a good sample size.Dew-Point Test Results. and the electrically powered "Cryocool" unit from the insulating glass laboratory of Bodycote Materials Testing Canada Inc. but this could not be arranged for a variety of reasons. 4). Dew-point measurements were made in 10~ brackets below 0~ Temperature brackets were used rather than precise temperatures to speed measurements. Ideally. temperature measurements should have been made at the low end of the range. ideally. dew-point temperature measurements should have been made in January. North Elevation. The bracket boundaries were based on the range of mean daily temperatures for the month of January for Toronto. The indoor temperature was a constant 21~ Dew-point measurements were made with ice cubes. This represented 15% of the total of 240 IG units in the two curtain walls. at -11. which proved to be more durable. Field (ln-Situ) Measurements Dew-PoimTemperature <0oC 0oC -10oC -20oC -30oC to to to to -10*C -20~ -300C -400C Fog Fo8 Fo8 Fos Fog Fog Fo8 Fo8 Fo~ Fo8 Fo[ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Fo[ Fo8 No Foa Detected Fog Fog Fog Fog No Fo~ Detected I IF~ I I Figure 5 . -10~ It is therefore expected that dew-point measurements made would be somewhat high and a resultant estimate of remaining time to fog at -I~ would be less. Unit No. Most measurements were made with the E 547 equipment. the test apparatus as described in ASTM E 547 (Fig.44 THE USE OF GLASS IN BUILDINGS Twenty (20) IG units in the north elevation curtain wall and 16 IG units in the south elevation curtain wall. The ice cube was used to check for dew-point temperature above 0~ and the E 547 and the "Cryocool" equipment were used for testing at temperatures of-10~ and below. Outdoor temperatures at the time were about -3~ at the high end of the range of January mean daily temperatures.1~ This range was obtained from the "Climate Normals" records available from Environment Canada for the 1937-1990 period. were tested in the late winter of 2001.1"(2 or the low end of the first bracket.

5 and 6. Clearly.5 era long sealant failure at setting block.. However. as expected (Figure 7). water vapour had entered the units and the desiccant was becoming saturated.. Blend of molecular sieve and silica gel desiccants. The difference in modal values suggests that IG units in the south elevation can be expected to fail before IG units in the north elevation. 11.30oc Point. Aider conditioning at standard temperature. more testing was required to confirm the trends that emerged from the dew-point measurements made.9 16 Fog Figure 6 . 1 2 3 4 5 6 7 8 9 10 11 12 Field {ln-Situ)Measurements DewDew-Point Temperature <0*C 0*C . Three units were successfully removed from the south elevation and delivered to the Insulating Glass Laboratory at Bodycote Materials Testing Canada Inc. Clearly.1 Local sealant failure at setting blocks.ments Moisture Comments Content.o 1 I -I~ 16. % by Weight 45 Unit No. As expected. two distinct bell-shaped curves emerge. all but one being in the south elevation. filled two long sides. Results for the 1977 units are tabulated in Figs.1 13 14 15 Fog Fo8 Fog 0oC 16. ON INSULATING GLASS UNITS Laborator~Measuro. the client required immediate direction so it was decided to remove several units of the original 1977 units for further examination and testing in the laboratory. lab dew-point measurements were warmer than those . This corresponds well with observed IG unit failures. Blend of molecular sieve and silica gel desiccants. The modal value for the north elevation IG units was in the -10~ to -20~ bracket and the modal value for the south elevation was in the 0~ to -10~ bracket. Blend of molecular sieve and silica gel desiccants. separately for the north and south elevations. dew-point measurements were again made. When graphed in the form shown in Figure 2 (frequency histogram combined with a psychrometric chart). two long sides.1 0 * C -200C . This should not be surprising for 24 year-old I(3 units. filled two lons sides.TOROK ET AL. +6~ 17. South Elevation.Dew-Point Test Results. tO ~ to to to -10*C -200C -300C -400C No Foil Detected ] F~ Visible Condensation {Failed Unit) No Fog Detected Fog Fos Foil Visible Condensation (Failed Unit) Fog Fo$ Fog .

. Desiccant was a blend of silica gel and molecular sieve. .. . the difference between field dew-point and lab dew-point measurements was not consistent for the 3 units tested. and thus raising the dew-point temperature. Based on available desiccant manufacturer's data sheets . . . The latter was felt most likely. .' ] I'. because discolouration and partial loss of adhesion of the perimeter sealant was noted at setting block locations.76 m (2 1/2 ft). .' . ~ - I I~ r 0 Outdoor cold weather temperature ranEe Temperature. 13 mm (12 inch) wide cavity). . 0 conaensat. . and filled two long sides (unit dimensions were approximately 1." . Desiccant was removed from each unit.9% by weight (Fig. Desiccant saturation was determined to vary from 16. 'I cavity gas fill moisture .1-17. . increasing the moisture content of the cavity. _ temperature Jrequency /~ / : t x=? Soath Elevation I measured dew point 4t I ] . .46 THE USE OF GLASS IN BUILDINGS measured in the field (Fig.'"[". ~ ." ~ r -10 I. causing additional moisture gain. 0.83 m (6 fl) x 0. 6 mm (1/4 inch) thick glass lites.. . prEN-1279-2. 6). . This could have been caused by error in field measurement of dew-point temperature or rupture of the perimeter seal during removal and transportation. . Time for peak o f south elevation measured dew point frequency distribution curve to move into outdoor cold weather temperature range is unknown. The technique followed was similar to that described in the proposed European insulating glass unit test standard. . Glass in Buildings.". weighed. However. .temperatureJrequency[ I/ distribution / 'I I t ~ s . . ~ ' I .'" / . then dried and weighed again.0016 -40 -20 Figure 7 - Results o f ln-Situ Dew-point Testing at Municipal Building The perimeter sealant for the original 1977 units was polysulfide (single seal).X I content without . Part 2: Long Term Test Method and Requirements for Moisture Protection.I'. /~ .on I ~ -30 . . This is to be expected because of desiccant reflux. . . 10 North Elevation measured dewpoint . Insulating Glass Units. ./ I l. . .. 6).

More accurate estimation could be made if dew-point measurements were taken periodically. time could be calculated. we advised the owner to schedule repairs (and IG unit replacement) to that elevation as soon as funds permitted. performance oflG units could be . the rate of moisture gain is not likely to be constant such that estimation of remaining life span is not likely to be accurate based on a single measurement. repairs (and replacement of IG units) could be delayed. If repeated periodically. 14]. and suggests that the proposed test methodology is valid. another winter passed. that removal and reinstaUation could result in increased seal stress. This generally confirms our findings. The recently revised provincial Condominium Act requires boards of directors for condominium corporations to establish reserve fund plans to fund future capital repairs and replacements. The basic technique was outlined by Spetz in the mid 1980s and the method of field determination of IG unit cavity gas fill dew-point temperature was formalized in ASTM E 576. An opportunity exists with recent legislative changes in Ontario. Dew-point testing oflG units in buildings with large expanses of glazing could be done as part of on-site reviews for reserve fund planning. However. this meant that repeated testing to confirm the notion of advancement of the dew-point curve and estimate to failure could not. if it occurred. Because the south elevation IG unit in-situ dew-point curve modal value was in the ambient outdoor temperature range in which fogging. Further field testing is required to validate the technique. Given the warm dew-point temperatures measured in the field and in the lab. be confirmed. would be considered unacceptable. Fortunately. The range of mean daily temperatures for the coldest month of the year is proposed as a guideline for this temperature range. ON INSULATING GLASS UNITS 47 for the types of desiccant found (molecular sieve and silica gel). over many years. there were some administrative delays in the approval of the work so in the meantime. we advised the owner that if funds for replacement were limited. As it turned out. whereas there were no new failed IG units in the north elevation. Because of the many factors that contribute to water vapour ingress into an IG unit. the peak of the histogram curves vs. increasing the rate of moisture gain and reducing the time to failure. Timing could be estimated by further testing. might be considered objectionable. this indicated the desiccant was very close to saturation(between 18% and 20% by weight) [10. which must be periodically updated to maintain them current and accurate. sufficient funds were made available to make repairs to both the south and north curtain walls. we visited the building and observed that the number of IG unit failures (fogging) in the south elevation doubled. Conclusions In-situ dew-point testing has been proposed as a method to estimate performance after IG units have been in service for some time. and the rate of temperature increase of. for instance. because the north elevation IG unit insitu dew-point curve was colder. Canada. and discolouration and adhesion failure of perimeter sealants at setting blocks. .if it occurred. The results could be plotted as frequency histograms. therefore. 13.TOROK ET AL. This rate could then be used to estimate when the dew-point temperature of any point in the curve would coincide with the selected range outdoor ambient air temperatures in which fogging. Unfortunately. it was reasonable to concur with the consultant who made the preliminary review of the building. During that winter.

American Societyfor Testing and Materials (ASTM) STP 1168. Ontario. 1987. L. Scientific Basis and Application of International IGU Durability Tests. S. 1996. R.." Building and Environment. June 5th and 6 th. N. Ed. McGraw-Hill. November 15. Spetz. K. 1997. pp.. Kromminga.. NRCC 7042. J. 28 ff. pp 363-368. "The Role of Desiccants in Insulating Glass Units. August 15. G. "Performance of Sealed Double-Glazing Units. August 2000. Double-Glazed Window Units. L. J." Proceedings of Window Innovations '95." Handbook of Glass in Construction. October 1962.. A. 66--67. "The Role of Desiccants in IG Units. Insulating Glass Manufacturers Alliance (IGMA). C." Journal of the Canadian Ceramic Society." SIGMA-GRAM Technical Bulletin TB-2000520. E. L. A." Technical Bulletin TB-4000-01." Glass Magazine. PA. pp. 3-16.. June 1986. Sealed Insulating Glass Manufacturers Association (SIGMA).. Vol." 1GMAC Technical Seminar. 1995.." Symposium on Testing Window Assemblies. 1992. Nowak. K. A. "Desiccants. Amstock. V. "Evaluation of Factory-Sealed. "Design. and J. Francis. circa 1995. West Conshohocken." UOP Inc. 34.. 1996. 42. ASTMSTP 251. R. Gallion.. and Harris. November 1990. 297-321. J.. DBR-RP-168. 1959. pp 42-47. ASTM." Glass Digest. 1991. Glazing and Waterproofing. R. "Desiccant Works with Temperature to Prolong the Life of Insulating Glass Units. J. Ottawa. 38 ft. C.. "IGMA Unit Longevity Statement. "Frost Point Measurement: How a Frost Point Tester Can Be Used to Predict the Future Service Life of Insulating Glass Units in Buildings. J." Glass Digest. 43. August 1996. Spetz. 1990 "UOP Molsiv@Adsorbents. Solvason. "UOP Molecular Sieves. "Send it Back! Know How to Spot the Signs of a Badly Fabricated IG Unit. NRCC-5270. "The History. S. J.. pp 67-81." Science and Technology of Building Seals. Spetz. "Field Performance of Insulating Glass. Sealants. DBR-RP-85. pp 62-68. Lichtenberger. Toronto. August 2000. L. J. PA. G. pp. W. Spetz. 22-27.. Wilson. Solvason." UOP Inc. New York. pp.. References [ 1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [ 12] [13] [14] [15] [26] [17] Burgess. G." Insulating Glass Manufacturers Association of Canada (IGMAC) Technical Seminar. pp.." Glass Magazine. "Examining the Factors that Determine Insulating Glass Longevity. ASTM International.. Chicago. 31.. pp. 1999." Glass Magazine. "Zeroing in on Premature Failure of I(3 Units.48 THE USE OF GLASS IN BUILDINGS tracked and it may be possible to estimate and eventually confirm if the technique is accurate.. T. Dangieri.. Wilson. West Conshohocken. Parise. Giangoirdano. Fabrication and Performance Considerations for Insulating Glass Edge Seals. "SIGMA Field Correlation Study. W. g. T. . No.

For those products that cannot be simulated. and Shah. The simulated results were used as a benchmark for the analysis of the testing data. and the rating is a total product rating. Each round robin specimen was modeled per NFRC 100 (1997): Procedures for Determining Fenestration Product U-factors and NFRC Test Procedure for Measuring the Steady-State Thermal Transmittance of Fenestration Systems using NFRC-approved computer software tools WINDOW 4. B.astm. a Condensation Resistance (CR) Rating can be modeled and calculated simultaneously with U-factor. V. Shah Evaluation of the Condensation Resistance Rating as Determined Using the NFRC 500 Procedure Reference: Wise. sliding glass doors. . Fenestration products include residential. a test only option is also offered in the standard. This procedure emphasizes simulation calculations and allows the utilization of computer software in analyzing and calculating the Condensation Resistance Rating for comparative purlx)ses. skylights.1 and THERM 2. All of the NFRC round robin test specimens were tested at NFRC-accredited testing laboratories. D. door-lite. commercial and sitebuilt applications. the product is modeled and rated for a specific size at a net zero air leakage: meaning the product is sealed and air leakage effects on condensation index rating are not taken into consideration. Introduction NFRC 500 [1] provides a method of determining a Condensation Resistance rating for fenestration products. Block.Daniel J. This paper analyzes the results obtained through NFRC annuai round robin tests run from 1997 to 2001.1. "Evaluation of the Condensation Resistance Rating as Determined Using the NFRC 500 Procedure.. The total product is evaluated for condensation. Ed.. ASTMSTP 1434. For the purposes of rating. ASTM International. Abstract: This paper presents a comprehensive and detailed assessment of the National Fenestration Rating Council (NFRC) 500 Procedure. entrance doors." Use of Glass in Buildings. and curtain wail. J. including windows. 49 Copyright9 by ASTM International www. Solar Heat Gain Coefficient (SHGC) and Visible Transmittance (VT) ratings. PA. West Conshohocken. V. By using NFRC-approved software tools. Wise and Bipin V.

Calculation of Condensation Resistance Rating A Condensation Resistance Rating for a fenestration product is determined by analyzing each section of the fenestration system (i.1 [2] computer program was utilized in this study to obtain center-ofglazing heat transfer. 2) frame areas and glazing areas that have surface temperatures at or below the three prescribed dew point temperatures at 30%.2 ~ 10. standardized conditions were used for the evaluation and rating of the Condensation Resistance rating. (Equation 2) 3) Condensation Resistance rating of the frame. 50% and 70% relative humidity.7 m/s (15 mph) to generate film coefficient of 28. calculations were made to determine the following: 1) total length of each 2-D cross-section (Equation 1).9 ~ (37. edge-of-glazing.4 ~ (59.11 Btu/hr-ft2-F) 9 sky condition of 100% cloud cover The THERM 2.97 W/m2-K (5.3 ~ (50.. Physical Testing for Condensation Resistance Physical testing is used to obtain a Condensation Resistance Rating only in cases where a product cannot be accurately simulated using currently approved NFRC software tools.8 ~ (0 ~ 9 relative humidity (RH) of 30%. CRag (Equation 3) . For a specific cross-section.1 ~ (70 ~ 9 exterior ambient temperature of-17. center-of-glazing. frame and sash.7 ~ 9 wind speed of 6. the indoor surface is subdivided into smaller segments that are no larger than the size of mesh or grid used by the simulation program. The test can be performed simultaneously with the U-factor testing using the NFRC Test Procedure for Measuring the Steady-State Thermal Transmittance of Fenestration Systems. CReog. dividers. 50%.50 THE USE OF GLASS IN BUILDINGS Simulation Requirements for Condensation Resistance Since both temperature and surface film coefficient affect the results. Once the indoor surface is subdivided.e. and 70% providing dew point temperatures of approximately 2. The conditions were as follows: 9 interior ambient temperature of 21. thermocouples are attached to the interior surface of the test specimen in pre-specified locations. edge-of divider and center-of-glazing).5 ~ and 15. For condensation evaluation. CRf.and edge-ofglazing.

and it includes the self-viewing effects between products components. represents a range with 30% and 70% at the extreme ends beyond which human comfort is affected and 50% is considered the optimal humidity level for a home or workplace. the temperature measurement at locations.p = dew point temperature + 0.WISE AND SHAH ON NFRC 500 PROCEDURE 51 S= i (t. center-of-glazing (includes dividers and edge-of-divider).. i Li = length of the surface segment. = temperature of the surface segment. or edge-of-glazing Details on how to determine each Condensation Rating for each section and the whole product are given in NFRC 500. 50%. When the computer modeling is performed with NFRC-approved software.5 F t. -to)*L where (1) t4. The relative humidity called out in NFRC 500. and 70%.and CR~og. The final Condensation Resistance Rating is the minimum value of the CRf. . 30%. measurement of segment areas at or below the dew-point temperature. and subsequent Condensation Resistance calculations are an automatic routine of the software.. CP~og. The simulation also accounts for detailed radiation exchange between the product and the baffle.100 (3) where Ak = area of each fenestration section A = total area of flame. i L = total length of the surface SS= j 3 (2) *A* }1/3t.

radiation heat transfer accounts for almost two thirds of the total surface heat transfer coefficient. This is done because on the interior (indoor) side. NFRC has addressed the size issue for each fenestration type and provides referenced sizes for comparison purposes. while in reality. The product tested in 1997/1998 was composite an aluminum/wood frame window. The third product. namely. for analysis of thermal performance of frame and edge-of-glazing cross sections. which is utilized on indoor window surfaces.52 THE USE OF GLASS IN BUILDINGS Ratings and Sizes Size is an important characteristic when comparing U-factor.1 [4] computer program was utilized to calculate center-of-glazing thermal performance for both windows. WINDOW 4. The current size for each type can be found in NFRC 1002001. low performance frame and high performance glazing. There is a difference between U-factor and Condensation Resistance values. and sidelites/transoms. incorporating high performance double glazing. or vice versa). one significant part of the window could be under condensation in the form of water or even frost. edge-ofglazing. the products were also simulated using NFRC-approved sot~ware per NFRC 500.[3] One size for each fenestration type has been designated for the following general categories: windows. including Condensation Resistance. The minimum of three areas is taken for Condensation Resistance because condensation is local in nature. have been used to analyze the three round robin products. Thus. In all cases. potentially damaging the window and/or surrounding wall structure. Analysis of the Round Robin Tests The proposed Condensation Resistance equations. and Condensation Resistance (CR) ratings from product to product. Solar Heat Gain Coefficient (SHGC). The THERM software program incorporates detailed multi-element radiation heat transfer model. as found in the NFRC 500. radiation heat transfer is only a small fraction of the total surface heat transfer coefficient. These center-of-glazingmodels were then imported into THERM 2. the former is an average index and the latter is the minimum of three areas. the product tested was an aluminum (no thermal break) frame horizontal slider window with very high performance triple glazing. Table 1. The same product size will be utilized for all reported indices. Size is important because of the area ratios of the frame. because extremes may be present in the same product (e. a two-dimensional finite element computer program. It forms in areas that are at or below the dew point. The Condensation Resistance model in THERM also incorporates a . and center-of-glazing. entrance doors. An average number would indicate acceptable performance.1. while on the exterior side.g. double doors/glazed wall systems/sloped glazing. In 1999/2000. an average rating for condensation is meaningless. also known as HeatMirror product.. tested in the 2001 test round robin. Visible Transmittance (VT). was a thermally-broken aluminum fixed window with high performance glazing.

the round robin Condensation Resistance value for the edge-of-glazing (reported value.WISE AND SHAH ON NFRC 500PROCEDURE 53 convection heat transfer model in the glazing cavities.. When compared to test results.ogareas of condensation evaluation. as well as a multi-element radiation heat transfer model. 1999/2000 and 2001 round robin testing and one simulation run for each of the robin test specimens. even when run by different simulators. The 1997/1998 results show good agreement in the CRf and CP. with the exception of the CP-~ in the 1998 test round robin. It is significant to note that the data obtained on tests done on identical units from one year to the next produced different results. in this case) shows a difference of almost 8 points from 1997 to 1998. . Table 1 summarizes the calculated Condensation Resistance values from the 1997/1998. simulation results show good agreement for the center-of-glazing values in all three cases. The 1999/2000 and 2001 test round robin specimen CR values showed variance from the simulation results for the CRf and C P ~ . The simulation values have the advantage of producing consistent results. For example. The 1997 Cr value was in close agreement with the simulated value.

6 71.9 49.9 6.7 66.7 32.1. revised 2001.9 50. Round Robin Year 1997/1998 1999/2000 2001 CRf CRcog CR~og CRf CR~og CReog CRf CReos CReog 69.0 N/A 59.2 N/A 41.7 45. 2002. Conclusion Condensation resistance tests done in adjacent years on identical units have produced different results. [2] THERM 2. A PC Program for Analyzing Windows Thermal Performance in Accordance with Standard NFRC Procedure. When compared to test results.4 First Year Test Second Year Test Simulated Value Note: The values in the tables are the averages of the laboratories that participated in the round robins for the years stated. References [1] NFRC 500.1. Procedure for Determining Fenestration Product U-factors.2 1 8 .0 69.0 48.9 68. accurate Condensation Resistance ratings than that obtained through testing.8 68. 8 68.9 31.1 52. It was observed from the 1999/2000 and 2001 round robin data simulated CRf was considerably lower than the tested value and the CR~ s was considerably higher. while testing relies on pre-determined single point thermocouple locations. Thus. The simulation values have the advantage that they produce very consistent results. simulation results show consistently good agreement for the center-of-glazing values. 3 67. [4] WINDOW 4. This outcome is attributed to the fact that simulation accounts for every segment on the cross section of frame and edge of glazing. Procedure for Determining Fenestration Product Condensation Resistance Values.Condensation resistance average values. which may or may not capture an average performance of that region.54 THE USE OF GLASS IN BUILDINGS Table 1 . [3] NFRC 100.1997.9 43. it can be concluded that the NFRC 500 simulation method produces more consistent. .0 71.9 1 8 .7 67.8 N/A 70. A PC Program for Analyzing Two Dimensional Heat Transfer Through Building Products Procedure. even when run by different simulators.


and Jagota. if we look to the requirements of architectural laminated glass we find that high compliance and an elastomeric response to be unnecessary in many applications. Inc. Van Duser. A. The mechanical properties of a stiff interlayer lead to benefits in three main areas: 1) glass strength. S. In this contribution we discuss the structural performance of laminates made with such a new interlayer: SentryGlas | Plus. a compliant interlayer can hinder structural performance in some cases such as in applications where low glazing deflection is specified. In this contribution we consider a new polymer system for i All authors:E. If we relax the requirement of high compliance after glass breakage.Stephen J. 2) sufficient compliance to ensure minimal head trauma for a passenger that strikes the windshield. A. ASTMSTP 1434. West Conshohocken. balustrades (cantilevers). V.I.. temperature performance Introduction Architectural laminated glass is dominated by the use of Polyvinyl Butyral (PVB) interlayers..DE 19880-0356. such as Butacite | This domination can be attributed to the long successful history of PVB use in the automotive industry for safety glass. an elastomeric interlayer with the correct elastic response to cushion passenger impact was needed. Strength benefits are realized in applications where significant bending stresses develop during loading. Enhanced post glass breakage performance expands the design envelope for glass in horizontal and sloped applications such as overhead glazing. Anthony Smith. However. C. and Anand Jagota Structural Performance of Laminated Glass Made with a "Stiff' Interlayer Reference: Bennison. both before and after glass breakage. "Structural Performance of Laminated Glass Made with a "Stiff" Interlayer. PA. such as: laminates with two sided support.A. Block.. Bermison. 3) temperature performance.astm. many more polymer systems now look attractive for laminated safety glass. All of these structural properties are maintained to higher temperatures as compared to conventional PVB. 2) stiffness and creep resistance. 2002. ASTM International.. Indeed. 57 Copyright9 by ASTM International www. Abstract: The growing demand for laminated glass in building facades and interiors is driving the development of new interlayers that extend the performance of laminated glass. Smith. This latter feature is a major reason why plasticized PVB was developed. glass floors and stairs.. Alex Van . ~C.Wilmington. This new interlayer is significantly stiffer and tougher than traditional PVBs and provides enhanced structural performance in many applications. DuPontde Nemours& Co. glass strength. bolted glass and laminates supported on three sides.J. ed. The primary structural requirements of automotive safety glass are: 1) impact resistance from external objects." Glass in Buildings. Keywords: creep resistance.

The polymer. or strength characteristics. Note that for the glass design stress specified by the horizontal dashed line all laminates display greater load bearing. It has a totally different molecular architecture to PVB and for most applications SentryGlas | Plus (SGP) is below its glass transition temperature (T8 ~ 55~ We describe several examples of the performance benefit of laminated glass made with SentryGlas | Plus (SGP). is significantly stiffer than conventional PVB and displays elasto-plastie stress-strain behavior.. 3) enhanced performance under long duration loading or at elevated temperatures is required. Figure 1 shows the results from one such experiment. In this experiment a grid of 15 3-gage rosette strain gages were attached to the~lass surface of a series of laminates made with either PVB (DuPont Butacite ) or SentryGlas | Plus. 2) post-glass breakage performance is important.g. Strength Performance The benefits of enhanced interlayer stiffiaess for laminate strength can be readily appreciated from a simple experiment in which glass stress during loading is measured directly as a function of polymer type and thickness. such as 1 and 2-sided and point support conditions and/or concentrated loading. e. in overhead glazing. Maximum principal stress was measured at each gage location . 1--Measurement of glass stress as a function of uniform applied pressure for 4sided support monolithic glass plates and several laminate plate builds. SentryGlas | Plus (SGP). The laminate plates were supported on four sides and loaded with uniform pressure. This structural interlayer is particularly effective in applications where: 1) glazing is subjected to bending loads.58 THE USE OF GLASS IN BUILDINGS laminated glass. FIG.

a laminate made from 3 mm glass / 2. for specified loading/support and rate/temperature conditions.Strength design chart for a SentryGlas| Plus -based laminate. Note the specified probabilily of breakage (0.oos / \ .29 mm Butacite| PVB / 3 mm glass shows significantly less glass stress development at a specified applied pressure. 2--. For example. PlateLength(in) o 50 100 ' .BENNISON ET AL.29 mm SentryGlas| Plus / 3 nun glass shows significant strengthening over monolithic glass and laminated glass made with PVB. Figure 1 shows the development of maximum principal stress in the glass during an experiment. load duration (3s) and temperature (50 ~ . i. 2) carrying out finite element analyses of glass stress development [10]. Briefly. 4) combining stress analyses with a statistical (Weibull) glass breakage model [11-13] that incorporates a time-dependence for glass strength [14. a laminate made from 3 mm glass / 2.e. First. at the glass design stress denoted by the horizontal dashed line the SentryGlas| Plus -based laminate can sustain almost 2x the applied pressure than the equivalent monolith. glass thickness.52mm SGIP/3'ram'/ 9 Load(kPa) zao~. Of course the overall laminate thickness is greater than the monolith but in many design codes around the world this laminate would be considered to be structurally inferior to the equivalent monolith. 3-Second =~ 150 i 2000 0 i i i i L i i t 0 1000 2000 ~000 4OO0 Plate Length (ram) FIG.15]. This result is consistent with much research demonstrating that 4-side supported. A design chart for a SentryGlas| Plus -laminate is shown in Figure 2. Second. The chart plots allowable applied uniform pressure contours for various plate sizes. Third. As a guide to design with SentryGlas| Plus -based laminates we have constructed strength charts that aid the selection of polymer type and thickness. Several features may be noted from these data. the upper curve consists of observations from a nominal 6 nun monolithic glass and shows non-linear stress-pressure behavior that is characteristic of large deformation of a 4-side supported plate. Pb = o. the laminate is stronger than the equivalent monolithic glass plate. 3) validating selected analyses against controlled loading experiments [8]. These charts have been computed using a procedure described in detail elsewhere [8]. -3 mm11.008). uniform pressure loaded laminated glass often displays equivalent or enhanced strength over monoliths [1-7]. ON STIFF INTERLAYER 59 along with plate deflection during loading. the procedure consists of: 1) establishing a constitutive model for the interlayer by dynamic mechanical analysis [9].

Note that the 5 m m / 2.29 mm SentryGlas Plus / 5 mm laminate demonstrates greater strength than a 10 mm monolithic glass and essentially equivalent strength to a 12 mm monolithic glass.29 mm Sem~yC. Comparison of this chart with one for an equivalent 6 mm monolithic glass shows strengthening over the whole field of plate sizes. Accordingly.~ I~ 500 400 300 Glass Edge Design Stress= 14 MPa T ~ 23~ I 0. The chart plots allowable pressure loading contours as a function of plate dimensions for a breakage probability of 0.60 THE USE OF GLASS IN BUILDINGS This chart has been constructed in ASTM El300 [16] format using the glass strength parameters and minimum glass thickness specified in that standard [17]. The main feature to note from these measurements is that a beam made from a 5 mm glass / 2. Interestingly. In this figure the force required to produce a specified glass stress has been measured as a function of loading rate in a three-point bend geometry. 1000 8oo 7O0 v v v 60O .29 mm SentryGlas| Plus / 5 rnm laminate significantly outperforms a 10 mm monolithic glass beam in strength properties. Again the total laminate thickness is greater than the equivalent monolith but many building codes around the world neglect this and .001 200 0. I I I I I 2000 -0- 5 ram/2. 3---Measurement of glass stress development for monolithic glass beams and SentryGlas | Plus -based laminate beams tested in three-point bending.008. At these higher aspect ratios a greater degree of bending deformation occurs and suggests that the benefits of SentryGlas| Plus may be realized in loading conditions where bending stresses dominate.0001 I I I I 0. the degree of strengthening depends on plate size and aspect ratio with the greater strength benefits appearing at higher aspect ratios.52 mm SentryGlas| Plus / 3 mm laminate under 3 s wind loading at 50~ A series of such charts for common laminate constructions is used in an iterative fashion to select the optimum laminate build for specified wind-load conditions. Figure 3 shows results of one set of experiments.1 1 10 100 Displacement Rate. dS/dt (ram/s) FIG. For a specified glass design stress (14 MPa) the figure plots the applied force as a function of rate. Note that the chart is constructed for a 3 mm glass / 1. a series of beam bending experiments have been carried out in which glass-stress development as a function of loading has been measured directly using attached strain gages.du(a)P]us 5 mm / I0 nun Monolithic Gins 12 mm Monolithic Gtass k.01 0.

In bendingdominated loeding/support conditions the interlayer shear properties play a significant role in glass stress development as pointed out by Hooper [18. Figure 4 shows the results of a pressure deflection test during loading after glass breakage. assuming that the plate is acting as a pure membrane. Under such conditions SentryGlas| Plus-based laminates are expected to significantly outperform PVB-based laminates. Of course the laminate contains significantly less glass and has a weight advantage over the monolith.29mm Se~tryGlas Flus/3mm t cR) 3mm/2. ~'--Deflection-appliedpressure behaviorfor two laminates tested under uniform pressure. The PVB-based laminate displays significantly greater compliance than a $entryGlas| Plus -based laminate. 50 i ! i i i i i 9 9 40 3mm/2. consisting of 3 ram heat strengthened glass / 2. We demonstrate the benefits of enhanced interlayer stiffness of SentryGlas| Plus -based laminates in the following test.19]. after glass breakage. 4-side support. reveals that the effective modulus of the PVB laminated plate is on the .BENNISON ET AL.29mm PVB/3mm 3O 2O 10 0 2 4 6 8 10 12 14 16 Applied Pressure. P (kPa) FIG.762 m x 1. loaded under uniform pressure to glass breakage. Post-Glass Breakage Performance SentryGlas| Plus was originally developed to enhance post-glass breakage performance in a cycling test component of new hurricane performance standards that are being adopted in certain regions of the USA. Basic plate analysis [20] of the experiment. ON STIFF INTERLAYER 61 penalize laminates against equivalent monoliths. and then the pressure cycled to study the maximum (center) deflection response of the fragmented laminate. When comparing strength performance of the laminate against a 12 mm monolithic glass it can be seen that the monolith and laminate are essentially equivalent. 0. Laminate plates.29 mm interlayer / 3 mm heat strengthened glass have been dry glazed into a frame.219 m.

62 THE USE OF GLASS IN BUILDINGS order of 400 MPa. Figure 5 shows the final . on the order of 55 . 5--. and glass/polymer adhesion. The reduced plate compliance reduces the tendency for interlayer tearing during loading of a cracked laminate and diminishes in-plane motion and tendency for glazing pullout from the framing system. Consider now the SentryGlas| Plus laminate in Figure 4. which is determined by glass strength and glass type. Qualitatively. However. The advantage of this higher Tg is demonstrated in a design example for an overhead glass canopy in which laminated glass is required for safety performance. other factors contribute also to the resulting laminate compliance.60~ the stiffness advantage of SentryGlas | Plus versus traditional PVBs is maintained up to and exceeding the glass transition temperature. Note that the full potential stiffening effect of SGP is not realized due to these other factors and glass fragments play a significant role in plate deformation characteristics. These other factors include fragment size. It is expected that the stiffness of a cracked laminate will scale with interlayer modulus. Analysis of this test reveals that the laminate modulus post-. Thus the stiffness properties of SentryGlas Plus result in a 30-fold increase in plate stiffztess. post-glass breakage.Computed deformed shape of a point-supported laminate canopy after 30 years at 40 %7under self-weight. Temperature Effects The polymer structure of SentryGlas| Plus results in a glass transition temperature. The question addressed in the example concerns the long-term creep performance of the point supported overhead structure under self-weight at 40~ We ask the question: how will the canopy deflect over time and what effect will changing the interlayer from PVB to SGP have on performance? FIG. We approach this problem using our t'mite element design methodology used to construct wind load design charts discussed earlier. small glass fragments or high glass strength will reduce laminate stiffness and low glass/polymer adhesion will reduce laminate stiffness. breakage is on the order of 12 GPa.



deformed shape of a PVB-based laminate canopy after self-weight loading for 30 years at 40~ One advantage of our finite element based approach is that we have constructed constitutive models for our polymers that have full time-temperature superposition capabilities that allow efficient calculation of rate and temperature effects [8,9]. Figure 6 shows the predictions deflection behavior. Loading: self-weight, 40~ 12 mm glass/1,52 mm interlayer/12 mm glass DuPont Laminate Design Engine 50
--O-P V B - D u P o n t B u l a c i t e TM

40 =" .o 30

- - < 3 - - S G P - D u P o n t S e n t r y G l a s TM Plus











I0U I01 102 I(P I& I0-s 10r 107 !0s 10'~ Time, t (s) FIG. 6----Computed maximum deflection o f the canopy shown in Figure 5 as a function o f time under self-weight at 40 ~ Note that the SentryGlas | Plus-based laminate canopy is predicted to display essentially constant deflection over time as compared to the PVB-based laminate, which is predicted to steadily deflect over time. Consider first the upper curve for a PVB-based laminate. It is predicted to deflect over time until reaching a steady state condition after a million seconds or so, The lower curve in Figure I5 shows the predicted response for a SentryGIas| Plus -based laminate of the same build. Note that overall deflections are lower than those predicted for PVB-laminates and that the deflection response is essentially stable with time for these conditions. Again the enhanced stiffness of SentryGlas| Plus versus PVB results in significant performance enhancements in deflection response over time at elevated temperatures.

We have shown that a new interlayer, DuPont's SentryGlas| Plus, demonstrates significant structural performance advantages over traditional PVB in many applications. The enhanced structural performance results from increased interlayer stiffness and higher glass transition temperature compared to conventional PVB. The primary attributes resulting from a stiffer interlayer include: I) enhanced strength, particularly where bending stress states dominate laminate deformation; 2) enhanced laminate stiffness both in pre and post-glass breakage conditions; 3) enhanced



temperature performance. Such performance attributes present architects and engineers with more design options for optimum performance glass facades and structures. It is expected that these attributes will also extend the design possibilities for laminated glass.

1. Behr, R.A., Minor, J.E., Linden, M.P., Vallabhan, C.V.G., "Laminated Glass Units Under Uniform Lateral Pressure," Journal of Structural Engineering, 111 [5] 1037-50 (1985). 2. Behr, R.A., and Linden, M.P., "Load Duration and Interlayer Thickness Effects on Laminated Glass," Journal of Structural Engineering, 112 [6] 1141-53 (1986). 3. Vallabhan, C.V.G., Minor J.E., and Nagalla, S., "Stresses in Layered Glass Units and Monolithic Glass Plates," Journal of Structural Engineering, 11311] 36--43 (1987). 4. Das, Y.C. and Vailabhan, C.V.G., "A Mathematical Model for Nonlinear Stress Analysis of Sandwich Plate Units," Mathematical Comput. Modelling, 11 713-19 (1988). 5. Behr, R.A., and Norville, H.S., "Structural Behavior of Architectural Laminated Glass," Journal of Structural Engineering, 11911] 202-22 (1993). 6. Vallabhan, C.V.G., Das, Y.C., Magdi, M., Asik, M., Barley, J.R., "Analysis of laminated glass units", Journal of Structural Engineering, 119[5] 1572-1585 (1993). 7. Van Duser, A., Jagota, A., Bennison, S.J. "Analysis of Glass/Polyvinyl Butyral (Butacite~) Laminates Subjected to Uniform Pressure" Journal of Engineering Mechanics, ASCE, 12514]435-42 (1999). 8. Beunison S.J., Davies P.S., Jagota A., Van Duser A., Smith C.A., Foss R.V., "Structural Performance of Laminated Safety Glass", presented at Glass Tech Asia 2000, Singapore. 9. Ferry, J.D., Viscoelastic Properties of Polymers, 3'd edition, John Wiley & Sons, (1980). 10. ABAQUS| version 5.8, (1998) Hibbit, Karlsson & Sorensen, Inc., Pawtucket, R.I. 02860-4847. 11. Weibull, W., "A Statistical Distribution Function of wide Applicability," J. Appl. Mech., 18 293 (1951). 12. Davidge, R.W., "Mechanical Behavior of Ceramics," chpt 9, p 132-39, Cambridge Solid State Science Series, Cambridge 1979. 13. Beason, W.L., Morgan, J.R., "Glass Failure Prediction Model", Journal of Structural Engineering, 110 [2] 197-212 (1984). 14. Brown W.G., "A Practicable Formulation for the Strength of Glass and its Special Application to Large Plates", Publication number NRC 14372, National Research Council of Canada, Ottawa (1974). 15. Reed D.A., Fuller E.R. (Jr.), "Glass Strength Degradation Under Fluctuating Loads", Journal of Structural Engineering, 111 [7] 1460-1467 (1984). 16. ASTM Standard E 1300-97 "Determining Load Resistance of Glass in Buildings" in 1997 ASTM Annual Book of Standards, American Society for Testing and Materials, West Conshohocken, PA. 17. Norville, H.S., Minor, J.E., "Strength of Weathered Window Glass", American Ceramic Society Bulletin, 64 [11] 1467-1470 (1985).



18. Hooper, J.A., "On the Bending of Architectural Laminated Glass", International Journal of Mechanical Sciences, 15, pp. 309-323 (1973). 19. Norville, H. S., King, K. W., Swofford, J.L., "Behavior and Strength of Laminated Glass", Journal of Engineering Mechanics (ASCE), 124 [I] 46-53 (1998). 20. Timoshenko, S., Woinowsky-Krieger, S., Theory of Plates and Shells, McGraw-Hill, New York (1959).

revised procedures to determine load resistance of insulating glass. This paper describes the development of NFL charts I Post-Doctoral Research Associate. and three sides. Block.. Lubbock.. three-sided support.S. M. H. . 2 Director. Texas 79409. ASTMSTP 1434. two. This paper discusses the development of load resistance charts for rectangular glass supported on three sides with one side free. Department of Civil Engineering. 2. It presents the nonlinear finite element model used to determine stresses and deflections. and procedures to determine load resistance of rectangular glass supported on 1. Keywords: glazing design. and 3 sides. West Conshohocken. Abstract: The forthcoming version of Practice to Determine the Load Resistance of Glass in Buildings (E 1300-02) will contain several new features. EI-Shami ~and H. maximum stress Introduction The next revision of ASTM Practice to Determine the Load Resistance of Glass in Buildings (E 1300-02) contains nonfactored load (NFL) charts for glass supported along one. Scott Norville2 Development of Design Methodology for Rectangular Glass Supported on Three Sides to Resist Lateral Uniformly Distributed Loads Reference: E1-Shami. Lubbock. The paper describes load-induced stresses along the edge of an unsupported side and over the glass surface for glass supported along three sides. Department of Civil Engineering. Texas Tech University. These include new procedures to determine load resistance of laminated glass. ASTM International. V.Mostafa M. Texas Tech University. The load resistance charts use maximum edge stress as the criterion for determining load resistance. and Professor. M. and Norville. "Development of Design Methodology for Rectangular Glass Supported on Three Sides. finite element analysis. 66 Copyright9 by ASTM International www.astm. Box 41023. It compares probability of breakage based upon surface stresses with that derived from the maximum edge stress criterion. Texas 79409. PA. The value of maximum edge stress derives from a failure prediction approach for glass edges rather than surface stresses. Glass Research and Testing Laboratory." The Use of Glass in Buildings. Box 41023. Ed. Glass Research and Testing Laboratory.



and corresponding deflection charts for rectangular glass simply supported along three sides with the fourth side free. It describes the nonlinear finite element model ffEM) used for the analysis, the boundary conditions, the maximum edge stress criteria for determination of NFLs, and a simple example that illustrates the use of the NFL charts and compares them with results obtained using a linear analysis.
The Finite Element Model

In this section the authors describe the stiffness matrix formulation for the nonlinear FEM for thin plate analysis. The authors employ Mindlin plate theory [1]. Mindlin plate theory makes the same assumptions as does Von Karman theory [2] except that the final rotations of lines normal to the plate's middle surface are obtained by adding the respective derivatives of the lateral displacement function w(x,y) with respect to x and y with the corresponding vertical shear deformations. The displacements used at any point (x,y,z) are:

u.(x,y,z)=u(x,y)+ zOr(x,y ) v.(x, y,z) = v(x, y)- zO,(x, y)
w. (x, y, z) =w(x, y)


where u(x, y), v(x, y) and w(x, y) denote the displacements of the middle surface and Ox and Oy denote the rotations of the normal in the undeformed plate in they-z and xz planes, respectively, after deformation. Figure la shows an element before deformation. Figure Ib shows the same element after deformation and illustrates that, according to Mindlin plate theory; lines normal to the middle surface before bending do not necessarily remain normal to the middle surface aider bending. Mindlin plate theory represents the relevant nonlinear strain vector as:

s = 7"~


to j



denotes the linear component of membrane strain,



Z~ W

--0 X

~ !

v=-ZO x

~'YZ=!~ W,y-x~ "'~



.~W,y , y,v


l a-Undeformed Element

Z~ W~


v=-zO x

'wYy-~~ "~
kL..f Figure lb.

, y,v

Deformed Element: Solid Lines Indicate Rigid Body Displacement and Dashed Lines Indicate Final Position After Element Deformation



{~o}=~y,.-0~.y denotes the bending strain,




{~:)~={,, +~. -o, +w/
denotes the shear strain, and mr I w 1 }r



denotes the nonlinear component of membrane strain. In terms of the shape function for a 9-node quadrilateral element, the displacements and rotations are:
9 9 9

u = E N,(~,~)~,,

~ = EN,(~,~),,,
ifl 9 i=l

~ = E N,(r



where N,(r 17), i = 1,9 denotes the shape functions of the 9-node quadrilateral elemem [3]. The terms r andr] denote coordinates used to transform the quadrilateral element into a square. The nodal displacement vector


w, o., o,


o.. o.}~


has 5 degrees of freedom [dot'] per node. Hence, with 9 nodes, each element has 45 dof. Since, both displacements and rotations are interpolated using quadratic functions, with three nodes on the boundary, both displacements and rotations have complete compatibility along the inter-element boundaries of the quadrilateral elements. Figure 2 shows a quadrilateral elemem with 9 nodes and indicates the dofat one node. The following equations define the relationships between the strain vectors and nodal displacement vectors.

{~:)f ,


.... V x Figure 29-Node Quadrilateral Element with Degrees of Freedom Indicated at Node 1.] and the stress-strain relationship is (13) . [B]= [Bol+ [B.. . and 11) The equations above combine to form a new matrix for linear relationships: 0 Thus. . the complete relationship between the strain and nodal displacement vectors is the sum of the linear and nonlinear parts. Y ~yl xl .. .70 THE USE OF GLASS IN BUILDINGS z ~ .

.v2) ' vEt s (1 8) D~2 = D2' = v ~12(1 Gt 3 ' (19) Da3 = and 12 .v ) / 2 0 0 D33 0 0 0 0 0 0 (17) D21 D22 0 D44 0 0 0 /955 E=71. (15) [zr]. 1.4 xl06 psi) denotes Young's modulus of elasticity.. The individual elements of [Db] are as follows: Et 3 Dl~ = D22 = 12(1 . in which[D] denotes the elasticity matrix (14) [D]= In the elasticity matrix E'"I E.22 denotes Poisson's ratio. Gt (21) . such as glass. and t denotes glass thickness.2 For linear elastic materials.EL-SHAMI AND NORVILLE ON RECTANGULAR GLASS 71 {r = [D]{ 6}. (20) D44 = D55 = .2) and "D. v=0..(i_. [D h] = 0 0 0 /)12 0 0 0 1 o (16) 0 (1.~ + D~ .7 GPa (10.

[K~][3]. The tangent stiffness matrix. [Kr]. Several authors [4-6] have demonstrated the FEM's efficacy in analyzing thin monolithic plates. in which (24) [Ko]-. Figure 3 shows the boundary conditions for a plate supported along three sides with one side unsupported in detail.]+[K. .]-V Several authors have described Mindlin plate theory and solution techniques associated with it. The equilibrium equation in its incremental form is: (22) [Kr]{AU}={AF} (23) where [Kr] denotes the tangent stiffness matrix.[BoY[D][BoldV. Boundary Conditions Since the glass is simply supported along three sides with one side free. it allows the glass to rotate at the supported sides. and the initial stress stiffness matrix.]. V (25) (26) and [x. The authors use Newton-Raphson techniques [1] in their solutions. the model allows the simply supported sides to slip in the direction of the undeformed plane of the glass.].72 THE USE OF GLASS IN BUILDINGS E G = ) 2v-1 ~ ( denotes the modulus of rigidity. [G]=[Ko]+[K. the nonlinear stiffness matrix. and [AF] denotes the nodal force increment vector. [Kj]. {AU}denotes the nodal displacement increment vector. and it prevents any deformation normal to the unreformed plane of the glass along the supported sides. consists of three parts: the linear stiffness matrix. including E1-Shami et al [4] and Vallabhan and EI-Shami [5]. noting that the authors conducted the analysis on half the plate and employed symmetry about the glass centerline that lies perpendicular to the unsupported side. [K.(. Therefore.

= o At p. lnt Cr v.~. Application of their failure prediction methodology to the general case of an edge under stress indicates that a 60second equivalent failure stress that would initiate fracture at the edge with a probability of 0.o. . The authors used this maximum edge stress criterion to develop monolithic load resistance (LR) charts for glass supported along one.Y // C o X at po=t a. and three sides. w. Ox=O D . . o~ =o At point D.EL-SHAMI AND NORVILLE ON RECTANGULARGLASS 73 Z Simply Supported A w. Walker and Muir [7] developed a failure prediction model based upon edge stress analyses. v.6 MPa (2400 psi) as the maximum edge stress that would initiate fracture with a probability of breakage of 0.008 is approximately 13.008. Converting this slress to a 3-second duration equivalent stress gives a value of approximately 16.OyOB/ w. Edge Stress Criteria The maximum edge stress criterion used for these calculations comes from work performed by Walker and Muir [7] in which they performed bending tests on glass louvers to determine edge failure strengths.8 MPa (2000 psi). o~. two. Ox= O Figure 3- Boundary Conditions for a Plate Simply Supported Along Three Sides with One Side Free.

1 .9 psf) loading having 3 second duration is P~ = 1.9 psf).) dimension unsupported under a lateral load of 1. The maximum lateral deflection also occurs at this location. v v Figure 4- Contours of Maximum Principal Tensile Stress for a 965 x 1930 x 6 mm (38x 76x 1/4 in.) Lite Supported Along Three Sides with One Side Free. . The authors used the stress distribution depicted in Figure 4 in conjunction with the failure prediction methodology with m = 7 and k = 2. x 76 in. as in the linear case.65x10 "s. The LR for this lite is very close to 1.0 kPa (20. based upon the maximum edge stress criteria of 16.86x 10"53N'Tm to compute the probability of breakage based upon a ~2 fracture originating on the surface.0 kPa (20.74 THE USE OF GLASS IN BUILDINGS Stresses in Rectangular Glass Supported Along Three Sides Figure 4 shows contours of maximum tensile stress for a nominal 6 mm (1/4 in. nonlinear analysis indicates that the maximum tensile stress occurs at the mid-length of the unsupported side. For this lite. This strongly indicates that fractures occurring along the edge of the unsupported glass side provide a limiting criterion for determining LR.6 MPa (2400 psi). This is the methodology to compute probability of breakage for glass supported along four sides. The authors observe that.9 psi').0 kPa (20.) with one 965 mm (38 in.) glass lite having rectangular dimensions of 965 mm x 1930 mm (38 in. the probability of breakage under a 1.

4 psf)..... ..) glass will have insufficient LR.. " ~' '~' '" " "' 0. From comments above. Depending on the LR determined. .'~ 2o!i 10 i ' 0 ~ ' ' ' 50 0 0 0 1000 2000 3000 4000 5000 Length of Parallel Supported Edges (mm) Figure 5- Nonfactored Load Chart for 8 mm (5/16 in.1 40 4. (b) try a larger thickness..-~...~"~.0 x 1.....9 psf . The glass type factor (GTF) for annealed glass under short duration load is 1. or (c) try a smaller thickness.~ ..-. . Nonfaetored Load (kPa) 2000 E 60 1. Figure 5 shows the nonfactored chart for this nominal thickness with a horizontal line projected from 965 mm (38 in..i-~ . The designer picks a trial thickness designation and determines its LR for the given dimensions.i ...) with one 965 nun (38 in.) along the vertical axis and a vertical line projected from 1930 mm (76 in.. ~ .8 psf)..) Glass Simply Supported Along Three Sides.'~.. ..-~---~--... 6 mm (1/4 in...75~1-I-//T" 160 180 200 2500 (27) i 80 70 ! tU 0 12...~" ~'~' '" ' '~'" .. Three Sides Simply Supported f . ! i "Pb =0-008 ... Length of Parallel Supported Edges (in.{-~. .:~ .6 kPa (33....0 kPa (41.. Experience with ASTM E 1300-02 significantly hastens this process... ~ ......:'..EL-SHAMI AND NORVILLE ON RECTANGULAR GLASS 75 Design Example The authors would now like to determine the required thickness for an annealed glass lite having rectangular dimensions of 965 mm x 1930 mm (38 in..) glass....... i ~ ~ .../ lkPa = 20...) 0 100 20 40 60 80 1O0 120 140 ' " 1 ' ' ' " / I .....6kPa = 1..! 'i .. / " ..) long side unsupported to resist a uniform loading of 2. x 76 in. hence the authors try 8 mm (5/16 in... The LR is: LR = GTF x NFL = 1..) along the horizontal axis. the designer earl (a) use the trial thickness.. i ~ ~.. i / / ! 3-Second Duration . .4psf)....7': I1 i ' . i "6 tt..00 30 ~:-+. They also wish to determine the approximate maximum deflection for the lite.. . ~-~.....0.i.6kPa(33..~---~i -J--. ~ 1500 "~ [IJ ~) 9 . .. The lines intersect at about 1. The process of determining the required thickness using the NFL charts from ASTM E 1300-02 is iterative.

While developing the NFL charts.20 kPa. The authors next try glass with a thickness designation of 10 mm (3/8 in. For glass supported on three sides. Hence. The particular redimensionalization depends upon the support conditions.) Glass Simply Supported Along Three Sides.- 0 1000 2000 3000 4000 0 5000 Length of Parallel Suppofled Edges (mm) Figure 6- Non-Factored Load Chart for 10 mm (3/8 in. This LR is greater than the specified loading. Each load chart in ASTM E 1300-02 has a corresponding deflection chart. the redimensionalized load takes the form: .20kPa = 2.) annealed glass is insufficient. The authors could have used a smaller thickness designation and specified heat strengthened or fully tempered glass with a smaller thickness.) 110 100 90 80 "O 0 20 40 60 80 100 120 140 160 180 200 2500 2oo0 E 70 60 50 40 30 20 10 0 500 LU . The value of NFL is approximately 2.8psf) (28) Length of ParalleI Suppofled Edges(in.20kPa(41. so annealed glass with 10 mm (3/8 in.J 1500 LU LL 1000 .). the LR is: LR = 1. Presentation of deflections in this form saves some calculation on the part of the designer but it also requires some familiarization for efficient use.0 x 2. The horizontal axes of the deflection charts contain various forms of "redimensionalized" loading.) glass will suffice. Figure 6 shows the appropriate NFL chart with the lines projected on it. For glass supported on three sides. the authors also developed deflection charts. the maximum deflection occurs at the center of the unsupported glass side.76 THE USE OF GLASS IN BUILDINGS The value of LR for 8 mm (5/16 in.= LL tr-.

. .. For this example... The sloping lines.... ........) Glass Three Sides Simply Supported Deflection va (Load x L 4) 0.2 0. w denotes the uniform load.../! ~'/ .. ..0.. and L denotes the length of the unsupported side of the glass.. . fl .P"i iI.... . ..:=:... . m2) [L Denotes Length of Free Edge] Figure 7... :e'v" 35 E E 30 . The authors described the development of the three sided NFL charts for ASTM E 1300-02........8 ] 10. The authors "read" the approximate maximum deflection from the intersection of the horizontal line with the vertical axes as 4.0 mm (3/8 in... the redimensionalized load is: {v = w x L 4 = (2.5...... ... ..ft 2) [L Denotes Length of Free Edge] 30 ' 40 50 60 70 80 I~ ' F L~[ ! '1 I ' I f 90 100 110 '1 ' I i ' j ~ .r / I l......2 mm (0.... so 45 40 0 1' ~ 10 20 .17 in..8 1.20kip .6 0.o . y . ..o ~ 20 15 10 5 ....... .. = 'i 1 ! .965m) 4 = 1... Previous .)Glass Simply Supported A l o n g Three Sides... ......... ....'.. I I .......6 1.. 1. Conclusions Deflection Chart f o r 10 m m (3/8 in. Load x L4 (kip.fl2 (30) Figure 7 shows the deflection chart with a vertical line projected from 1. E 0..20 kip't~) and a horizontal line projected from the intersection of the vertical line with the deflection line for AR > 1.. .EL-SHAMI AND NORVILLE ON RECTANGULAR GLASS 77 }} = w x L4 (29) in which ~ denotes the redimensionalized load. curving for small thickness designations. e- 15 25 -Z/i..00kPa) * (0... ..... I..0 0 5 10 15 20 25 30 35 40 45 Load x L4 (kN...4 . r Z /i y . ." ..... Their discussion included descriptions of the FEM used for analyses and the design failure criteria for glass with one or more unsupported sides.4.. ! .. .. ......... /q/ i r ....73kN'rrl2 (4..73kN .J..4 I s / r/ 1~+/ .).... ..0 t ... relate the redimensionalized load to maximum deflection. 7 .m ~ = 4. . ..

O. M. New York." McGraw-Hill. S. M. D. 1989. August. March. and Muir. G. [4] E1-Shami. The formulation presented herein uses Mindlin plate theory coupled with nonlinear von Karman relationships. "Comparison of Nonlinear von Karman and Mindlin Plate Solutions. Texas Section. I. London. [3] Zienkiewicz. L. "An Investigation of the Bending Strength of Glass Louvre Blades. Texas Section. R. C. Texas. A. April 49." John Wiley & Sons. New York.M. Texas. M. "Solid Mechanics: A Variational Approach... H. Vallabhan.. S.78 THE USE OF GLASS IN BUILDINGS analyses [4] of thin glass lites used Kirchhoff plate theory coupled with nonlinear von Karman relationships to obtain a solution. Sydney. The Mindlin plate theory accounts for shear deformation which is insignificant in thin glass lites but becomes significant for thick glass lites. For thin glass lites supported on four sides. [2] Dym. Houston. and Shames... References [1] Cook.. M. and Norville. C.." Proceedings: 9th Australian Conference on the Mechanics of Structures and Materials. In addition the authors presented an example that illustrates the use of the NFL charts and deflection charts for rectangular glass with three sides simply supported and one side free. L. [7] Walker. M. Spring Meeting." Proceedings of the American Society of Civil Engineers. R. Australia. V." Journal of Engineering Mechanics. and Plesha.. the results obtained using Mindlin plate theory are identical to those obtained using Kirchhoff plate theory [4]." Proceedings of the American Society of Civil Engineers. . G. C. "The Finite Element Method. V.S. 2001. 1977. C. "A New Model for Analyzing Architectural Laminated Glass. American Society of Civil Engineers.. H. and Kandil.. San Antonio. [6] E1-Shami. Spring Meeting. "Concepts and Applications of Finite Element Analysis.. 1973." McGraw-Hill. G. 1997. M. Malkus. 1984. in review. D.M. E. [5] Vallabhan. E1-Shami."Comparison of Rectangular Glass Plates with Area-Equivalent Trapezoidal Plates.

Box 41023. ASTMSTP 1434." The Use of Glass in Buildings. Box 41023. EI-Shami. These window 1Director. R.astm. The paper summarizes the experimental data. Atlanta NAS Implementation Center. Jackson. Glass Research and Testing Laboratory. ASTM International. Ed. The authors have developed a design methodology for these lites as discussed in this paper.. Terminal Platform. Georgia. Lubbock. They lean outward to provide relatively unobstructed and distortion-free views of runways and taxiways for ground traffic controllers. Lubbock. airport control towers. weathered window glass lites. 30320. M. Lubbock. and Professor. Abstract: Large trapezoidal annealed window glass lites glaze airport control tower cabs. In developing the methodology. Keywords: glazing design. Texas 79409.H.. under uniform pressure. loadinduced stresses in large trapezoidal window glass lites. It presents the design methodology and illustrates it with an example. Block. G. "Design Methodology for Large Trapezoidal Window Glass Lites. Department of Civil Engineering. airports. Texas Tech University. El-Shami. Texas Tech University. the authors conducted failure tests of large. Atlanta. Glass Research and Testing Laboratory. finite element analysis. S. 4 9 9 9 Senior Project Engineer. H. They used a nonlinear finite element analysis technique for thin plates coupled with failure prediction techniques that underlay ASTM Practice for Determining the Load Resistance of Glass in Buildings (E 1300-00) to develop their design methodology.S. l Mostafa M. 3 Research Assistant. West Conshohocken. Texas Tech University. trapezoidal. Federal Aviation Administration. 2 Ryan Jackson. V. ANI-340. The fact that they lean outward requires the lites to have a trapezoidal surface area. PA.. Texas 79409. maximum stress Introduction Large trapezoidal window glass lites and window glass constructions glaze control tower cabs at all U. obtained from the U. Department of Civil Engineering. and the load resistance of large trapezoidal lites relative to that of rectangular window glass lites. and Johnson. M. failure prediction. 2 Post-Doctoral Research Associate. 2002. Federal Aviation Administration. 79 Copyright 9 by ASTM International www. Texas 79409. Department of Civil Engineering. PO Box .S. Scott Norville. Glass Research and Testing Laboratory.. 3 and George Johnson4 Wind Load Resistance of Large Trapezoidal Glass Lites Reference: Norville. Box 41023..

"Standard Practice for Determining the Load Resistance of Glass in Buildings. At present. presently under design. may use large trapezoidal window glass lites supported only along their top and bottom edges. The FAA provided numerous specimens from which GRTL selected two relatively small but consistent samples. laminated glass (LG) with annealed plies. GRTL staff conducted nonlinear stress analyses in conjunction with destructive tests of large. The top side had a length of 2705 mm (106.80 THE USE OF GLASS IN BUILDINGS glass lites may consist of annealed monolithic glass. and design load resistance from Weibull parameters in ASTM E 1300-00. Monolithic Specimens Separatedfrom IG Units GRTL staff separated the individual lites from the IO units and tested them to failure under uniform pressure. GRTL staff began with the IG units since they arrived first. For both window glass samples. . GRTL stored them until inception of the project. This paper presents: (a) experimental results from failure tests. in U. One sample consists of nine IG units that had been in service approximately 21 years.S. Experimental Procedure and Data The first step in this project consisted of obtaining samples comprised of several large trapezoidal window glass specimens from airport control tower cabs undergoing demolition or replacement. The Glass Research and Testing Laboratory (GRTL) has undertaken a project to develop a methodology for the FAA to use when designing window glass in control tower cabs. (c) strength comparisons between load resistance data for the sample of large trapezoidal lites and load resistance data from published research. FAA window glass designers use approximate methods that typically result in a thickness larger than necessary to resist design wind loads. The top and bottom sides of the lites were parallel. or insulating glass (IG) units fabricated with annealed lites. (b) analysis results. weathered trapezoidal window glass specimens under uniform load.5 in.). The other sample consists of eight relatively new large trapezoidal LG lites. framing members support all four sides of the trapezoidal window glass lites and window glass constructions.S. and (e) a simple design example to illustrate the procedure for a trapezoidal lite supported on four sides. The authors note the removal from a control tower cab and subsequent storage of these IG units occurred some time before their shipment to GRTL facilities. Ground traffic controllers require annealed glass because it provides nearly distortion-free viewing." (d) contours of maximum stress in large trapezoidal lites. In producing the design methodology. The FAA shipped them to GRTL approximately six months before this project began. The bottom side had a length of 2310 mm (90 in. the U. GRTL had no role in their removal and realizes that the lites underwent significant handling that probably caused some surface damage. Instead. airport control tower cabs.). Currently. New towers. Federal Aviation Administration (FAA) has no regular design methodology to guide in the selection of proper glass thickness for the large trapezoidal lites to resist design wind loadings. The design methodology relates uniform loading having a specified time duration to probability of breakage.

e. The lites had a nominal thickness of 10 mm (3/8 in.7 psi). one fracture originated along the sloping edge. i. The authors summarize the failure pressure data for the monolithic specimens as follows: the mean failure pressure for 13 specimens was 2. The coefficient of variation for the 60-second equivalent failure loads was 34. The authors converted the raw failure data to 60-second duration equivalent loads as defined in ASTM E 1300-00. Curtain Walls.1%. and the remaining fractures originated in the right angle comers. Analog signals from the pressure transducer and the linear variable displacement transducer went to an analog/digital (A/D) converter. and Doors Under the Influence of Uniform Static Loads by Destructive Methods.6 psi). or shipping prior to their delivery to GRTL facilities. two originated in central areas of the trapezoidal lite. researchers carefully measured and recorded the location of the fracture origin.). They monitored center of glass deflection using a linear variable displacement transducer.7 psi).730 kPa (15.82 kPa/min (38.NORVlLLi:: ET AL. Three fractures originated in the obtuse comer of the trapezoid. and then measured their locations from a common reference at the lower left comer of the glass when mounted in the test frame. and the coefficient of variation for the failure pressures was 32. The mean 60second equivalent failure load for the specimens was 2.38 kPa (49. The A/D converter sent digital signals to a personal computer at a rate of 2 Hz. The graphing allowed GRTL staff to monitor the load and control its rate on a continuous basis during each test. The top and . a lateral load. Laminated Glass Specimens The series of tests on LG specimens followed the same fundamental procedure as described above except that this series required no separation of lites. The authors note that the variability for the loads was high." Researchers evacuated air from the vacuum chamber at a controlled rate to produce a uniform pressure difference. They increased the lateral load linearly with time from the inception of loading at a rate of 1. the standard deviation of the failure pressures was 0.2%. acting across the window glass lite. GRTL staff located fracture origins for both the inside and the outside glass plies. Handling during this period most certainly inflicted significant damage to the glass surfaces that resulted in lower load resistance values. The computer simultaneously wrote the data to a file and graphed it on the computer monitor. Following fracture. Surface damage results in experimental data falling below typical load resistance values and possibly having higher variability. the process of separating the lites from IG units induced evert more damage. Fracture origins occurred in the following locations. The standard deviation for the 60-second equivalent failure loads was 0. This test series contained eight specimens. Researchers monitored pressure using a piezoelectric pressure transducer calibrated against a mercury manometer before each test.14 kPa (44.3 psf)..0 psf/min) until the specimen fractured. Researchers placed each lite in a test chamber that provided support conditions and glazing stops in compliance with ASTM E 997-00 "Standard Test Method for Structural Performance of Glass in Windows. The authors again point out that GRTL staff did not supervise the lites' removal from the airport control tower cabs. ON TRAPEZOIDAL GLASS LITES 81 The vertical dimension of each lite was 2110 mm (83 in. subsequent handling. In addition.754 kPa (15.). In addition.

All other steps remained the same as described for the monolithic test series.). 2. The mean 60-second equivalent failure load for the specimens was 20. They compared deflection data with finite element model (FEM) predictions. High values of maximum principal tensile stresses also occur near the other comers. Ten of the 13 fracture origins from the monolithic specimens occurred in areas of relatively high maximum principal tensile stress.5 kPa (469 psi').4]. occurs in the area near the obtuse angle of the trapezoid. Acting for a 60second duration. Values of maximum principal tensile stresses on the contour lines in Figure 1 are in MPa.8 MPa (3300 psi).2 kPa (422 psf).4 psf).). It uses Mindlin plate theory [5] that considers shear deformations through the thickness of the plate. The standard deviation for the 60-second equivalent failure loads was 3. mounted them to the test chamber. GRTL staff classified all fractures on the tension surfaces of the LG specimens as occurring in the center of the specimens. Glass Deflections GRTL staff measured deflections during all tests using a linear variable displacement transducer (LVDT). The bottom side had a length of 2030 mm (80 in. The authors converted the raw failure data to 60second duration equivalent loads as defined in ASTM E 1300-00. The authors observe that the highest magnitude of the maximum principal tensile stress.5%. 3. The lites had a nominal thickness of 38 mm (1-1/2 in.86 kPa (86. The authors summarize the failure pressure data for the LG specimens as follows: the mean failure pressure for eight specimens was 22. the standard deviation of the failure pressures was 4. In Mindlin plate theory. The coefficient of variation for the 60-second equivalent failure loads was 20.008). one of the authors conducted a nonlinear FEM analysis for the trapezoidal plates. and the coefficient of variation for the failure pressures was 18.80 kPa. The authors discuss the comparison below.3 kPa/min (319 psf/min) sampled at a rate of 2 Hertz.7 psf). a plane normal to the middle surface before deformation remains straight but not necessarily normal to the middle surface after deformation. GRTL staff tested the LG specimens using a load rate 15. Figure 1 presents contours of maximum principal tensile stresses on the surface of the trapezoidal lite associated with a uniform lateral load of 1. The top side had a length of 2260 mm (89 in. and loaded them. this load corresponds to a probability of breakage of 8 lites per 1000 (Pn = 0. using the failure prediction methodology in ASTM E 1300-00. The vertical dimension of each lite was 2260 mm (89 in.). The nonlinear FEM predictions of deflection agree with experimental measurements to within about 5% for both monolithic . Several references describe the model [1. Numbers along the sides of the graph in Figure 1 indicate distances in mm. Stress Analysis While GRTL staff destructively tested the individual lites of the IG units. approximately 22. GRTL staff removed the LG specimens from their packing crates. In summary.).15 kPa (86.82 THE USE OF GLASS IN BUILDINGS bottom sides of the lites were parallel.5%. None occurred in any proximity to the obtuse comer or other high stress areas.

the authors produced maximum principal stress contours using equivalent loads that.y)v~ (x.y).008.00 1300.NORVILLE lET AL.00 =.V' Figure 1-Contours of Maximum Principal Tensile Stress (MPa) in a Ttrapezoidal Lite. 7. 500. and ~m~(x. The failure prediction methodology simply represents a Weibull probability distribution for window glass strength as given in the following equation: P. c(x.y) denotes maximum equivalent principal tensile stress at .00 1000. y)r d-~dy area (2) in which m and k denote the Weibull parameters. This observation provides a high degree of validation to the nonlinear FEM analysis technique. The risk function takes the following form: s--.00 2000. correspond to a probability of breakage of 0. Hcc(x.00 2500. ON TRAPEZOIDAL GLASS LITE$ 83 and LG specimens. Failure Prediction In the previous section. =l-exp[-B] (1) in which Pa denotes probability of breakage and B denotes a risk function. To determine the probability of breakage.y) denotes a stress correction factor at location (x. the authors coupled the failure prediction methodology [6. if acting over a 60-second duration.8] with the nonlinear finite difference stress analysis.

in other words.0~1 J 0. the failure prediction methodology indicates that the probability of breakage decreases or.5~ represents probability of breakage versus 60-second equivalent failure load for a trapezoidal lite with the dimensions of the monolithic specimens tested in this research. the load resistance associated with a given probability of breakage increases.5 3 3.5 ~ represent probability of breakage versus 60-second equivalent load for trapezoids with the acute angle of the sloping side decreasing. Figure 2 shows curves representing probability of breakage versus 60-second duration equivalent loads for various 0.5 Lold 2 IkPII) 2.7 and k = 2. geometries of lites based upon the Weibull parameters in ASTM E 1300-00 and the thickness of the monolithic specimens. namely m -. the authors used values of the Weibull parameters consistent with those in ASTM E 1300-00. The two curves on the fight ((0-=45~ and 0=39.84 THE USE OF GLASS IN BUILDINGS location (x. Although the magnitude of the principal stress in the obtuse comer increases.y).004 0. moving to the right (0=78. . The conclusion is not exhaustive for all conceivable polygons that the rectangle might encompass.5 Figure 2. For the probabilities calculated herein. The next curve.5 '1 1. The curve farthest to the left (0-90 ~ represents probability of breakage versus 60-second equivalent load for an encompassing rectangular lite.01 0.86 x 10"53N7m 12(1.002 0 0 0. its area decreases.37 x 10-29in) 2 lb'7). As the lite moves from a rectangle towards a triangle.0~6 ~ 0.Probability of Breakage Curves versus 60-Second Equivalent Load for Rectangular and Trapezoidal Lites. at least for trapezoids and triangles. It indicates that designing for an encompassing rectangular shape should be eonservative.

the Weibull parameters from the trapezoidal monolithic sample indicate that it might be weaker than rectangular samples for whatever reason. and then using a Kolmogorov-Smirnov goodness-of-fit test to determine how well the constructed probability of breakage distribution modeled the experimental distribution. constructing a curve representing probability of breakage versus 60-second equivalent load. the authors surmise that magnitudes of wind loads used for design exceed magnitudes of wind loads that design wind events produce on in-service window glass lites. . regardless of the degree of care taken. NorviUe and Minor [7] used a comparison similar to this. Finally. This observation indicates that significant numbers of window glass lites designed using ASTM E 1300-00 should fracture during every occurrence of a design wind event.) and nominal 6 mm (1/4 in. even during a design wind event.e. Design Procedure General Design Criteria Figure 2 indicates that basing the design of trapezoidal window glass. Figure 3 shows this curve along with curves based on Weibull parameters from other weathered glass samples [7] and those that form the basis for the non-factored load charts in ASTM E 1300-00. It also indicates that the Weibull parameters in ASTM E 1300-00 result in significantly higher load resistance than any data from weathered window glass samples would justify. the authors believe that the handling window glass samples undergo in testing.T6)adequately described the load "4' resistance distribution of the monolithic specimens separated from the IG units.NORVlLLE ET AL. Hence. Finally. i. In addition. The most prominent reason is that the design wind load.001. Furthermore. simply supported along all edges. Jackson [9] also found that the Weibull parameters m = 4. Figure 3 indicates that the monolithic specimens discussed above display lower load resistance than do any previously tested samples. a load having constant magnitude for 60 seconds never occurs. ON TRAPEZOIDAL GLASS LITES 85 The authors next used the statistics from the failure loads to estimate WeibuU parameters for the experimental data from each test sample described above.37 x 10-20lb'4"Sin.98 x 10"36N sm7"6(1. and a probability of breakage of 0. Jackson [9] gives the details of this analysis. The authors feel that this breakage does not occur for several reasons.) thickness using these Weibull parameters.8 and k = 7. The authors constructed a probability of breakage curve versus 60-second equivalent failure load for a lite with rectangular dimensions 1524 x 1219 mm (48 x 60 in. significantly reduces the load resistance of the test specimens. That breakage does not occur. Jackson [9] found that the Weibull parameters from ASTM E 1300-00 adequately described the load resistance distribution of the LG sample. ASTM E 1300-00 Weibull parameters. the design of trapezoidal lites for airport control tower cabs simply supported along all four sides will be based on the load resistance of an encompassing rectangle. on the dimensions of the encompassing rectangle should prove conservative. an airport control tower cab is a critical facility.. This procedure consisted of an iterative procedure of trying a set of values for m and k.

4 Weathered Glass Window [7] d3 O Ix 0.32 kPa (49.6 0.4 psi).'>~ ~ 0.) and 2290 mm (90 in. Currently GRTL personnel are producing load charts and computer software to facilitate this design methodology for the FAA. ~/ ~ / / 0. The length oft_he side perpendicular to the horizontal sides is 2540 mm (I00 in. Design Example The authors wish to determine the thickness designation for a LG trapezoidal life. Select a trial thickness designation of 19 mm (3/4 in. The charts are not ready for publication at this time.6 0.2 0.0 3"1>' i t i i i i 0. 6 I 8 I 10 I Monolithic Trapezoidal Specimens .001 is 2.31 MPa (1350 psi). For 60second duration loadings the maximum design stress based on an edge stress analysis as Walker and Muir [10] presented is 9. to resist a 60-second duration wind loading of I kPa (20.2 MPa (1620 psi).9 psl). 1.0 0. this value is 11.).86 THE USE OF GLASS IN BUILDINGS Pressure (kPa) 0 2 . the maximum tensile stress along the edges of the unsupported sides will control the design.8 o= 0O "~ .2 0./. . simply supported along all four edges.4 0. For 3-second duration loads.8 Pressure (psi) 1. The horizontal sides have lengths of 2540 mm (I00 in. The component .4 1.).2 1. The non-factored load associated with a probability of breakage of 0.6 Figure 3.0 1.. For trapezoidal lites supported only along the top and bottom sides.). . f ~ f /~.Probability of Breakage versus 60-Second Equivalent Load for a Common Lite Size. The lite slopes outward at I 5~ from the vertical.0 4 1 .

). The long duration load resistance far exceeds the component of glass weight acting normal to the glass.) under the design wind loading. (6) (5) Without showing the calculations. the authors found that the encompassing rectangle provides a bounding geometry for design of glass with trapezoidal and triangular surface areas.46 kPa (30. It presents the basis for the design methodology.75 for short duration loading and 0.62kPa = 1.32kPa = 1. the authors select a new trial thickness designation of 16 mm (5/8 in. Therefore. The aspect ratio is 1.8 psi). Conclusions This paper describes the development of a design methodology for large trapezoidal window glass lites.4 psi). which is 1. The aspect ratio is 1.45 for long duration loading.45x 1.90 for short duration loading and 0. (4) (3) The short duration load resistance is greater than the combination of the wind load and the weight component acting normal to the glass. Hence the required glass thickness designation is 16 mm (5/8 in.53N'Tm12(1.86 x 10. even though ASTM E 1300-00 tends to predict higher window glass load resistance.001 is 1.0 and the b/t ratio is 134.729 kPa (15.NORVILLE ET AL. the glass type factors (GTF) are 0.8 psi).122 kPa (2. a trial thickness designation of 12 mm (1/2 in. Therefore. 12lb'7).2 psi). wind loads do not fracture window glass during design . They observe that.) will not have adequate load resistance.04 kPa (21. m = 7 and k = 2.).62 kPa (33.45 for long duration loading. It also presents a design example.45x 2.32kPa = 1. Going through the computations for this thickness designation: Short Duration Load Resistance = 0.75x2.0 and the b/t ratio is 159. in ASTM E 1300-00 tend to predict weathered window glass load resistance higher than that indicated by data reported in the literature. which will function along the lines of the methodology used in ASTM E 1300-00. In developing the design methodology. although the authors do not wish to generalize to other geometries at this time. This leads to a short duration load resistance of: Short Duration Load Resistance = 0. ON TRAPEZOIDAL GLASS LITES 87 of glass weight acting normal to the glass is 0. The component of glass weight acting normal to the glass is 0.74 kPa (36.3 psi) and a long duration load resistance of: Long Duration Load Resistance = 0.55 psi).29 in.37 x 10. The approximate maximum glass deflection will be approximately 8 mm (0.62kPa = 0. The authors also note that the Weibull parameters. GTF are 0. Consequently.9x1. The non-factored load associated with a probability of breakage of 0.103 kPa (2.3 in.12 kPa (23.4 psi) and Long Duration Load Resistance = 0.15 psi).

"Comparison of Rectangular Glass Plates with Area-Equivalent Trapezoidal Plates. 1135-1144. and Minor. R. and El-Shami. [3] EI-Shami. pp. M. in review." Master of Science Thesis. Houston... W. 1985.. Texas Section Spring Meeting. and Norville. Texas.. and Norville. Ed. M. H. "Development of a New Glass Thickness Selection Procedure. J. [2} EI-Shami.. 1997 [4] Vallabhan." Bulletin of the American Ceramic Society.. M. 1989. October. 1990.. L. 64(11). H." Journal of Structural Engineering... V. "Load Resistance of Trapezoidal Window Glass. "IX. Submitted in Partial Fulfillment of Degree Requirements." John Wiley & Sons. M. American Society of Civil Engineers. 36. M. West Conshohocken. R." Journal of Engineering Mechanics. V. Texas. Block.. "The Strength of Weathered Window Glass. 1984. "A New Model for Analyzing Architectural Laminated Glass. W." Journal of Wind Engineering And Industrial Aerodynamics. American Society of Civil Engineers." Proceedings. March. G. D. Acknowledgements The authors gratefully acknowledge the U. S. L. and Norville. E. New York. "Concepts and Applications of Finite Element Analysis. American Society for Testing and Materials. [7] Norville. Malkus. 110(2). V. M. H. and Morgan. 197212. They offer opinions as to why design wind events putting loads on inservice window glass do not cause the predicted failure rate of 8 lites per thousand... S. G. R. April. [8] Beason..S. S. [5] Cook. pp. Lubbock. C. S. and Plesha. Department of Civil Engineering. and Kandil. M.. M. Texas Section Spring Meeting. American Society of Civil Engineers. FAA for providing monetary support for this work. S. ASTM STP 1434. American Society of Civil Engineers." Proceedings. M. "Development of Design Methodology for Rectangular Glass Supported on Three Sides to Resist Lateral Uniformly Distributed Loads. E. in review.. H. . 2001. Elsevier Science Publishers. December. A. 1467-1470. San Antonio. J. D. PA. "A Comparison of Nonlinear von Karmen and Mindlin Plate Solutions." The Use of Glass in Buildings. S. Vallabhan. [6] Beason.. pp. [9] Jackson. 2001. C.. "Glass Failure Prediction Model. Texas Tech University.88 THE USE OF GLASS IN BUILDINGS wind events. References [1] E1-Shami.

M. Sydney. G. 1984." Proceedings. .."An Investigationof the Bending Strength of Glass Louvre Blades. Australia." 9thAustralian Conference on the Mechanics of Structures and Materials. and Muir. ON TRAPEZOIDAL GLASS LITES 89 [10] Walker. August. L.NORVILLE ET AL. R.

3417 73rd Street." The Use of Glass in Buildings. S.Stephen M.. ASTMSTP . TX 79423. "Window Glass Design Software. M. Suite K-3. Ed.. Standards Design Group. For instance.. computer technology.. The paper describes recent rapid improvement in computing technology and the corresponding increase in complexity of engineering design standards. West Conshohocken. The paper goes on to discuss specifically modifications that SDG incorporated into its window glass design software in conjunction with the forthcoming revision of E 1300. Abstract: Standards Design Group. Each advance in technology has led to increased productivity for engineers and designers. Morse l Window Glass Design Software Reference: Morse.astm. 90 Copyright9 by ASTM International www. It addresses those changes that have made possible software that reduces the computations in ASTM Practice for Determining Load Resistance of Glass in Buildings (E 1300) to a few simple input operations. Block. (SDG) produces software to aid engineering and design professionals in the building industry with lengthy and often tedious design procedures. PA. This paper discusses the impact of technological advances on window glass design methodology. design standards. It provides justification for development of software that incorporates long computation procedures with logical decisions to achieve designs with just a few simple input operations on the part of the designer and engineer. Inc. V. computer software. ASTM International. engineers and designers have employed many tools to assist them in performing calculations. It presents design examples to illustrate some of the differences that will result with this revision. 2002. Keywords: glazing design. ASTM El300 Introduction Over the years advances in technology have significantly altered the role of engineers and designers. Inc. Lubbock. 1president.

and laptop computers became more prominent. Electroniccomputers in those days had were huge machines thatfilledentirerooms. Computer sizesbecame smaller. Electroniccomputers also had the ability. The computers produced output in printedform.and more programming languages became available. ithas removed them from the onerous tasksof performing stringsof simple calculationsto accomplish analyses and designs. In the late 1960s and early 1970s.MORSE ON WINDOW GLASS DESIGN 91 A Brief History of Engineering Computing Prior to about 30 years ago. engineersand designersused sliderulesto perform basic mathematical calculations. What has allof thisdone to computational abilities engineers and designers? of Simply put. While toolsto perform computations progressed from sliderulesand mechanical adding machines to programmable calculators. The earlydesktop versionsused a disk operatingsystem thatwas not particularly make numerous decisions. desktop computers became compact and reasonablypriced. In the modem world. Programmable calculatorsprovided engineersand designerswith the ability to perform chains of computations and make simple decisions.with the resultthatone of thesemachines occupied nearly every designer'sor engineer'sdesk. includingmultiplication. engineers and designers have several options. exponentiation. But by the middle to late 1990s. In the late 1960s. In the middle to late 1970s. electroniccomputer was becoming the more accessibleand easierto use. The sliderule leftplentyof room for errorand provided no assistance whatsoever in interpretingdesign procedures. they provided the engineer and designerwith the ability accomplish more than simple to calculations. These machines had computing power thatexceeded thatof the huge machines from the preceding decade. programmers could only communicate with thesehuge devices through machine language. These proved much faster.and on the basisof these.and considerablymore versatilethan sliderulesin performing calculations. programming languages such as BASIC and F O R T R A N appeared and provided an interfacethatengineers and designerscould use in place of machine language.and trigonometric calculations. The slideruleproved quicker and more versatilethan eithertables combined with pencil and paper or mechanical adding machines for performing calculations.more accurate. Most significantly they can write programs that embody numerous calculations with decision trees that allow the programs to accomplish complicated analysis and designs. Sliderulessimply helped engineers and designers perform calculations. within the scope ofprograrns. data input changed from card decks to electronic communications through keyboards and monitors. enabling them to become more efficient. methods of communicating with these machines alterthe calculationpaths. In the 1980s. They used card decks to provide input instructionsand data. computers enabled engineersand designersto perform huge seriesof calculationsrapidly.thereby increasingthe speed of input and output. As desktop. Floatingpoint and scientific calculatorsmade theiradvent in the early 1970s.As programmable calculatorsappeared in the mid-1970s. Engineers and designerscould view output on monitors ratherthan waiting for printedoutput. Despite being cumbersome to use. . This frees the engineer and designer to devote their time to other considerations. desktop computers began to gain prominence. Initially. largecorporations and universities access to electroniccomputers.

at least. Beason and Morgan [1] published a paper on failure prediction for window glass. software that performed the calculations in E 1300 1998 and 2000. Reliability methods replaced straightforward working stress approaches in concrete and steel analysis and design. look up factors. and produce designs increases. including stress duration. In 1998. In other words. relative humidity. E 1300 grew into 12 charts. Norville and Minor [2] published a paper that used failure prediction methodologyto assess the strength of weathered window glass. The beauty of the SDG software lies in the fact . In wind load computations. and window glass age proved to have measurable effects on window glass strength. The failure prediction methodology required significant computational power.92 THE USE OF GLASS IN BUILDINGS A General Overview of Standard Development At the same time as computer hardware and software underwent these significant advances. temperature. Improved computing technology allowed the development of more sophisticated models of physical phenomenon. window glass geometry. Inc. As models of physical phenomena gained in complexity. a good standard is simply an algorithm for the mathematical calculations necessary to produce a design. too. Other parameters. The 2002 edition will contain approximately 84 charts.. but duration shortened as improving technology facilitated improved methods to measure wind speeds and their resultant pressure on buildings. they need not go repeatedly through all the machinations entailed in these design standards. The author feels that a good design standard should produce the same results for the same input conditions and professional judgments of the designer. 6 tables. the time needed to read charts. 5 tables embodying many faotors. The failure prediction methodology replaced maximum stress theories and aforementioned simpler methods of assessing window glass strength. for its time. As complexity increases. The research underlying both of the papers relied heavily on advancing computer technology in experimental research as well as analysis techniques. standard design methodologies followed a parallel path. The author also feels that while engineers and designers must know and understand the design standards they use. load magnitude became higher. even when the user is familiar with the standard. In addition. The impact of computers on research efforts revolved around the ease with which computers allowed researchers to gather and record experimental data. SDG's first venture consisted of Comprehensive Window Glass Design. (SDG) set about formulating software that performs the computations in various standards for the design industry. and 15 procedures to accomplish window glass design. to provide assessments of window glass strength. The failure prediction methodologyincorporated all factors known to affect window glass strength. E 1300 fits into this definition. window glass design required 12 charts and a table of factors. Research on window glass strength and behavior indicated that strength models relating failure load to window glass thickness failed to adequately describe window glass strength. and special procedures for different window glass types and constructions. Standards Design Group. Consequently. state of stress. Window glass design in the 1970s and early 1980s required one glass strength chart and a table of factors. Furthermore. design practice for window glass began changing. the time required for the professional's learning curve increases. With the advent of the first edition of E 1300.

2 and 3-sided supported single glazed lite designs. Changes in the 2002 Edition ofASTM E 1300 The revision for the 2002 edition includes the addition of 1. one for each single glazed lite support and construction combination replacing the complex polynomial method. The subsequent double glazed insulated unit combinations are represented in the same format.Monolithic Monolithic ..Laminated Laminated.Laminated X X X ' Monolithic . Additionally 42 deflection charts have been added. and laminated over laminated double glazed insulating units have been added. . These changes are Incorporated in over 42 charts..MORSE ON WINDOW GLASS DESIGN 93 that it can keep pace w/th changes in standard design practices such as E 1300 and incorporate all complications that accrue as standards become more complex in nature.Laminated E 1300-02 Design Methods 1-sided 2-sided Single Glazed Lite Monolithic Laminated (Symmetric) Laminated (Asymmetric) 3-sided x x x 4-Sided x x x x x x x x x Double Glazed Insulating Unit Monolithic . design methods for asymmetric laminated IRes. Table I - Summary ofchangesbetweenE 1300-00and E 1300-02 E 1300-00 Design Methods l-sided 2-sided . Table I summarized the additions in the 2002 edition.Monolithic represents an double glazed insulating glass unit constructed with two monolithic lites.. Also.Monolithic 1 Monolithic . 3-sided 4-Sided x x Single Glazed Lite Monolithic Laminated (Symmetric) Double Glazed Insulating Unit Monolithic .

Multiply NFL by GTF to get the LR of the lite. Calculate the aspect ratio (AR) of the lite by dividing the long side length (a) by the short side length Co): AR = a/b. Determine the NFL from the appropriate chart for the glass thickness and size. Determine the non-factored load (NFL) from the appropriate chart for the glass thickness and size. Multiply NFL by GT to get the LR of the lite. Determine the glass type factor (GTF) for the appropriate glass type and load duration. Determine the NFL from the appropriate chart for the glass thickness and size.00 For Monolithic Single Glazing: 1. For Single-glazed Laminated Glass Constructed with a PVB Interlayer Simply Supported Continuously along Four Sides where In-Service LG Temperatures do not Exceed 50 oC (122 ~ 1. 2. Design Procedure for E 1300 . and t = plate thickness (ram) or (in. 4. 2. 3.94 THE USE OF GLASS IN BUILDINGS Design Procedure for E 1300 . Multiply NFL by GTF to get the LR of the lite. Determine the glass type (GT) factor for the appropriate glass type and load duration.02 For Monolithic Single Glazing Simply Supported Continuously Along Four Sides: 1. Procedure for Calculating the Approximate Center of Glass Deflection The maximum glass deflection as a function of plate geometry and load may be calculated from the following polynomial equations. Determine the NTL from the appropriate chart for the glass thickness and size. Calculate the flexibility ratio (b/t) by dividing the short side length (b) by the laminated glass thickness designation (t). 3. Determine the GTF for the appropriate glass type and load duration. 3.). MultiplyNFL by GT to get the load resistance (LR) of the life. Determine the GT factor for the appropriate glass type and load duration. For Single-glazed Laminated Glass (LG) Construction of Two Glass Plies of Equal Thickness and Glass Type Bonded Together with PVB Interlayer: 1. 2. 5. 3. w=t*exp(ro +r1 * x-I. 2.r2 *X 2) where w = center of glass deflection (ram) or (in. (1) .).

4 x 106 psi).0 mm (1/4 in. and E 1300-00 is b a s h on a 60-second load.908 * (a / b) + 0. Example 1: Monolithic Window Glass Design In this example. the author projects a horizontal line from the 1600 mm (63.).t4]} (2) (3) (4) (5) q = uniform lateralload (kPa) or (psi). E 1300-00 contains 12 non-factored load charts.).5 = 1.1. The first will be a monolithic single glazed lite supported on all four sides.2.0822 * (a / b) 3 x = ln{ln[q*(a/b) 2 ~E. b = short dimension (ram) or (in.83 * (a /b) . the glass type factor equals 1.5 kPa (31 psi) (6) Since the design load is a 3-second duration load. the author wishes to determine the appropriate glass thickness required to glaze a vertical opening having rectangular dimensions of 1600 mm (63 in.7 x 106 kPa) or 00. For an AN monolithic lite under short duration load.0 mm (1/4 in.) by 1600 mm (63 in.83 * (a / b) + 1.0 in.0969 * (a / b) 3 r 1 = -2.Following the steps outlined in the procedure section.). the basic NFL is 1.50 k/'a (31 psf). a = long dimension (ram) or (in. Using E 1300-00 .3.) point on the horizontal axis.0 mm (1/4 in.0 in.) AN glass is computed by the product: LR = GT x NFL = 1. first using E 1300-00 and then E 1300-02.485 .) point on the vertical axis and a vertical line from the 1600 mm (63.MORSE ON WINDOW GLASS DESIGN 95 ro = 0. Design Examples The author will illustrate some of the differences between E 1300-00 and E 1300-02 through two examples. for a monolithic single glazed lite design. These two lines intersect on the 1. one for each nominal lite thickness.815 * (a / b)2 _ 0.) to resist a 3-second duration design loading of 2. the design load must be converted from a 3-second duration to a 60-second duration . The second will be a laminated single glazed life supported on all four sides. Additionally.553 . therefore.2067 * (a/b) 3 r 2 = 1. To determine the load resistance for this lite.11 * (a / b) 2 _ 0. The LR for the 6. Figure 1 presents the non-factored load chart for a nominal thickness of 6.00.2 kPa (45 psi).). The author starts the procedure by trying a monolithic annealed window glass lite having a nominal thickness designation of 6. after the appropriate glass thickness has been determined the author will also calculate the approximate center of glass deflection under that loading. E = modulus of elasticityof glass (71.29 + 5.0 x 1.17 * (a /b)2 + 0.50 kPa (31 psf) load line. For each case the design will be performed.

kPa*(7. which is greater than the equivalent 60-second duration design load. The author finds in E 1300 the load duration factor for a 3-second load to a 60. the AN glass thickness is insufficient.) [ire is determined to be 2. ro = -2. The load resistance for the 8.577.OOB I00 150 200 La oO O oo ~ 3000 E I kPe = 2 0 .mm)')]} = 1.Non-factored Load Chart for Nominal 6-ram (1/4 in.2.0 kPa (42 psf). To increase the LR.8 kPa (38 psf) (7) Since LR is less than the specified design loading.43 + 0.42.) lite will work.42. The equivalent 60-second duration design load is the 3-second duration load divided by the load duration factor for a 3-second load: Q6o = 2.43 w = 7.second load is 1.31 * 1. therefore.31 x = ln{ln[(2.7"10 ~ .kPa* ( 2 . which is 8. mm * exp(-2. g p s f C 100 C E t" 2. 5( 2000 .oo r. 50 a.21 for annealed glass.0 mm (5/16 in.54 in.7 nun (0.13.284.) Glass (E1300) in order to compare it to the load resistance. the 8.2 kPa / 1. The approximate center of glass deflection is calculated using the polynomial method as follows: a/b = 1600/1600 = 1. 5 6 .) (9) (10) . r2 = 0.).21 = 1.d rC i0 O0 0 1000 2000 3000 (ram) 4000 C 0 P l a t e Length 0 5000 Figure 1 . rl = 1.284 + 1.0 (8) Subsequently.mm2)~)/(71. 1 0 ' .0 m m (5/16 in.0 m (5/IB ~n) S m Nonfactored (Pro) PD = O.96 THE USE OF GLASS IN BUILDINGS Plate Length (inl 0 50 8.577 * 1.0 m m (5/16 in. the author will perform the same procedure using the next greater lite thickness.435) .

and the design is acceptable.5 kPa (52 psf) and the 3.--~k~ . / 1 N I r'-k I. Figure 2 presents the non-factored load chart for a nominal thickness of 8. z0o/l I~l A. the author starts the procedure by trying a monolithic annealed window glass lite having a nominal thickness designation of 8. one for each nominal lite thickness. Plate Length (in.0 mm (5/16 in. for a monolithic single glazed simply supported continuously along four sides design.) lite is computed by the product: LR = GT x NFL = 1.) 0 50 100 1 150 2O0 140 8.0 mm (5/16 in. I I~'III g l. For a AN monolithic lite under short duration load.~ ~ I~ 2~ 3000 == i 4~ ~P~T~/ 1000 0 0 5000 Plate Length (mm) Figure 2 .ooAdhL~. the LR for the 8.). Therefore. the GTF equals 1..0 in."~ 7.) Gla:ts Nonfactored Load 120 : Pb = 0. These two lines intersect between the 2.7 kPa (56 psf)..oo .'1 Ill 5~~.0 mm (5/16 in. the author will "eyeball" interpolate the non-factored load.0 mm (5116in.. .0 kPa (63 psi') lines.0 in. The LR for the 8. Since the intersection is between two load lines.) Glass (E1300-02) To determine the load resistance for this lite.9 psf 1.00.o0 ' o.7 = 2.Non-factored Load Chart for Nominal 8-ram (5/16'20oo L~! .Using the steps outlined in the procedure section. Therefore.) point on the horizontal axis. I ~ .~t 9 9 3000 9 oo 8o 60 40 20 0 3-Second Duration . the basic NFL is 2.).MORSE ON WINDOW GLASS DESIGN 97 Using E 1300-02 . E 1300-02 contains 12 non-factored load charts for monolithic 4-sided supported lite.0 x 2.7 kPa (56 psi') (ll) Since the design load is a 3-second duration load and E 1300-02 is also based on a 3second load one does not need to factor the design load.) lite is greater than the design load.~.0 mm (5/16 in.~ I~t'q I ~A"qkl ~-. E 1300-02 .) point on the vertical axis and a vertical line from the 1600 mm (63. the author projects a horizontal line from the 1600 nun (63.008 I kPa = 20.

. the approximate center of glass deflection is 14 mm (0.2 ..). _0. Figure 4 presents the main design input window. Once the designer enters in the desired parameters he or she simply clicks on the calculate button.6 I _~.. the author projects a vertical line from the point on the horizontal axis corresponding to the value of Load x Area 2 in kN-m2 (kip-ft').Deflection Chart for Nominal 8-mm (5/16 in. / 0~ 0 ~ 80 I Oe. 40 35 1. Using Window Glass Design 2002 ..2 I Ioo 200 Load x A r e a 2 (kN" m 2) Figure 3 . I .7 I/ / " t ... the load resistance and the approximate center of glass deflection.jr.. .. .) Glass (E1300-02) One can determine the approximate center of glass deflection using the deflection chart in a similar fashion to the load charts.... Figure 5 shows the results window returned by the software which includes the design load. ~ ~11 .4 30 E" _i .2 kPa x (1600 mm x 1600 ram)2 = 14...55 in..v.. Figure 3 presents the deflection chart for nominal thickness of 8. rectangular dimensions..~ .0 mm (5/16 in..42 kN'm 2 (34.98 THE USE OF GLASS IN BUILDINGS provides a new set of deflection charts for each lite thickness. i .4 .eo../1 i g t- 25 2O 9 s ~ 15 10 84 5 ...)... f / " ii I ~ 1.. .. and loading conditions. ~ 9 / ~J ..88 kip'ft2) (12) The deflection is the point on the vertical axis where a horizontal line projected from the intersection of the aspect ratio and the vertical line meets the vertical axis.o. Load x Area 2 = 2.' S ~ / I . ..Ooa.-.. I 120 160 Four Sides Simply Supported ~io. where the designer defines the glass construction.~ i~.The following figures illustrate the so.k i108 o. Load x Area 2 (kip" ft 2) 0 100 200 300 400 500 . Therefore..~ I ~ / ~ -'t40 l~" / _ r . To determine the deflection for this lite.ware interface.

MORSE ON WINDOW GLASS DESIGN 99 Figure 4 .Window Glass Design lnput Window (Window Glass Design 2002) Figure 5 . .) by 1930 mm (76 in. The sketch window shows a dimensioned drawing of the glass construction and the details window shows specific factors used in calculating the load resistance. Example 2: Laminated Window Glass Design In this example.4 kPa (50 psf). The software also provides a detailed two page report that summarizes all of the information contained on the four design windows. the author wishes to determine the appropriate glass thickness required to glaze a vertical opening having rectangular dimensions of 965 nun (38 in.Results Window (Window Glass Design 2002) Two additional windows are available to the designer.) to resist a 3-second duration design loading of 2.

I I.) AN glass is computed by the product: LR -. o o / I II ~L L ~ I I~t"1 P ~ I IJ2.GT x NFL = 0.0 in.) point on the vertical axis and a vertical line from the 1930 mm (76.100 THE USE OF GLASS IN BUILDINGS Additionally./rl I P I. Using E 1300-00 . "0 Z al 100 I I g 4 " l I I I .5(/~] k l I "F. therefore.) point on the horizontal axis.-4-I~1 J J r ~ ~ r ' l ~ L .OOB r /111111 III 3000 r. Plate Lengt}~ (in) 5O i i i r i l # J i 100 ~50 200 B.50 kPa (52 psf).-. the basic NFL is 2. for a single-glazed laminated glass construction of two glass plies of equal thickness and glass type bonded together with PVB interlayer.).5o/]%jll ~ I%L. the glass type factor equals 0.~! I-Jr~T I 1000 . the author starts the procedure by trying a glass lite having a nominal thickness designation of 8.0 mm (5/16 in.~11 r'. The author finds the load duration factor for a .9 kPa (40 ps0 (13) Since the design load is a 3-second duration load. For a AN laminated lite under short duration load..) Glass (El300) To determine the load resistance for this lite.Following the steps outlined in the procedure section.75..O mm (5/15 in) Crass i .~V1 I I I./] I /'11~ I I I l.Non-factored Load Chart for Nominal 8-ram (5/16 in.I I I J I I 1 I I I I I I I I o 1000 2000 3000 4000 5000 Plate Length (mm} . L .0 in. the author projects a horizontal line from the 965 mm (38. The LR for the 8. after determining the appropriate glass thickness. and E 1300-00 is based on a 60 second load.~ l I 3-oo~I~1 ~YLI~b.5 kPa (52 psf) load line.0 mm (5/16 in. ! ! ! ! ! ! !l! Figure 6 ..75. 1 kPa = 20. Figure 6 presents the non-factored load chart for nominal thickness of 8. _l .. the author will also calculate the approximate center of glass deflection under that loading..0 mm (5/16 in. .~V1 I 1..424 ~ _.~'1 I I . 25/] I P I J -/1 T I I I.g Dsf /q I I I I I I L~I o.. These two lines intersect on the 2.~ I 2 .~I~L-~L~T--~. one must convert the design from a 3-second duration to a 60-second in order to compare it to the load resistance. 0 I [ I IIIII Ii iJl~r-I I I I In.75 x 2. 2000 /1%Fl-d J/Itl I I.~11 I I I I I J..).5 kPa = 1. -'/]' '!'" :J )J"')'i Nonfmctonea Loar PI~ 9 m o .

). the AN glass thickness is insufficient. 2.0 mm (5/16 in. . Therefore.008 4 .0 /--IN. ~" E ~" ~0oo .MORSEON WINDOWGLASSDESIGN 101 3-second load to a 60-second load is 1.0 mm (5/16 in.0(]. / / ~ .18 in..~ i . Figure 7 presents the nonfactored load chart for laminated glass with a nominal thickness of 8.--q--l-I- 1ooo o o 0 ' 1000 2000 3000 Plate Length(ram) 4000 5000 Figure 7 .4 kPa / 1.) 1Re will with stand the design load. ~ T .). which will be 10.) Laminated Glass (E1300-02) .21 = 2. 5 0 . is 4.0 mm (3/8 in.Following the steps outlined in the procedure section. the author starts the procedure by trying a glass 1Re having a nominal thickness designation of 8.). for singleglazed laminated glass constructed with a PVB interlayer simply supported continuously along four sides where in-service LG temperatures do not exceed 50* C (122 ~ F).). ' . Pb = 0.'g_ 8mm(5116in.'.0 kPa (42 psf) (14) Since LR is less than the specified design loading. the 10.0 mm (3/8 in.ond duration load divided by the load duration factor for a 3-second load: Q6o = 2.21 for annealed glass.~ m n 5O 1 io . which is greater than the equivalent 60-second duration design load.Non-factored Load Chart for Nominal 8-ram (5/16 in.6 mm (0.. 150 0 50 Plate Length(in) 100 150 i 200 i ' .~= / / 1 kPa = 20. .5 kPa (73 psf). ..100 3 SecondDuration . The approximate center of glass deflection. one for each nominal 1Re thickness.) lite is 3.0 mm (3/8 in. The equivalent 60-second duration design load is the 3-s~. ~ 50oc (122oF) . To increase the LR the author will perform the same procedure using the next greater IRe thickness. Using E 1300-02 . calculated using the polynomial method..)PVBLiminate NonfactoredLoad ~ FourSidesSimplySul:ported 1.9 psf e. The load resistance for the 10. E 1300-02 contains 12 non-factored load charts for laminated 4sided supported IRes.

8 mm (5/16 in.0 0.2 kPa (67 psf)./ / 400 1.0 kPa (63 psf) and the 4.) Laminated Glass (E1300-02) One can determine the approximate center of glass deflection using the deflection chart in a similar fashion to the load charts. the GTF equals 1.102 THE USE OF GLASS IN BUILDINGS To determine the load resistance for this late.) late is greater than the design load so the design is acceptable.2 kPa (65 psf) (15) Since the design load is a 3-second duration load. .0 mm (5/16 in. Therefore.0 mm (5/16 in. one does not need to factor the design load. the author projects a vertical line from the point on horizontal axis corresponding to the value of Load x Area 2 in kN-m 2 (kip-t~). 0.4 1._= v E 20 e- 0. For a AN laminated late under short duration load.8 30 E r- oAT~ .) late is computed by the product: LR = GT x NFL = 1.0 180 Load x Area =(kN.).Deflection Chartfor Nominal 8-ram (5/16 in. 300 .0 in.) point on the horizontal axis. These two lines intersect between the 3.6 10. the LR for the 8. the basic NFL is 3.0 x 3. The deflection is the point on the vertical axis where . (Load x Area= 0.0 mm (5/16 in. Load x Area2(kip.ft 2) 0 40 .0 in.r / / / f . Therefore.4 Four Sides Simply Supported Defllmtlon vs..2 1.0 kPa (84 psf) lines.2 500C (122"F) -" 0 1 0 30 60 90 I 120 [ 1 150 I 0.00. the author projects a horizontal line from the 965 mm (38. E 1300-02 provides a new set of deflection charts for each late thickness.) point on the vertical axis and a vertical line from the 1930 mm (76. The LR for the 8. and E 1300-02 is also based on a 3second load. Since the intersection is between two load lines the author will "eyeball" interpolate the non-factored load. Figure 8 presents the deflection chart for laminated glass with a nominal thickness of 8. 100 i 200 . To determine the deflection for this late.) PVB Laminate --.m =) Figure 8 .2 = 3.

Again the designer simply enters in the desired parameters and clicks the calculate button.MORSE ON WINDOW GLASS DESIGN 103 horizontal line projected from the intersection of the aspect ratio and the vertical line meet the vertical axis.).5 mm (0.Window Glass Design Input Window (Window Glass Design 2002) Figure 10 .The following figures illustrate the software interface.37 in. the approximate center of glass deflection is 9. Figure 9 . Using Window Glass Design 2002 . Therefore. Figure 9 presents the main design input window and Figure 10 shows the results window returned by the software.Results Window (Window Glass Design 2002) .

On the other hand. standards either will include large amounts of complex calculations or include many charts and tables which summarize the complex calculations. although faster to use.the designs are more consistentand accurate. Since the program willperform allthe calculationsvery quickly. Bulletion of the American Ceramic Society..E. the calculations comprising them increase in complexity. Vol. "Glass Failure Prediction Model. Both methods have pros and cons. the lastedversion of SDG's window glassdesign software includes the numerical data that forms the basis for the non-factoredload chartsas well as allcalculationsused in E 1300. when used correctly. In general. 64 No. Depending on the design. "The Strength of Weathered Window Glass. Since the software does the actualcalculationsand includes the underlying data from the charts.alleviatingthe differencesdue to "eyeball"interpolationof charts. For the standards that employ complex sets of calculations. A benefit of using this type of standard is that they will provide. 2. and Minor. Window Glass Design 2002. L. As with E 1300 the non-factored load charts are easy to use.the time spent on calculationsby engineers and designerswill be greatlyreduced. R. pp 1467-1470.104 THE USE OF GLASS IN BUILDINGS Conclusion As standards continue to change and evolve to become more accurate.. . Vol 110. but the values of the loads may vary by more than + 5% due to the way the designer "eyeball" interpolates." Journal of Structural Engineering.. 11. and Morgan. provide opportunity for variation in results due to individual interpretation. the most accurate designs that the current research and technology support. the designer gets the best of both situations. The software performs the calculations of the complex design procedures. This is where the advantage of software comes in. S. J. An additionaladvantage in using soRware for design calculationsis the reduced learningcurve for designers and engineers due to significantrevisionsin the standards. November 1984. J. References [1] Beason. Through software. standards that consist of many charts to be "eyeball" interpolated. February 1984." Bul. while only requiring minimal input from the user. this may be the difference between using a smaller economical lite and one that is larger and more costly. H. W. This software produces consistentresultswhere everyone obtainsthe same designs for the same input. many designers and engineers may have difficulty becoming proficient with the design procedures which are very time consuming. [2] Norvillo. No.

if part of the glass is shielded from the direct effects of the sunlight by the edge support system or by a shadow pattern. Thermal breakage involves the occurrence of thermal gradients induced by uneven heating of glass by solar irradiance or other heat sources. windows. Keywords: glass.Rockwall. V. This uneven heating situation causes tensile stress to develop along the edges of the glass. L. ' AssociateProfessor. an edge strength failure prediction model for annealed glass is presented which allows a designer to determine if the level of risk associated with an estimated edge stress is acceptable or unacceptable. some of the energy is absorbed by the glass.astm. Lynn Beason t and A. 2002. When sunlight impinges upon a glass plate." Use of Glass in Buildings. and some of the energy is transmitted through the glass. West Conshohocken. Ed. "A Thermal Stress Evaluation Procedure for Monolithic Annealed Glass.W.TX 77843. The energy that is absorbed by the glass increases the temperature of the glass above the previously existing equilibrium condition. In addition.. and if the support system can accommodate the thermally induced expansion of the glass. no stresses will be induced by a uniform temperature increase. 2Principal. W. This procedure is based on the results of a finite element parametric study and engineering judgment.. thermal stress. William Lingnell ~ A Thermal Stress Evaluation Proeedure for Monolithic Annealed Glass Reference: Beason.LingnellConsultingServices. The uneven heating of the glass will give rise to thermally induced in-plane tensile and compressive . design Introduction One of the most common factors that leads to the breakage of window glass in buildings is thermal breakage.TexasA&M University. and Lingnell. the glass will be heated unevenly. This paper presents a theoretically based procedure that can be used to estimate the level of tensile stress induced in a glass plate exposed to this typical thermal loading situation.CollegeStation. When these thermally induced tensile stresses interact with critical edge flaws.. 105 Copyright9 by ASTMInternational www. The classic thermal breakage situation occurs when the edges of a glass plate are shielded by the window frame system or significant building protrusions. ASTM International. 1270ShoresCourt. If the glass is uniformly heated.CivilEngineeringDepartment. Abstract: Thermal breakage caused by the uneven heating of glass by solar irradiance is one of the most common causes of annealed glass breakage in buildings. Block. while the central portions of the glass are not. A. W.TX 75087. some of the energy is reflected from the surface of the glass. However. PA. ASTM STP 1434. thermal breakage can result.

A generalized theory to explain the breakage of glass subjected to thermal stresses must first incorporate a procedure to calculate the magnitude of the maximum tensile stress along the edges of the glass. This is accomplished through the use of finite element analysis (FEA). the edges of the glass are subject to higher tensile stresses than the surfaces of the glass away from the edges. This latter objective is accomplished through the use of an edge strength failure prediction model (ESFPM). In the current discussion it is assumed that the glass plate is installed in a typical frame that provides four sides of continuous lateral support with an edge bite (glass engagement into frame system) that protects a perimeter strip of glass from direct exposure to the sunlight. exterior shading conditions. there is very little information available to explicitly describe the development of thermal stresses in glass. and other factors. The purpose of this paper is to present a rational procedure that can be used to evaluate the thermal stress effects of sunlight on rectangular glass plates. Therefore. As a result. Secondly. the stress concentrating flaws along the edges of a glass plate are generally more severe than surface flaws away from the edges. Hence.106 THE USE OF GLASS IN BUILDINGS In a typical thermal breakage situation. These two items are combined to develop a procedure that allows a designer to determine the risk associated with a particular thermal stress situation. Thermal stress breakages tend to originate at the edges of glass plates as the result of the interaction of thermally induced tensile stresses with edge flaws or edge damage. One of the most commonly discussed examples of thermal stress breakage occurs on clear cold winter days. interior shading devices. In general. the temperature outside of the . During the cold night the glass cools to an equilibrium condition that depends on the temperature inside the building. If the glass is supported in some other manner. most thermal stress evaluation procedures are based more on empirical methods and experience than explicit engineering formulations. It is also assumed that the edges of the glass are free to slip within the support system so that normal thermal expansions and contractions may be accommodated. While not as common. In addition. the procedure must include a method to account for the occurrence and severity of edge flaws. the level of residual edge and surface compression inherent in either heat-strengthened or fully tempered glass tends to preclude thermal stress breakage problems. the magnitude of the edge bite. thermal stress breakage is associated with relatively low levels oftensile stress. Thermal stress breakage is exacerbated by the heat absorption characteristics of the glass. The thermal stress evaluation procedure presented in this paper focuses only on shading caused by the edge bite. Although the generalities of the solar induced thermal stress problem have long been understood by those working in the glass industry. the majority of thermal stress problems are experienced by annealed glass that has a low level of residual edge and surface compression. These circumstances combine to create a situation where thermally induced glass breakage usually initiates at the edges of glass plates. Solar Induced Thermal Stresses in Glass Plates In-plane or membrane stresses are introduced in glass by uneven heating of the glass plate. severe surface damage or flaws can also trigger thermal stress fractures. the character of the stress distribution may be significantly different than that discussed herein.

it expands. the interior surface film coefficient.BEA$ON AND LINGNELL ON MONOLITHIC ANNEALED GLASS 107 building. breakage of the glass at the edge of a thermally loaded glass plate usually results in a crack that is normal to both the vertical and horizontal projections of the edge of the glass as shown in Figure 2. Because of the orientation of the edge stress. As the exposed area of the glass warms. it becomes clear that the edge block must be in uniaxial tension or compression. both the top and bottom surfaces of the glass are exposed to the same stress. Because thermally induced stresses are membrane stresses. This is due to the fact that the stress components on the free surfaces of the block are all zero as shown. As the crack propagates inward.[ O~ Figure 1 --Thermally induced edge stress. the unheated edges of the glass around the perimeter are forced into tension. When sunlight begins to shine on the glass. Ifa small block of thermally loaded glass is examined along the edges as shown in Figure 1. As the exposed area of the glass expands. o /. the exposed area of the glass is heated. the stress distribution in the glass plate becomes more complicated and the crack may branch into different directions depending upon the transient state of stress within the glass as it fractures. and the exterior surface film coefficient. .

and the nature of the support conditions. and the elevation angle of the sun in winter is more favorable for loading vertical glass plates than it is in the summer. the initial temperature of the glass has little to do with either the occurrence or severity of the thermal stress. The first step is the development of the FEA model that consists of nodes and elements that represent the geometry of the glass plate being studied. the convection/radiation conditions. The most probable reason that thermal breakage tends to be associated with winter conditions is that solar irradiance in the northern hemisphere is at a maximum during winter months. In the third step. Although there are several different FEA codes avail- . Then. the temperature data corresponding to a given point in time is transferred from the thermal analysis to a stress analysis.108 THE USE OF GLASS IN BUILDINGS Figure 2 --Thermally induced crack. Finite Element Analysis of Thermal Stresses FEA provides a convenient mechanism for standardizing a procedure that allows a quantitative analysis of thermal stresses induced in glass by uneven heating caused by solar irradiance. Rather. The initial temperature of the glass has little effect as long as the glass is in a steady state condition when it is first subjected to the sunlight. The stress analysis allows the stresses associated with the particular temperature distribution to be calculated. While the above scenario is the most frequently quoted example of the occurrence of thermal stresses. the severity of the thermal stress associated with a particular situation is primarily dependent on the amount of energy absorbed. the model is used to perform a transient or steady-state analysis of the heat flow in the glass. thermal stresses can become critical at any time when the proper conditions exist. The result of a thermal analysis is the variation of temperatures throughout the body of the glass at discrete points in time. While winter provides excellent conditions for inducing thermal loads in glass. In most FEA codes the general thermal stress analysis procedure is conducted in three major steps.

However. The writers have made estimates of the interior and exterior surface film coefficients for glass based on information presented by ASHRAE and engineering judgment [2]. the surface film coefficients employed in the analyses discussed herein. Then. it is believed that FEA analyses conducted with these assumed coefficients provide conservative results for the glass design situations discussed herein. Thermal analyses can be conducted for both steady-state and transient conditions.55 W/(m2-K) for the exterior surface and 8.BEASON AND LINGNELL ON MONOLITHIC ANNEALED GLASS 109 able for use to analyzethermal stresses in glass. thermal conductivity. Thus. the surface film coefficient has a significant effect on the rates of temperature change and the magnitudes of the temperatures that develop throughout a glass plate that is subjected to given set of conditions. and density. the FEA data presented in this paper were generated using the ALGOR version 12 program [ 1]. This collection of nodes and elements is referred to herein as the finite element model. Therefore. A steady state analysis is used to determine the equilibrium conditions in a long-term heat flow problem with constant conditions. In the current effort. A transient analysis is used to determine the variation of the temperatures in glass as a function of time. Refrigerating and Air-Conditioning Engineers (ASHRAE). Under some conditions. The results of the transient thermal .04 W/(m2-K) for the interior surface. Thermal Analysis The first step in formulation of a thermal analysis is to define the properties and characteristics of the materials involved in the problem under study. The surface film coefficient is used to model the transfer of heat energy between a glass surface and its surrounding environment based on the temperature differential betwoen the two. These values are 13. In this process discrete nodes throughout the geometry of the plate are graphically defined in the context of a three dimensional orthogonal coordinate system. a steadystate analysis was conducted for each situation with the glass exposed to constant indoor and outdoor temperatures to establish an equilibrium condition with no solar irradiance. actual surface film coefficients can be greater or less than these assumed values. these nodes are used to create elements that define the extent of the glass plate under consideration. In this paper all units are expressed in SI units. Formulation of the FEA Model The first step in formulation ofa FEA model to represem a particular situation is to formulate the geometry of the problem. The surface film coefficients are expressed in terms ofunits of W/(m 2 . The resulting equilibrium temperatures were then used as initial conditions for a transient analysis of the glass plate exposed to solar irradiance. The material properties of interest include the specific heat.K). It is the magnitudes of the temperature differentials throughout the body of the glass plate that ultimately determine the magnitudes of the thermal stresses within the glass plate. Following the precedent established by American Society of Heating. represent the combined effects of radiation and convection [2]. it is necessary to define realistic surface film coefficients for exposed glass surfaces.

and the thermal coefficient of expansion for glass is taken to be 8. The thermal stress failure prediction model is patterned after the GFPM. expressed in GPa. of a brittle material can be expressed as follows: . Poisson's ratio. instead of surface stresses and surface flaws.82 x 10"6 m/m/K. Both the ESFPM and the GFPM are based on a statistical failure theory for brittle materials that was originally presented by Weibull [4]. In most instances. and the thermal coefficient of expansion. The linear static stress analysis allows the calculation of the distribution of stresses induced in the glass given the distribution of temperatures and the prescribed support conditions. The Weibull theory suggests that the probability of breakage. a . the Poisson's ratio for glass is taken to be 0. the modulus of elasticity for glass was taken to be 71. the resulting stress conditions will be substantially different. Thus. Edge Strength Failure Prediction Model Most current glass thickness selection charts presented in the United States are based on the glass failure prediction model (GFPM). The thermal stress model thus developed is referred to herein as the edge stress failure prediction model (ESFPM). By examining a wide range of times.110 THE USE OF GLASS IN BUILDINGS analysis provided the variation of temperature in the glass as a function of time for the given set of conditions. but is allowed to slip freely in the plane of the plate. these boundary conditions were used to generate the results presented in this paper. expressed in m/m/K. except that it relates the probability of failure to the distribution of edge stresses and the characteristics of the edge flaws. The GFPM recognizes that laterally loaded rectangular glass plates with four-edge support fail as the result of the interaction between surface tensile stresses and stress concentrating surface flaws [3]. the maximum tensile stresses and the critical flaws associated with thermal breakage are almost always located along the edges of the glass. g . This process is conducted for different times to determine the variation of the maximum edge tensile stress as a function of time. To calculate the thermal stresses for a particular situation. However. The required material properties of interest are the modulus of elasticity. the variation of temperatures through the glass as a function of time are used as input for a linear static stress analysis. If the edges of a glass plate are prevented from slipping in the plane of the plate. The magnitude and distribution of the thermal stresses induced in a glass plate is extremely sensitive to the type of edge support conditions.7 GPa. Stress Analysis Formulation of a stress analysis for glass requires a definition of its material properties and a specification of appropriate boundary conditions. Pb. E. In the current situation.22. the boundary conditions associated with the edge of a glass plate are such that the edge is prevented from motion perpendicular to the surface of the glass plate. the magnitude and time of occurrence of the maximum edge tensile stress can be estimated. Therefore. a modified failure prediction model must be used to model the thermal stress situation.

Using this value of m. es Le . in application of the ESFPM. Equation (3) is the form of the ESFPM that was used in development of the information presented in this paper. = 8.2 m (150 mm x 8) from the total perimeter.~. Thus. As indicated.5 m perimeter. the effective length of the perimeter of the glass exposed to thermal stress is found by subtracting 1. Thus.5 m be conservatively treated as though they have a 1. (p -1. Experience suggests that it requires a distance of approximately 150 mm from the comer for the edge stress to reach its maximum value. the integral is summed along all glass edges that are subjected to tensile stresses. B. . (x) is the maximum principal tensile stress along the edge of the glass.BEASON AND LINGNELL ON MONOLITHIC ANNEALED GLASS 111 Pb = 1-e-B (1) where B is a risk function that is evaluated by integrating the combined effects of flaw severity and tensile stresses experienced by the glass plate. It is recommended that the m edge flaw parameter be set to a value of 7 for glass design purposes based on expected coefficients of variation associated with glass failure strength data [3]. it is assumed that all of the perimeter except for the 150 mm zones around the corners are subjected to the same maximum principal edge stress. rr.93 x 10. N'7be used to model glass edges. it is recommended that a value of k. In the case of thermally loaded glass plates.2) is the effective length of the perimeter. and results of a large number of edge strength experiments. and the other factors are as previously defined. No. assumes that the risk o f failure in a thermal loading situation is controlled by the length of the edge subjected to tensile stress: B--k.54 m u. It is believed by the writers that these edge flaw parameters are reasonable for most thermal design situations associated with glass thicknesses up to and including 6 mm. It is recommended that plates with a perimeter length less than 1. and rr. This allows equation (2) to be rewritten as follows: m (3) where p is the length of the perimeter of the glass expressed in meters.x(X)]'• (2) where t~ is the duration of the thermal stress expressed in seconds. This means that the recommended minimum effective perimeter length is 0. The following expression for the edge strength risk function.~ x . the edge stress is zero at the comer and increases to a maximum value that remains relatively constant along the edge.3 m. To use the ESFPM for the design of thermally loaded glass it is necessary to have edge flaw characteristics that are representative of the glass edge conditions that are to be expected in a given installation. k. and m are the edge flaw characteristics.

. and the edge strength risk function..oo'~ 14 --'-t POB = 0. . it is the writers' opinion that 60-minutes is a reasonable duration to assume.001 POB = 0. tr~._._____~ 0 0 5 10 15 Pedmeter (m) 20 25 30 Figure 3 --Probabilityof Breakage (POB)Chart If the assumed edge flaw characteristics and load duration are substituted into equation (3) the following relationship results: B = 5..~. before the ESFPM can be used for thermal design it is necessary to establish a reasonable duration for thermal loadings.2)[trm~ ]7 (4) This equation can be solved for the maximum edge stress. . p . .36 x 10-53(p -1. as a function of the glass perimeter.. . a 60-minute load duration is used in the formulations presented herein. POB = 0. This judgment is based upon the writers' previous experience observing thermal loadings situations and FEA results.002 .0001 | 1. Therefore.112 THE USE OF GLASS IN BUILDINGS Finally. . . While no exhaustive effort has been made to establish a definitive duration for thermal loading. 0 0 ~ o... \ 6 ~ ' ----.-__.004 POB = 0.008 POB = 0.i... thermally induced edge stresses usually last more than 60 seconds. B. i. to yield the following: . While it is reasonable to use a 60-second duration loading for wind loading situations. 20 18 " ' t ~ ' ~ 16 0 .

it is necessary to develop a detailed FEA model that accurately represents the situation under consideration including a detailed description of the geometry and exposure conditions. an acceptable POB that is consistent with the project expectations must be determined. particularly those projects that involve large glass plates. and a set of edge conditions that are believed to represent many conventional applications.0001 (1 lite per 10 000 lites) to 0. . The most favorable edge conditions occur when the glass edge in the glazing pocket is insulated so that no heat flows through the glass boundary within the edge bite. results of three simplified FEA models are presented herein. To determine the maximum allowable edge stress for a particular situation simply enter the horizontal axis of Figure 3 with the glass perimeter and project upward to the desired POB and then project to the left to estimate the allowable edge stress. As shown in Figure 3. a large number of glass plates. it is not always economically feasible or practical to expend the resources necessary to accomplish this for each individual project. Then.5 MPa as the POB varies from 0. While developing a detailed FEA model to calculate the edge stress is the most desirable approach. Determining Thermal Stress The best method to calculate the edge stresses for a given situation is to use FEA techniques as described in the previous section to model the specific situation under consideration.BEASON AND LINGNELL ON MONOLITHIC ANNEALED GLASS 1 13 1 tYm~ =I1"87X 1052 (p-l.008 (8 lites per 1 000 lites) and the perimeter varies up to 30 m. First. a rational decision as to whether or not the glass meets expectations must be made. The TSEP is accomplished by performing three steps. These three models were selected to represent the most favorable edge conditions. it is necessary to estimate or calculate the level of edge stress induced in the glass subjected to a defined set of conditions. It is recommended that a detailed FEA approach be used for all significant projects. To do this. Therefore.2)B ]~ (5) Equation (5) was used to develop Figure 3. or unusual glazing and building conditions. The details of the TSEP are discussed below. This condition occurs when the glass is perfectly insulated in the edge bite area. Finally. the most unfavorable edge conditions. This means that the only mechanism for heat flow into or out of the glass in the edge bite region is through conduction within the glass. These results allow a reasonable estimate of the magnitude of the edge stress to be made without performing a detailed FEA for each specific problem. the allowable edge stress varies from about 5 to 17. Thermal Stress Evaluation Procedure The purpose of this section is to present a thermal stress evaluation procedure (TSEP) that can be used to evaluate the thermal stress performance conditions associated with annealed glass. which presents the variation of the maximum edge stress as a function oforobability_ofbreakage (POB) and perimeter length.

Best Case 20 E d g e Bite. In development of the model that represents the most unfavorable conditions it was assumed that the glass boundary within the edge bite remains at its initial equilibrium temperature throughout the entire thermal exposure.. ~ --Worst Case Assumed ( Conditions .~ 5 10 I. In this conventional model it is assumed that the edge of the glass is supported by the rubber gaskets on both sides for the full extent of the glazing pocket with proper setting and edge blocks so that the glass does not directly contact the ahtminum. m m 30 Figure 4 --Thermal Stress Factor Chart The conventional conditions selected for the current analysis are based on generic characteristics of glazing systems commonly used in commercial construction situations. This closely represents the case where the edge of the glass is in contact with a massive heat sink.76 mm thick. 35 m 3O f" F B tiE g 15 i.114 THE USE OF GLASS IN BUILDINGS The most unfavorable edge conditions occur when the glazing pocket is assumed to have a large thermal mass (heat sink condition) that effectively prevents the edges of the glass from warming when exposed to a solar heat load. Such a situation occurs if the edge bite is encased in concrete. The rubber gaskets are assumed to be 4.. This case represents the most critical situation for the development of thermal stresses. The rubber and aluminum surfaces facing the exterior and interior . These characteristics consist of aluminum framing members with rubber perimeter gaskets on the interior and exterior surfaces of the glass in the glazing pocket.

008 (8 per 1 000). However. respectively. It is believed by the writers that these conditions represent realistic conditions for many practical glazing situations. the acceptable POB for a particular situation should be based on a number of factors including the number oflites in the building and the consequences associated with thermal failures. a brief example demonstrating the use of the procedure is presented. . width. For wind design applications an acceptable POB associated with the occurrence of the design wind event typically ranges from 0.0001 (1 per 10 000) depending upon the application. Formal Thermal Stress Analysis Procedure The first step in evaluating a thermal stress situation is to determine the thermally induced stresses affecting the plate. Figure 4 presents the variation of the thermal stress factor (TSF) in terms of edge bite for thicknesses up to 6 mm glass. Then. In lieu of this process. The units associated with the TSF are kPa/(W/m2).001 (1 per 1 000) to 0. The TSL is obtained by multiplying the total solar absorptance of the glass expressed as a decimal fraction by the intensity of the solar irradiance on the glass surface expressed in W//m . The final selection of the acceptable POB is the responsibility of the glass designer. Logically. 9 Select the type of glass to be analyzed. The design professional should determine the acceptable POB for each application based on each unique situation. the information presented earlier in this paper can be used to estimate the thermally induced stresses. and height. Acceptable Probability of Breakage The concept of acceptable POB for glass subjected to uniform wind loads is well accepted within the glass design community. The maximum thermal stress experienced by the edge of the glass is found by multiplying the TSF by the total solar load (TSL) on the glass given in W/m 2 . a clear definition of the acceptable POB for glass subjected to thermal loadings is not as well defined. The following steps have been organized into a step-by-step procedure to assist in application of the information presented in this paper. Both of these steps are explained below.BEASON AND LINGNELL ON MONOLITHIC ANNEALED GLASS 1 15 building environments are assumed to have the same surface film coefficients as the exterior and interior glass surfaces. Results from FEA models for each of these three cases were used to develop Figure 4. In addition. However. a decision must be made as to whether or not the thermally induced stress for the current situation presents an unacceptable risk or not. Determine the size of the lite. Evaluation of the ThermalStress The most accurate way to estimate the thermal stresses for a particular glass plate is to perform a detailed FEA of the situation under consideration. it seems reasonable to the writers that the designer might wish to contemplate the use of POBs as low as 0.

n . The POB is to be established by the designer. of the lite by summing two times the width.) (6) 9 Determine the solar load by multiplying the maximum solar irradiance (SI) expressed in W/m s by the total solar absorption as follows: SL = SI x As 9 (7) Determine the glazing type that best represents the glass under consideration based on information presented earlier in this paper.. and two times the height. W. 9 Finally.. As is found by subtracting the total solar transmission (Ts) expressed as a decimal fraction and total solar reflection (Rs) expressed as a decimal fraction from 1. O'o. determine the estimated thermal stress by multiplying the thermal stress factor. the allowable stress.h. crt. Then. 9 If the calculated thermal stress. is less than or equal to the allowable stress. selection of the proper POB is based on a range of variables and engineering judgment.0-(T~ +R. Determine the perimeter.. TSF. for the selected edge conditions and edge bite using Figure 4. Select a defensible POB that is consistent with the goals of the project.. then the risk of breakage is judged to be acceptable. is less than the allowable stress. cro. by the solar load as follows: o't. cr. As stated earlier.~t = TSFx SL (8) Evaluation of Risk of Breakage The first step in the evaluation of the risk of breakage is to establish an acceptable probability of breakage (POB). of the lite as follows: P=2xW+2xH (9) 9 Enter the horizontal axis of Figure 3 with the perimeter of the glass plate. H. . TSF. At the intersection of the vertical line with the acceptable POB curve extend a horizontal line to the left to find the allowable stress. Project a vertical line up to the curve that represents the POB selected for the application.. cra. The following steps describe the evaluation procedure.ow. 9 Determine the thermal stress factor.. P.~..o~.11 6 THE USE OF GLASS IN BUILDINGS Determine the total solar absorption (As) from the solar optical properties of the glass.0 as follows: determined using Figure 3. = 1.. t . the risk of breakage is judged to be acceptable.ow. If the calculated thermal stress.

is greater than the allowable stress. it might be desirable to perform a more accurate estimate of the thermally induced stresses through a detailed application of FEA. o'aUow. the conventional case is acceptable by a narrow margin. then the plate experiences differential heating. cr. This means that the best case situation is acceptable by a wide margin. However.5 m x 2. about 26 kPa/(W/m 2) for the conventional glazing system. the mechanics of the development of the thermal stresses have not been well understood. a 6 mmx 1. The TSEP incorporates results of FEA to estimate the magnitude of the thermally induced stresses and an edge strength failure prediction model (ESFPM) that can be used to judge the significance of the calculated thermal stresses. The more detailed FEA might result in a more acceptable outcome than the more conservative procedure discussed above.. The allowable edge stress is found to be about 11.ot. and 32 kPa/(W/m 2) for the worst case situation. designers have been forced to deal with thermal breakage using anecdotal and empirical results. the probability of breakage might be unacceptable. It is found that the TSF is about 14 kPa/(W/m 2) for the best case situation. The total solar load experienced by the glass is then calculated to be 403 W/m 2 . If the assumptions are judged to be inappropriate for a particular . then the risk of breakage is judged to be unacceptable. or if there are exacerbating circumstances such as shadows. The COl'responding edge stresses are then calculated to be 5. Thus.6 MPa for the best case.3 MPa for a probability of failure of 0. Otherwise. If the risk of breakage is judged to be unacceptable. it is the responsibility of the designer to examine these assumptions. Procedures and guidance are presented in this paper that allow a designer to estimate the magnitude of the thermal edge stresses induced in a glass plate. and the worst case situation is not acceptable. Conclusions If the middle area of a glass plate is heated by sunlight while the edges of the glass are shielded by edge support conditions. and to make sure that the assumptions are compatible with the application under consideration.5 MPa for the conventional case. It is assumed that this glass plate is subjected to a thermal load of 630 W/m 2 . Further. the glass being examined should be heat-treated or another type of glass should be selected.9 MPa for the worst case. and 12. This procedure is based upon assumptions and idealizations that are believed to be reasonably representative of many situations.64 is evaluated. 10. This paper presents the basis for a thermal stress evaluation procedure (TSEP) for annealed glass. it is assumed that it has an edge bite of 19 ram.h~. If a lower probability of failure is desired because this glass is installed in a major structure with hundreds of windows.008 (8 lights per 1 000) from Figure 3. 9 Example of Thermal Stress Analysis In this example. While this general phenomenon has been widely recognized within the glass industry.4 m glass plate with a total solar absorptance of 0.BEASON AND LINGNELL ON MONOLITHIC ANNEALED GLASS 1 17 If the calculated thermal stress. The thermal gradients that develop in the glass induce tensile stresses along the edge of the glass that can he sufficient to initiate failure of annealed glass in some cases.

GA. L. Atlanta. and the effects associated with the use of insulating glass. No. Acknowledgments The writers gratefully acknowledge support provided by Cardinal IG and Visteon Float Glass Operations for some aspects of the research reported herein. and Bracci. 2. 124. 150 Beta Drive. "A Statistical Theory of the Strength of Materials. it is recognized that this new approach is largely untried.Version 12. This is accomplished through the application of the ESFPM. L. the effects of interior shading devices. The ESFPM was used to develop a relationship between allowable thermal edge stress. "Basis for ASTM E 1300 Annealed Glass Thickness Selection Charts. Sweden. This chart allows a designer to make a decision as to whether or not a particular thermal stress is acceptable or not." American Society of Heating. W. [2] "ASHRAEHandbook.. While the general procedure presented herein is a major step forward. PA. 2001. [3] Beason. Inc.." Ingeniors. Handiingar NR151. however. The proposed thermal stress evaluation procedure presented in this paper allows a designer to make a reasonable decision for different glass applications. the effects of the use of low emissivity coatings.. . Refrigerating and Air-Conditioning Engineers. the thermal stress procedure does not address the effects of shadows on the development of thermal stress. The writers are currently in the process of incorporating these and other factors into the TSEP. 2001. It is believed that the assumptions and parameters that are incorporated in the analyses are reasonable. or if the simplified procedure yields results that are marginal. 215-221. 1998. Once the thermal stress in a particular situation is determined through the application of the simplified procedure or through a rigorous application of FEA. perimeter of the glass. Fundamentals. Vol. M. pp. ASCE. References [ 1] "ALGOR Finite Element Analysis Software .. and the acceptable POB.118 THE USE OF GLASS IN BUILDINGS application." Journal of Structural Engineering. T. a designer should consider formulating a complete FEA for the situation under consideration.. [4] Weibull."ALGOR.. Inc. J. the effects of interior heat traps.vetenskapsakademiens. Kohutek. the significance of the calculated stress must be evaluated. This relationship is summarized in the POB chart presented in Figure 3. W. Pittsburgh.. 1939. Stockhom.


Many buildings have glass areas too large to accommodate shutters. Abstract: There is an enormous inventory of buildings in the United States. Pilcher. Sciaudone. Before Alicia. The majority of these buildings have not been designed to resist hurricane effects." ASTM STP 1434. West Conshohocken. 121 Copyright9 by ASTMInternational www.. Institute for Business and Home Safety. Block. Beers. Sciaudone 3 Retrofitting Commercial Structures with Laminated Glass to Withstand Hurricane Effects Reference: Beer. Ed. 2002. Suite 208. Commercial structures have a history of damage during previous hurricanes. requires unique considerations. ASTM International. it was thought that high winds alone were responsible for hurricane damage. Associate Director of Engineering. Protecting commercial structures from hurricane elements. even if all newly constructed buildings were designed for hurricane effects. "Retrofitting Commercial Structures with Laminated Glass to Withstand Hurricane Effects. 2 and Jeffrey C. 22 State Farm Road... FL 33463. Use of Glass in Buildings. .. R E. t Mark A.. 2Underwriting/Operations Superintendent. 5700 Lake Worth Road. Pilcher. Even if shutters can be employed it may be impractical to Chief Executive Officer. Tampa. 1408 N. LA. including the Kendall area in Dade County. While mitigation efforts for newly constructed buildings are very important. However..astm. namely wind and windbome debris.Paul E. it will take many years before a significant percentage of commercial buildings are protected. Without retrofitting. FL 3361 I. namely high winds and windbome debris. C. Suite 100. Westshore Blvd. The vast majority of the built commercial building inventory is not protected. M. J. Monroe. Texas suffered extensive glass breakage during Hurricane AlMa in 1982. Florida during Hurricane Andrew in 1992. State Farm Fire and Casualty Company. Lake Worth. Many large buildings in downtown Houston. V. recently there have been several more instances of glass breakage during hurricanes have occurred. Glazing Consultants Inc. A. This was the first documented instance of windbome debris causing significant glass breakage in an urban area. this activity only benefits a small percentage of the built inventory.

storm shutters. hurricanes.Houston Skyscraper that Experienced Extensive Glazing Damagefrom Windborne Debris During Hurricane Alicia . This paper will present methodology for a cost-effective retrofit using laminated glass and the existing window frames. 25 percent of businesses that close following a natural disaster never reopenvl. missile impact test. on average. Laminated glass is often thicker than ordinary window glass and it must be attached or anchored to the window frame so it remains in place. laminated glass. This phenomenon. The challenge for retrofitting commercial windows with laminated glass is to find a cost-effective method that does not require replacement of the framing as well. However. And. if broken. most commercial building envelopes are not designed to resist hurricane effects. can have a dramatic effect on people being able to go back to their jobs and maintain "normal" lifestyles. Introduction Communities can survive natural disasters only if their businesses survive. Often the best solution for a commercial building is to use impact-resistant laminated glass. namely high winds and wind-borne debris. The Insurance Information Institute reports that. With laminated glass. The key to keeping these businesses from closing is limiting the damage they experience as the direct result of a storm.122 THE USE OF GLASS IN BUILDINGS install them in advance of a storm. hurricane protection. Keywords: cyclic pressure cycles. impact resistant glass. which diminishes the tax base. there is no special preparation required in advance of a storm as the protection is always in place. Figure 1 . The use of laminated glass for hurricane protection requires special designs. it provides invisible protection because it appears as ordinary window glass. in turn. retrofit.

Yet few of their most important structures hospitals. the business will be able to reopen sooner. but are not a perfect solution for several reasons. And laminated glass offers invisible protection that does not detract from the . building owners may choose to protect their businesses with impact resistant window and door products. Norfolk and Jacksonville are vulnerable to hurricanes. because they are not frequently opened and closed. It will take many years before a significant percentage of commercial buildings are protected. especially if the windows are above the ground floor. ON HURRICANE EFFECTS 123 When Hurricane Alicia struck Houston. Retrofit solutions are necessary to reach the massive inventory of existing commercial windows vulnerable to wind-borne debris. hardware can become rusted and the shutter difficult to operate. This was the first documented instance of wind-borne debris causing significant damage in an urban area2. many large commercial buildings suffered extensive glass breakage. stronger frames and more secure anchorage back to the structural system of the building. most people assumed that high winds alone were responsible for hurricane damage to commercial buildings. New Orleans. These impact resistant products consist of laminated glass. These systems require no preparation in advance of a storm . West Palm Beach. many commercial buildings have glass areas too large to consider using protective shutters. it is more important to protect the building envelope to prevent destruction of the interior and contents of businesses. government buildings. If the envelope stays intact. While keeping the structure intact is important. Along the hurricane coasts. Texas in 1982. the addition of shutters may dramatically alter the appeararlce of the building. In order to alleviate these concerns.have wind-borne debris protection. strong but flexible sealant to hold the glass in place. even if all new buildings include impact-resistant features. First of all. Before this event. ASCE 7 requires the designer to either provide protection of glazed openings from wind borne debris or design for internal pressures. Need for Retrofit Methodology Coastal cities like Houston. Storm shutters are a viable option for protecting low-rise buildings. the people who are supposed to ensure that the shutters are properly closed and ready for the storm may be more interested in securing their own homes and evacuating the area. corporate offices and multi-family residences .the protection is always in place.BEERS ET AL. Second. new buildings typically represent about 2% of the building stock. Third. Tampa. Fourth. In any given year. Figure 1 shows some of the damage that occurred during Hurricane Alicia. they only protect new buildings. have shown that wind-borne debris plays a major role as well3. And fifth. The damage from Alicia and several more recent hurricanes. including Hurricane Andrew in 1992. The International Building Code now refers the designer to the procedures of ASCE 798 for the calculation of wind loads on structures 4. While these code developments are very important. deploying shutters in the face of an approaching storm may prove difficult.



building's appearance. However, new impact resistant window systems.can be prohibitively expensive to install in existing buildings. In response to repeated problems, engineers are beginning to develop retrofit solutions to incorporate laminated glass into existing window frames. The laminated glass used is often thicker than ordinary window glazing, and as such, presents some unique design considerations. The challenge for retrofitting commercial windows with laminated glass is to find a cost-effective method that does not require replacement of the framing as well.

Overview of Methodology Most existing commercial buildings were built following the requirements of earlier building codes. Many do not comply with current building code requirements.
Recent changes in building codes, such as The South Florida Building Code and The International Building Code now require that windows, doors and glass comply with higher wind loads and impacts from windbome debris. Wind pressure is calculated using ASCE 7 Minimum Design Loads for Buildings and Other Structures or by conducting a wind tunnel study where a scaled building model is subjected to a simulated wind field. Missile impact criteria is based upon performance tests where materials intended to provide hurricane protection are tested in a laboratory by subjecting them to missile impacts followed by a series of wind pressure cycles. A typical test sequence includes impacting three specimens twice with a nine pound two by four traveling at fifty feet per second, followed by subjecting the impacted specimens to 9,000 inward and outward (4,500 inward and 4,500 outward) acting pressure cycles6. Outside of recently constructed buildings in south Florida, few commercial buildings provide protection from impact from windborne debris. Thus, most existing buildings require retrofit or renovation to provide meaningful hurricane protection. To retrofit an existing building to provide hurricane protection, the glazing system must be evaiuated for resistance to wind pressure and protection from windborne debris. This includes analysis of the window frame, anchors and the glass product. The first step in the retrofit process is to determine appropriate design criteria. This includes calculating the wind design pressure and selecting missile impact performance criteria. The South Florida Building Code, International Building Code, Southern Building Code Congress SSTD 12-97: Test Standard for Determining Impact Resistance from Wind-Borne Debris and ASTM Standard Test Method for the Performance of Exterior Windows, Glazed Curtain Walls, Doors and Storm Shutters Impacted by Missile(s) and Exposed to Cyclic Pressure Differentials (E 1886) and ASTM Standard Specification for the Performance of Exterior Windows, Glazed Curtain Wails, Doors and Storm Shutters Impacted by Missile(s) and Exposed to Cyclic Pressure Differentials (E 1996) provide guidance with the selection of appropriate missiles, impact speed and cyclic pressure loading.



Some critical facilities such as hospitals, police and fire rescue stations, corporate data centers and buildings to be used as shelters require design criteria that exceed minimum building code requirements. ASTM E 1886 and ASTM E 1996 provide design criteria for critical facilities. Enhanced design criteria can also be developed through the use of a site-specific wind risk analysis by modeling the historic wind records for the site and selecting an enhanced missile based upon the debris threat at the site and precedent for similar projects. Once the design wind pressure has been calculated, the existing window frames and anchors must be analyzed. This can be done using theoretical calculations of the existing frames and anchors to determine their strength as compared to the design wind pressure. Normally the existing frames will not meet the current project design wind pressure requirements because they were built under an earlier version of the building code. Additional structural bracing for the window frames can be designed to strengthen the framing to the current project requirements. This often includes the application of structural steel elements to the interior or exterior of existing vertical framing members. The new steel usually must be anchored to the structure at the top and bottom. The steel and anchors can be painted to match the existing frame color or clad with matching aluminum sheet break metal. Existing anchors also are typically not compliant with the new project design wind pressure requirements. This is often due to inadequate sizing and spacing or corrosion. A new anchor schedule to comply with the current project requirements can be developed through calculations. A conservative anchor design would be to abandon the existing anchors and design new anchors without relying upon or using the benefit of the existing ones. Most building code jurisdictions require submittal of a full set of calculations illustrating the additional structural elements necessary to comply with the project design wind pressure. These calculations must be signed and sealed by a Registered Professional Engineer. Existing buildings predominately utilize annealed, heat-strengthened or tempered glass in monolithic or insulated configurations. These materials do not pass the minimum missile impact tests prescribed by building codes and standards 7. Therefore, it is necessary to replace the existing glass with a material that is capable of passing the project's missile impact test criteria. The new glass is typically a multi-layer laminated glass unit and is thicker than the existing glass it is replacing. The challenge is how to provide enough space in the existing frame to accommodate thicker glass and a silicone anchor bead to affix the glass to the frame. This is accomplished by utilizing an aluminum frame adaptor that fits into the space where the existing glass was and provides a wider glass pocket for the new glass and silicone. The adaptor decreases the daylight area of the glass by about 3 mm all the way around, but is not noticeable if painted the same color as the existing frame



material, Figure 2 shows examples of impact resistant glazing systems with and without the adapter.

Figure 2 - Retrofit GlazingSystem Building codes require performance testing of the new glass adaptor in the existing frame to show compliance with code and project requirements. Sometimes it is not possibte to get an exact replica of the existing framing, if the system is now obsolete. In these instances, framing that replicates the glass pocket condition should be used for the laboratory test. The combination of engineering calculations and laboratory testing of the retrofit assembly provides a system that can be accepted by building code authorities. More importantly, with proper design criteria, these systems will provide reliable performance during hurricanes at a lower cost than new windows and some shutter systems. Buildings that have been retrofitted with laminated glass provide additional benefits to the owner and occupants. The laminated glass system is always in place and requires no advanced preparation in advance of a storm. The retrofit can be designed to have little visual changes from ordinary windows and is not noticeable to most building occupants. Laminated glass also provides additional benefits beyond hurricane protection to include enhanced security from intruders, reduction of ultra violet rays, better acoustical performance and it is safer than ordinary window glass because it remains in the frame when broken.
Case Studies

Tarpon Springs Chamberof Commerce; TarponSprings, Florida

They were not suitable for retrofit and were replaced with South Florida Building Code approved assemblies. United States Virgin lslands This project was renovated following heavy damage from Hurricane Marilyn in 1995. Figure 3 . The selected design criteria was the wind pressure requirements for Miami. which at the time were more stringent than the loads for St.BEERS ET AL.Tarpon Springs Chamber of Commerce The building contained glazing along the street front and in an upper clerestory area in the rear of the office reception area. The project consisted of multiple buildings serving functions including reception. A retrofit laminated glass panel similar to a tested system was designed and installed. meeting space. Luxury Hotel. Thomas and the missile impact criteria from The South Florida Building Code. Evaluation by a Florida Registered Professional Engineer determined that the window frames and anchors did meet current wind load requirements and no additional bracing or anchors were required. The cost of the project was deemed to be similar to the cost of accordion or roll down storm shutters. . The requirements of ASTM E1886 and ASTM E1996 were used as the design criteria. ON HURRICANE EFFECTS 127 The Institute for Business and Home Safety and Tampa Bay Regional Planning Council selected this building as a demonstration project for commercial glazing retrofit in Florida. Tarpon Springs Chamber of Commerce is an important organization within the local community and it is important that they are ready to support local business during the aRermath of a storm. The common area buildings contained custom shop built wood framed windows and doors. restaurants and guest rooms. The building will not be used as a shelter during a storm." St. Thomas. It is in the downtown section of Tarpon Springs that is considered to have high windborne debris potential. They were heavily damaged during the hurricane and many exhibited wood rot. Florida using ASCE 7.

the hotel operators did not want shutters because the appearance and the logistics of deploying them as a storm approached.128 THE USE OF GLASS IN BUILDINGS Figure 4 . St. Additionally the airport is a critieai part of the country's infrastructure as it is the point of entry and departure for residents and visitors to the island. Additional anchors were installed as required. United States Virgin Islands The guest rooms contained aluminum sliding glass doors. Cayman Islands. British West lndies The Cayman Islands Government decided to retrofit the airport to serve as a hurricane shelter in the event a hurricane struck the island and all visitors could not be evacuated. The specified large missile is a 15 pound 2 x 4 timber traveling at 65 feet per second. A new aluminum framed laminated glass panel was installed to replace the existing tempered glass that was not shattered during the storm. Engineering calculations determined the need for additional structural bracing and anchors. The study recommended enhanced wind speeds be used to calculate the wind pressure design loads using ASCE 7 and an enhanced large missile. . The airport facility must remain operational immediately after a hurricane. which fared better in the storm. The terminal building was built in the 1970s and has had alterations and additions since. Additionally. Owen Roberts Airport. Thomas. A wind risk analysis was commissioned for shelter design criteria in the Cayman Islands. Steel reinforcement tubes clad in tubular aluminum extrusions were installed behind each vertical framing member and anchored at top and bottom to the structure. Most of the glazing systems are similar and there are many operable doors of various types.Luxury Hotel. The cost for the retrofit was less than several cost proposals for accordion shutters." Grand Cayman. A retrofit design was developed to add aluminum clad external steel bracing and re-anchor the frames.

There were some areas at the airport.NY. Because shutter application is not practical at many commercial buildings. that were not suitable for the glazing retrofit.InsuranceInformationInstitute.ChiefEconomist.Large Missile Testing Retrofit glazing panels for the glass and doors were designed and installed. In addition to saving money. retrofit is the only way a significant portion of these buildings will be protected from hurricane damage. Accordion storm shutters were installed at this location.New York. a majority of the terminal will contain laminated glass systems that are designed and tested to the enhanced Cayman Islands shelter criteria in combination with a few storm shutters. The case studies presented show that commercial buildings can be cost effectively retrofitted with laminated glass using the retrofit method described in this paper. As demonstrated in the case studies. Prior to final approval of the design.July2000. Conclusions Windows in existing buildings in hurricane-prone areas can be cost-effectivelyretrofit with laminated glass to provide protection from windbome debris. This work can usually be performed while the building is occupied and operational and does not damage interior or exterior finishes. Building owners need to be educated about this option. Most owners are not aware of the retrofit option. the methods presented in this paper have proven to be cost effective and practical. ON HURRICANE EFFECTS 129 Figure 5 . A laminated glass with a polycarbonate core was selected. the panels were tested using various laminated glass types to determine which would perform satisfactorily. there is much less disruption to the facility than when replacing entire window and door systems. [1] IRobertHartwig. Upon completion. .BEERS ET AL. such as a large breezeway. Personal Communication.

. Falls Church. 1999. . DC. 1994. Washington. TX. International Building Code. University of Western Ontario. MA. and the Building Envelope Invitational Seminar. (1985). "Wind Damage to Envelopes of Houses and Consequent Insurance Losses". Proceedings of Wind. P. Hurricane Alicia: One Year Later (Galveston. ICC. August 16-17. 1999." IBHS. Impact Resistance Standards. Industry Perspective. Schiff. [7] ~Institute for Business & Home Safety. Impact Resistance Standards. "Minimum Design Loads for Buildings and Other Sl~'uctures. Industry Perspective. 1984). T."American Society of Civil Engineers.. VA. Rain. MA.130 THE USE OF GLASS IN BUILDINGS [2] 2Minor. 2000. Boston. Boston. 1998. [4] 41nternationalCode Council. ASCE. S. J. [5] 5ASCE 7-98. "Natural Hazard Mitigation Insights. Reinhold. "Window Glass Performance and Hurricane Effects. "Natural Hazard Mitigation Insights.9." Proceedings.. Section 6. New York." IBHS.R.E.5.3. [3] 3Sparks. pp 151-164. [6] 6Institute for Business & Home Safety.

E.K. 4 Senior Instrumentation Specialist. Janney. Inc. "Standard Test Method for Performance of Exterior Windows. The intent of these standards is to minimize property or personal loss due to the breaching of fenestration elements in a hurricane. Kaskel 1. An existing building with a history of thermal stress glass breakage required a system to safeguard against the possibility of glass fallout. Chicago. and the pass/fail criteria were significantly different from those presemed in these standards. and Storm Shutters Impacted by Missile(s) and Exposed to Cyclic Pressure Differentials" and its accompanying specification.. Elstner Associates.E.. Testing was necessary to determine the effectiveness of an anchored-film system for this application.. ASTM International. the nature of anticipated wind loads.. 2002. Northbrook. Structures I Department. Ed.S. The film is anchored around the glass perimeter to the window frame. Inc. Wiss. 120 North LaSalle Street. .. Protective systems for both new and existing buildings have been developed in response to these new standards.. John E.. The aforementioned ASTM standards were consulted. IL 60602 2 Senior Engineer. J Senior Consultant. Curtain Walls. 131 Copyright9 by ASTMInternational www. West Conshohocken. "Standard Specification for Performance of Exterior Windows. Wiss. Pearson 2. Janney. Glazed Curtain Walls. IL 60062. This system relies on a polyester film adhered to the existing glass. PA. ASTM STP 1434. and Pelletier. One possible application for existing building glass is an anchored-film system. Test criteria were developed by the authors to simulate the cracked glass. and Roger E. Wiss. Doors. implementation. IL 60062. 330 Pfingsten Road. Janney. M. Elstner Associates. This paper summarizes the development. Abstract Recently adopted ASTM Standard Test E 1886-97. Mark K.astm." Use of Glass in Buildings. Elstner Associates. to apply appropriate loads and to measure the degree of fallout protection. Schrnidt 3. 330 Pfingsten Road. Northbrook. 330 Pfingsten Road. E 1996-99. Structures I Department.Bruce S. 3 Consultant. B. Chicago Branch. Inc.. Block. Schmidt. Elstner Associates. however. Pearson. Wiss. Northbrook. J. Inc. thereby demonstrating the feasibility and effectiveness of anchored-film systems for this unique application. IL 60062. Doors. Pelletier4 Testing of Annealed Glass with Anchored-Film Glass Retention Systems for Fallout Protection After Thermal Stress Cracking Reference: Kaskel. Suite 2000. the cause of cracking. and results of the test program. and Storm Shutters Impacted by Windbome Debris in Hurricanes" are used to determine the the performance of exterior fenestration elements when subjected to hurricane-like conditions. V.. "Testing of Annealed Glass with Anchored-Film Glass Retention Systems for Fallout Protection After Thermal Stress Cracking. Janney.

temperature differential. This may exceed the edge strength of annealed glass in some cases. The windows tested consisted of annealed glass measuring approximately 203 cm (80-in. Figure 1 shows window temperature differences measured at the center and edges during a typical sunny day during the spring. temperature differentials as high as 25~ (45~ were measured. "~dowT~ Data Apdl. fallout protection.oo 144or-oo Figure 1 . A q 10 t. which is the same thickness as the glass in the building.) high by 292 cm (115 in.) and 12 mm (V2-in. Temperature differentials of this magnitude can be large enough to produce edge thermal stresses greater than 2000 psi.-lO -20 -30 7-~. The glass thickness was 10 mm (3/8-in. load test.33d Roar. For this study..00 8-.~xeh 50 40 ~_30 . fenestration. Window glass has cracked on occasion and a system to safeguard against the possibility of fall-out of the cracked glass was investigated and tested. As can be seen in Figure 1. glass retention Introduction The windows of a Midwest high-rise building consist of gray-tinted annealed glass set into aluminum frames that are recessed from the face of the building's steel cladding. annealed glass. Because of the recessed configuration.00 g-~ir-00 io.. causing large temperature differentials between the center and edges of the glass.). various anchored-film glass retention systems were tested. shadows are cast across these windows.132 THE USE OF GLASS IN BUILDINGS Keywords Anchored-film.~.) wide and glazed in aluminum window frames.~r-00 n-~r-00 t~Vr-00 134g.Temperature difference between center and edge of window .

The use of anchored-film at the subject building was intended to safeguard against the possibility of fall-out of cracked glass. Applying an anchored-film glass retention system to these test specimen windows and allowing 21 days to cure. One condition that could cause glass to fallout involves the application of an external force to the cracked glass. Doors. and Storm Shutters Impacted by Missile(s) and Exposed to Cyclic Pressure Differentials" to specify a protocol for testing glass that has been cracked. The aforementioned ASTM standards do not address the subject conditions related to the cause of cracking. or a 40 g. Protective systems for both new and existing buildings have been developed in response to these new standards. since it potentially produces higher magnitudes and more cycles of load application than the other forces. the nature of anticipated wind loads. These standards consider the wind-borne debris that could impact the fenestration during the storm and the cyclic nature of the hurricane wind forces. Wind pressure is the most critical. the crack is induced by projecting either a 10 g. Test Specimens and Load Test The load test for the subject building consisted of: 1. The film is anchored around the glass perimeter to sound construction. ON FALLOUT PROTECTION 133 Industry Testing of Cracked Glass Annealed glass of the size and thickness used on the subject building will typically remain in the window frame after a thermal stress crack. the glass is tested through repetitive cyclic static loading. The intent of these ASTM standards is to minimize property or personal loss due to the breaching of fenestration elements that can commonly occur in a hurricane event. 2. vibrations. Another ASTM standard. nor an appropriate pass/fail criterion. "Standard Specification for Performance of Exterior Windows. intended to replicate the type of forces that can occur during a hurricane event. Doors. E 1996-99. This system relies on a polyester film adhered to the existing glass. missile at high velocity at the glass specimen. Curtain Wails. Following this impact. ASTM has recently adopted standard test E 1886-97. and Storm Shutters Impacted by Windbome Debris in Hurricanes" provides pass/fail criteria for the performance of glass subjected to the E 1886 test method. such as the window frame. Glazed Curtain Walls. Test criteria were therefore developed to simulate the cracked glass. In the case of E 1886. to apply appropriate loads and to measure the degree of fall-out protection. The passing criteria is intended to ensure that the fenestration will maintain the building enclosure.KASKEL El" AL. . and wind pressures. Load tests to simulate the critical cyclical wind pressures were performed on the test specimens after the glass was intentionally cracked. "Standard Test Method for Performance of Exterior Windows. One possible application for existing building glass is an anchored-film system. This external force can be due to normal pressure differences between the interior and exterior of the building by the air distribution system. Constructing test specimen windows that replicate the existing windows.

5. I Glazing Gasket T~icad "InFt~'ramoSide Figure 2 ." For these systems.Section through test specimen at jamb before the application of the anchored-film glass retention system (head and sill similar) Glass Retention System Application The anchored-film glass retention system consisted of a clear polyester composite film from 0.006 in.) thick glass and five had 12 mm (89 thick glass.) to 0. Twelve specimens had 10 mm (Vs-in.20 mm (0.) in thickness. Three of the four glass retention systems were "mechanically anchored. Each test specimen received an application of one of four different systems tested. The type of window frame used for the test was a readily available off-the-shelf system selected to resemble the glazing conditions of the subject building windows. A section of the test window is shown in Figure 2.134 THE USE OF GLASS IN BUILDINGS 3. 203 CM By 292 CM By 10 MM or 12 MM MOaoU~c A t o n e d Glazing SealantOr Glass (1/2 In.008 in. Shown) . Intentionally cracking the test specimen glass to simulate the type of initial thermal crack observed on the building. ~ie~ f ' . the polyester film was secured at the perimeter with an aluminum bar that was screwed to the . Test Specimens Seventeen specimens were constructed utilizing monolithic annealed glass sheets set into an aluminum window frames. Experienced and licensed installers approved by the glass retention system manufacturers performed all applications. Subjecting the test specimens to cyclical loads representative of code not clear at this point but later wind loads for the building. 4. Measuring the degree to which the anchored-film system prevents the cracked glass from falling out of the frame. adhered to the inside of the glass and anchored to the window frame.15 mm (0.

" In this case. ON FALLOUT PROTECTION 135 aluminum window frame (Figure 3).Section through adhesively anchored glass restraint system . Film et ~ - 11/ /1/ EXTERIOR~ INTERIOR Figure 3 . Due to the limits of the maximum width of a roll of polyester film. structural silicone sealant was installed around the window perimeter to bond the polyester film to the window (Figure 4). DashedLine Indicates Polyec. In all cases. the film was allowed to cure for at least 21 days before load testing. The fourth glass retention system was "adhesively anchored.Section through mechanically anchored glass retention system DashedLineIndicates PolyesterFilm AdhesiveSealAnt Anchorage 11/ r// EXTERIOR INTERIOR Figure 4 . all four glass retention systems contained a center vertical butt joint on each test specimen.KASKEL ET AL.

determined the magnitude and number of cycles of the test pressures to be apply to each specimen. Perterka. The adhesively anchored system was always anchored on all four sides. From these analyses. Inc. The two-sided condition consisted of the anchorage system installed at the top (head) and bottom (sill) of the test specimen. At the sides (jamb) the polyester film was extended to the sight line but was not anchored to the window frame.5 P .5 P 0. The foursided condition had the anchorage system installed along all four sides of the window frame.0. Table 1 .0 .3 P . 8 P 0. Test Protocol Cermak. filmed window condition [1]. these standards are developed for hurricane-prone areas and with design loads greater then proposed for this test protocol.05 kPa. three test protocols were developed for this project as tabulated below.6P 0.1.5 P ..05 kPa (-22 lb/flz) for windows near the comer or edges of the building. Their analysis to determine the magnitude took into account the temporary nature of a cracked. P = -1.136 THE USE OF GLASS IN BUILDINGS The mechanically anchored systems were installed with either two-sided or foursided anchorage. Petersen. . A reasonable test profile was established as a "scaled-down" application of these standards.2 P . All testing using the negative pressure loads was performed using P = -1. a wind engineering consulting finn.5 P Total Cycles Number of Cycles 575 9 180 17 180 9 575 1545 Pressure. P = -0.0. As previously discussed.0P-0.0 P 0.Negative Pressure Test Protocol Load Range 0. Determination of the cyclical application of the load was developed from a review of current cyclical tests that demonstrate the ability of a building panel to withstand fluctuating wind loads [2-3].0.0P-0.2 P .6P 0.8 P 0.58 kPa (-12 lb/fl2) for the middle portion or field of the building.

8P 0.Positive Pressure Test Protocol LoadRan~ 0. A programmable controller was used to monitor the pressure and communicate with the slide door motor to increase (close slide door) or decrease (open slide door) the interior pressure. The fan evacuated air from or forced air into the test chamber to create a press~e difference. A motor controlled sliding door was used to allow or restrict airflow through the opening in the reinforced panel (Figure 5).5P Total Cycles Numb~ofC~ 1~ 1~ 1~ 1~ 1545 137 P = -1. an electronic data acquisition system recorded the test pressures and center of glass deflection (Figure 6).25P--0.. The interior test chamber pressure varied as the airflow through the opening changed.6P 0. Two electronic pressure gages.5P ~.0P-0.0P-+0. Test Equipment A steel test frame was constructed to support each window in a similar fashion as in the building.6P 0.0P-0.05 kPa and +0. At the completion of a test.8P 0.2P-0. .Combination of Positive and Negative Test Protocol LoadRange 0.0P--0.25P 0. ON FALLOUTPROTECTION Table 2 .5P Total Cycles Numb~Cyeles 6~ 53 103 34 103 53 6~ 1546 P = +0. Loading was achieved by a fan blower connected to the test chamber.2P-0. A reinforced plexiglass panel was bolted to the back portion of the chamber.25P +0.5P-0. a static proof load of 150% of P was applied for I0 seconds. Each test took approximately 6 hours to perform.25P-+0.KASKEL El" At.58 kPa The duration of one cycle for any of the test protocols was approximately 15 seconds.5P-0. During each test. An overload pressure of 250% of P was then applied after the proof load for I0 seconds.0P 0. one for the low and one for the high pressures.58 kPa Table 3 .5P 0. were used to measure the internal pressure.3P-1. An opening was cut into a section of the plexiglass panel.

138 THE USE OF GLASS IN BUILDINGS Figure 5 - Sliding door to adjust the test chamber internal pressure Figure 6 - Computer controlled data acquisition system .

The specimen frame was clamped to resist displacement under application of the test load.Positioning test sample into test chamber .KASKEL ET AL. The test chamber had a plexiglass back panel to allow for visual inspection (exterior side in the actual building) of the specimen during testing. Figure 7 . each test specimen was placed into the steel-framed test chamber with the filmed glass surface (interior side in the actual building) facing the exterior o f the test chamber for visual examination (Figure 7). The test specimen was seated into the chamber creating a pressure seal between the test specimen and the test chamber. ON FALLOUT PROTECTION 139 Test Procedure After a cure period of at least 21 days from the date of installation of the anchored-film glass retention system.

The pattern and length of the crack was patterned after the initial thermal crack typically observed at the building (Figure 8). Id INDICIIA ~ .140 THE USE OF GLASS IN BUILDINGS Once the test specimen was situated in the chamber.Test specimen indicating initial crack pattern Figure 9 .N T E I q l . INE S E A M IN ~ L ~ FIIM Figure 8 . A test specimen during cyclical load application is shown in Figure 9 and a sample pattern of cracking at the end of a test is shown in Figure 10). additional cracks would develop. a crack was created in the glass with a glass cutter and impact tool.Test specimen during testing . As test loads of increasing magnitude were applied to the specimen.

Curtain Walls.One test specimen indicating final crack pattern During the test. ON FALLOUT PROTECTION 141 4. Repeated load cycling caused some grinding of the glass at the crack interfaces.KASKEL El" AL. Evidence of large glass fragments not retained by the anchored-glass film glass retention system were considered failure of the test. and Doors by Cyclic Static Air Pressure Differential." . FILr~ Figure 1 0 . At the end of the entire test. The application of the test loads to the test specimen was conducted by air pressure differences in general conformance with American Society for Testing and Materials (ASTM) test method. the total amount of glass shavings at the bottom of the test chamber was weighed and documented. resulting in small glass shavings collecting at the bottom of the test chamber. E 1233-97 "Standard Test Method for Structural Performance of Exterior Windows. the specimen was closely observed.

Four of the completed tests were performed on adhesively anchored systems. 14 where a piece weighing 1. since this was the more severe loading. which typically abut the glass at mid-width of the window. . 12). Both of these tests passed the cyclical load test. mid-point deflections up tol 5 cm (6-in.) were recorded. Most tests were conducted with the negative pressure load. although one of the tests failed the proof test. Three tests were performed with the specified positive pressure load (P = +12 psf). 6 and 7) and one four-sided mechanical anchorage (Test No. This relationship was not linear and was influenced by the number of cracks in the test specimen.75 kg fell out only during the 250% load test. Between 100 g and 1 kg of glass were weighed for each test. except for test No.05 kPa). after passing the cyclic load test.1. 6. 16) did not pass during the cyclic load test. Twelve of the completed tests were performed on glass specimens with an anchored-film glass retention system that was mechanically anchored to the window frame. Eight of the twelve completed mechanical anchorage tests had two-sided anchorage and the other four had four-sided anchorage. the accumulated glass shavings and fallout pieces were weighed. vertical strip of the film removed from the window at the center of the test specimen. Two additional two-sided mechanical anchorage tests. 3. failed the proof load test at 120% (Test No. 11) and 130% of the test load (Test No.142 THE USE OF GLASS IN BUILDINGS Test Results Test results for each sample are presented in (Table 4). 4. At overload test pressures. Thirteen tests were performed with the specified negative pressure load for comer conditions (P = . Two two-sided mechanical anchorage (Test No. The first of these tests was stopped shortly after beginning to check test equipment. Deflections increased as loads increased. At the completion of all testing. The intent of this modification was to simulate the condition in the building where the film would not be applied continuously across the glass due to obstructions from existing partition walls. After two of these three positive pressure load tests were completed. All adhesive anchorage tests passed both the cyclic load test and the proof test. 5. Eleven mechanically and adhesively anchored systems passed the cyclic load and proof test requirements that were established before beginning the tests. These tests demonstrate the following regarding the anchored-film glass retention systems: 1. the same test specimens were additionally subjected to the positive/negative pressure test. Two test specimens had a 10-cm. 2.

Fallout piece weighed 40 g. 17 12 mm 4 Sided(2) Positive Pos. Fallout piece weighed 90 g. Passed Passed * 10 cm. wide film strip removed at the center to simulate partition wall against glass.Load Test Results Test No. 1 2 3 4 5 6 Glass Thickness 10 mm 10 mm 10 mm 10 mm 10 rnm 12 mm Extent o f Pressure Cyclic 150% Proof Anehoralge Direction Load Load 2 SidedtL) Test stopped to check equipment 4 Sided(2) Negative Passed Passed 2 SidedO) Negative Passed Passed 4 Sided (2) Negative Passed Passed 2 Sided(l) Negative Passed Passed 2 Sided(l) Negative Fall-out at --cycle 655 Comments 7 10ram 2 SidedO) Negative Fall-out at cycle 843 Fallout piece weighed 480 g. 12" 13 14 15 16 10 mm 12 mm 10 rnm 10 mm 12 mm 4 4 4 4 Sided(z) Sided(l) Sided(1) Sided~ 4 Sided 0) Negative Negative Positive Positive Pos.KASKEL ET AL. (~)Mechanically anchored (Z)Adhesively anchored . Negative Passed Passed Passed Passed Fall-out at cycle 1379 Passed Passed Passed Passed --Fallout piece weighed 370 g./Neg./Neg. 8 9 10 12 mm t0 mm I0 mm 2 Sided0) 2 Sided0) 2 SidedCt) Negative Negative Negative Passed Passed Passed Passed Passed Fall-out at 120% I 1" I 0 mm 2 SidedO) Negative Passed Fall-out at 130% Fallout piece weighed 170 g. ON FALLOUT PROTECTION 143 Table 4 .

Doors and Storm Shutters Impacted by Missile(s) and Exposed to Cyclic Pressure Differentials. "Winds Speeds for Design of Temporary Structures." ASTM E 1886-97. [3] American Society for Testing and Materials. Curtain Wall. Glazed Curtain Wall. and Peterka." ASTM E 1996-99. W. "Standard Specification for Performance of Exterior Windows. The testing was conducted using annealed glass of the same size and thickness as actually used on the subject building.144 THE USE OF GLASS IN BUILDINGS Conclusions Laboratory testing of anchored-film glass retention systems was performed to determine the degree to which such a system could retain cracked glass under simulated wind loading conditions that were determined from a probability analysis considering the building's location and duration of exposure. D. and the interpretation of the test results. Doors and Storm Shutters Impacted by Windborne Debris in Hurricanes. The tests determined that an anchored-film glass retention system can provide an adequate level of glass fallout protection for a window that incurs a thermal crack and is exposed for one to eight days prior to being replaced. However. 1992." Structures Congress 1992 Compact Papers. the new test protocol testing established different criteria with regard to the induced glass crack. The test protocol used concepts developed in ASTM E 1886 and E 1996. ASCE. "Standard Test Method for Performance of Exterior Windows.A. [2] American Society for Testing and Materials. References [1] Boggs. These criteria were modified to more accurately test the performance of cracked glass with an anchored-film glass retention system at the subject building. J. the number and magnitude of applied static cycles. .


strategic use of glazing materials may be the answer. Smith and C. This is most evident in a crime Overland Park Kansas Police Department. The concerns people have over glazing materials in their homes or businesses often result in them taking actions that result in fire hazards. 147 Copyright9 by ASTM International www. Keywords: glazing.. The perceived vulnerabilities of glazing are self-evident. we must weigh actual threats while addressing the fear of clients." A seemingly simple solution? However. "if you want a fortress. A. and enhancing the overall quality of life. windows. Jones.. 2002.Michael Betten ~and Henri Berube The Advantages of Glazing in an Overall Security Strategy Reference: Betten. Burglar bars and double cylinder deadbolts are becoming more common in residential applications creating fire hazards and business owners frustrated with their inability to protect their assets from crime are resorting to unsightly. 12400 Foster. B. PA. ineffective alternatives. One homebullder recounted his personal feelings on the topic of residential security by saying. D. West . Berube." The Use of Glass in Buildings. "The Advantages in an Overall Security Strategy. H. egress capabilities.. ASTM International. The challenge is to balance the heightened need for security against the fortress mentality. security. KS 66213. the risk of detection / apprehension becomes unacceptable to the malfeasant. OverlandPark. criminal interviews and crime prevention practitioners suggest that if you want to increase security. and a statement about the community. Our role as crime prevention/security practitioners must focus on educating the public on effective security products while at the same time maintaining suitable aesthetics. The goal should be to increase security and not trade to protect against one threat at the expense of other life safety systems and strategies. ASTM STP 1434. burglary The prevailing public sentiment is that glazing materials in a business or home are extremely vulnerable to penetration. The general concept being that if criminals perceive that they may be observed. as individuals who routinely consult with citizens on security issues.astm. a sacrifice in aesthetics. don't put any windows in it. Research. Eds. but what evidence is there to support the fact that glazing materials can actually enhance an overall security strategy? Many theories that try to explain criminal behavior make mention of or reference an absence of witnesses or suitable guardians as providing the opportunity for crime to flourish and escalate the likelihood of criminal activity. M.

" (Cromwell. These measures. traditional unattractive target hardening approaches such as bars.. Traditional target hardening measures send a definite message about an area or neighborhood and can affect the overall quality of life. Interviews with burglars further support the claims made by CPTED practitioners. Wright 1994:110) Burglars perceive windows as a vulnerability of detection where possible witnesses can observe their activity and relay that information to the proper authorities. I wouldn't do this place. No specific information was mentioned regarding attacks on glazing. and territorial reinforcement (Crowe 1991:). barbwire. The three main components of CPTED are: natural surveillance.148 THE USE OF GLASS IN BUILDINGS prevention philosophy known as Crime Prevention Through Environmental Design (CPTED).Gary. But the question has to be asked. " (McKay 1997) CPTED practitioners rely heavily on natural surveillance because without it. and raised entrances that encourage observation by the entire population. One significant study on burglary found: "The location and type of windows both at the target site and at neighbors" houses were considered critical by almost all informants.Detectivewiththe Piano. Plano. they are not opposed to breaking glass to gain entry. creating a palpable crime deterrent and a sense of proprietary ownership in the residents. prominence is given to no obstructions. resident's windows are closed due to climate conditions. expansive windows. Key to the performance of these criteria is the concept of natural surveillance. the risk of detection is relatively low since the noise of breaking glass is mitigated by many closed windows. 2 The police department suggests glass breakage is a problem since most of the year. (Decker. residences that back up to woods." One informant stated: "Notice how that picture window looks out onto the street. . parks and isolated or unincorporated areas are at a more significant risk to glass breakage. For example. Therefore. "Many of the offenders turned to the second method. The National Burglar and Fire Alarm Association indicates burglars gain entry into homes 23% of time through a first floor window." Although burglars perceive windows to be a risk. Other research seems to contradict the assertion that surveillance prevents burglaries. Olson. Depending on what part of the country is referenced. why the contradiction? 2Griffin. TX listed a significantly high number of residential burglaries where glass was broken to gain entry. Avary 1991:35) Other research further supports this assertion. "When CPTED is applied to surveillance. glass breakage can be unusually high. and concrete barriers could be possible alternatives. on a survey offered to many law enforcement agencies. low landscaping. access control. a problem that can feed on itself. Wright and Decker (1994:1244/1250) found that burglars are not afraid to break glass. Glass breakage is more likely when the potential entry point is isolated from witnesses or neighbors. if not applied in conjunction with quality of life considerations can signal the beginning of urban decay and fear. The curtains stay open all the time and both the houses across the street can see straight into the living room.. For example. breaking the window.TX PoliceDepartment(2002).

This brings us back to the prevailing sentiment: Can overall security strategies be enhanced with the use of appropriate glazing materials? Before answering this question. This assessment would be established on a case-by-case basis. insurance premiums. another popular burglary prevention measure for both businesses and homeowners. D. just to name a few. Commercially other tangibles must be considered when a burglary occurs. Repairs to the facility. 1994:124) Unfortunately. To combat the false alarm problem. an additional expense for alarm owners. they do not keep intruders out! Most systems are designed to activate . one being false activation. some law enforcement agencies have stopped responding to alarms until the activation has been verified by private security services. By using appropriate security glazing materials. According to the UCR. Over time. have a few vulnerabilities. let's examine some of the other security alternatives often used by both homeowners and businesses.BE'I'3"EN AND BERUBE ON ADVANTAGES OF GLAZING 149 How then do we explain the seemingly contradictory evidence of windows being both a deterrent and a vulnerable point of entry? The answer probably resides in the context of the overall situation target selection.381 while a non-residential burglary was $1. and perceptions by customers.another alarm expense. sleep. Adding to the effect of a residential burglary is the fact that this is where people eat." The psychological harm of a burglary to the occupants is severe and difficult to measure. Will the additional expense pay dividends in the end and can security-glazing materials compliment overall security strategies? Alarms systems. The Uniformed Crime Reports (UCR) in 2000 reported 2. the use of security glazing materials is not comnlon. Burglars do not like multiple impacts on glass fearing that the second and third blow will attract to much attention to their actions (Wright and Decker.049.605. with nearly two-thirds being residential in nature. The assessment of these variables along with other factors not relating to windows will invariably result in a 'desirability assessment' on the part of a burglar. the advantages of windows can by maintained (or even enhanced) while the disadvantages (ease of penetration) can be minimized. it was typically enough to deter most criminals from completing the crime. there have been relatively few incidents where burglars have encountered security-glazing materials during the commission of their crimes. At most of those crime scenes (anecdotally). A common concern is. Nation wide 9598% of all alarms are false (IACP: 2002). If the glazing material was able to withstand several impacts. keep valuables and raise their children. it was quite obvious the criminal wasn't prepared for the additional work required to penetrate such security glazing materials. If occupancy (or lack of) can be easily established through assessment of occupancy clues.946 burglaries nation wide. it is not uncommon for an alarm user to accumulate more in false alarm fees over one year than the average dollar loss per burglary (Austin. down time. The average reported dollar loss per residential burglary was $1. how much will the glazing improvements cost compared to their worth? Many homeowners and businesses perceive security glazing materials as a questionable investment. most municipalities have enacted false alarm ordinances that fine alarm users for every false activation . The most crucial downfall of an alarm system is.: 2002). Furthermore. the windows of the intended target are vulnerable to attack while the windows of overlooking properties and other capable guardianship cues would act as a deterrent. "Home" is where people generally feel "safe.

How are surveillance opportunities to be maximized if security-glazing materials are not used? Criminals and crime prevention practitioners clearly recognize the need to increase surveillance opportunities to develop a comprehensive security strategy. 1968:29). Appropriate use of glazing materials based on threat and risk assessment is called for. Once the glass is attacked. For example. alarms etc. This impact must be considered in the development of an overall security strategy.). what good is the lighting with nearby buildings or houses without the ability of people to witness events? The purchases of closed circuit television (CCTV) systems have boomed in recent years.150 THE USE OF GLASS IN BUILDINGS after entry is achieved at which point the alarm is only as good as the response it generates. there may be a trend to minimize the use of glazing materials in buildings because of it's perceived vulnerability. Further. This will significantly enhance the overall security of a building (burglary prevention. The outcome may very well be that the question to ask is not . The burglar hears the alarm and knows the police are being notified. The effective use of glazing strategies includes the consideration of window location of both proprietary and adjoining facility windows. Lighting is primarily designed to enhance the visibility of an area. 2001. since the fallout of September 1lth. But. is to encourage residents to watch out for one another. planners and law enforcement need to consider the impact glazing materials have on the environment and it's impact on crime and perceptions of safety and security. emergency egress) while avoiding many of the signals of fear and urban decay. However. With a false activation rate in excess of 95%. But. triggering the alarm. Yet. to be effective. CPTED practitioners claim that with appropriate CPTED strategies. Glazing materials have an impact on criminal behavior. The increased penetration time makes apprehension more likely forcing the burglar to choose to flee while the going is good or assume additional risk that may not have been considered during the original target selection process. Architects. CCTV has its limitations and may not compare to the surveillance capabilities provided by the use of security glazing in homes and businesses. the alarm system can be a significant deterrent. 64). developers. the glass breakage sensor is activated. The premise behind Neighborhood Watch programs. supported by other security strategies (capable guardianship. quality of life can be improved (Crowe 1991:28-29). it is understandable why law enforcement does not respond to them as 'emergency' calls. evidence suggests that police need to respond to alarms within three minutes of the activation for apprehension (deterrence) to be probable and that such a response is unlikely (Berube 2001:56. what impact could this have on local crime trends? The concern of terrorist bombing attacks must be weighed along with the more prevalent crimes within a community. Using post alarm delay tactics effectively can increase deterrence while allowing for more reasonable police response times and thus provide a powerful deterrent to burglary (Berube. Small Business Administration. With the alarm activated. the burglar is confronted with a situation where they are probably on the exterior looking at glass that is still blocking their entry. With the use of an appropriate security glazing material. a business that invests in laminated glass and incorporates glass breakage detectors into their alarm system has a strategy where all components compliment one another. organized by law enforcement. Lighting is another security tool that could benefit from the expanded use of glazing materials.

Olson. Window and glazing products are an integral pan of that strategy. by working together in each of the related fields we will facilitate the development of these integrated strategies.United States Senate. (2001) An Examination of Alarm System Deterrence and Rational Choice Theory: The Need to lncrease Risk.BE'FTEN AND BERUBE ON ADVANTAGES OF GLAZING 151 "can we afford appropriate security glazing materials?" McKay T. "can we afford not to use security glazing materials?" Following the assumption that there will always be crime.. (1991). every home and business would incorporate a CPTED philosophy. Breaking and Entering. constantly monitored CCTV. D. Berube H. it is absolutely essential to take appropriate measures to minimize its effects. activate an alarm system and utilize securityglazing materials. KY. . some are obviously better than others and some can actually increase risk (double cylinder deadbolts/bars etc. (2000) Uniformed Crime Reports http://www. Architects. National Crime Prevention Institute. University of Leicester. Ideally. (1991). F. Crime Prevention Through environmental Design: Applications of Architectural Design and Space Management Concepts. J. Federal Bureau of Investigations. W. (2001) Overland Park (KS) Police Department False Alarm Statistics.fbi.).. Small Business Administration (1969). Unfortunately there is a perception that by applying only one of these security measures you can dramatically increase security. The question shouldn't be "can we afford appropriate security glazing materials. D. none are fool-proof. Security Architecture: The Right Designfor Reducing Crime.K. Leicester. MSc Thesis. Crowe T. but can we afford not to?" References Austin. (1997). Cromwell. London. U. developers.An Ethnographic Analysis of Burglary. And while there are positives to any crime-prevention approach. utilizing proper lighting techniques. P. December. Louisville. Crimes Against Small Business: A report transmitted to the Select Committee on Small Business . That said. and Avary. Security Management Magazine. England. Sage Publications. D. planners and law enforcement need to consider the impact glazing materials have on the environment and it's impact on crime and perceptions of safety and security. N. The reality is that each one of these measures is part of an integral strategy that needs to be catered to individual sites based on a threat assessment.

Virginia. http://www. 1994. .toc.. Boston.152 THE USE OF GLASS IN BUILDINGS International Association of Chiefs of Police (1993) False Alarm Perspectives. MA.htm Wright R. Northeastern University Press. Alexandria. and Residential Break. Burglars On The Job . USA.theiacp. and Decker S.

Jerry Razwick 1

The Relationship Between Sprinkler Systems and Glass

Reference: Razwick, J., "The Relationship Between Sprinkler Systems and Glass,"
Use of Glass in Buildings, ASTM STP 1434, V. Block, Ed., ASTM International, West

Conshohocken, PA, 2002.

Abstract: This paper explores the particular relationship between sprinklers and glass in
a building. There are many misconceptions about sprinklers' ability to prevent non-rated glass from breaking during a fire, when in fact sprinklers can actually cause some glass to break. Laboratory testing has shown that unless the conditions can be strictly controlled, sprinklers and non-rated glass are not an adequate combination to prevent the spread of a fire. Specific testing examples and illustrations are cited in the paper. Sprinklers are also "active" systems, depending on proper maintenance and activation sequences in order to work properly, whereas fire-rated glass is a "passive" system providing compartmentation around the clock.

Keywords: sprinkler systems, deluge, fire-rated, glass, thermal shock

Since the first automobile air bag was introduced in 1973, there have been many advances in the technology for car safety. Today, air bags are standard equipment in nearly all cars being manufactured in the United States, and the track record continues to grow for how effective the devices are in saving lives. Imagine for a moment that auto manufacturers became so enamored with air bags that they decided to abandon other safety features. Would you want to buy a car that had great air bags but a weak frame? Would you believe strongly enough in that one safety device to give up your anti-lock brakes? Probably not. And hopefully, you would not even consider a car that substituted air bags for safety belts - particularly since air bags without safety belts can cause serious injury when inflated during an accident. Such scenarios sound absurd in the context of automobile safety. But in the construction arena, something similar is happening on a regular basis. Sprinkler systems have become so popular as fire safety devices that in many cases architects are abandoning other important fire protection systems. Code officials are allowing "tradeoffs" - approving the use of sprinklers where the code would normally require fire-rated building materials.

1president, Technical Glass Products, 2425 Carillon Pt., Kirkland, WA 98033.


by ASTMInternational



Is this a safe practice? If sprinklers can offer equivalent levels of fire protection, it would be logical to conclude that other fire safety measures might be unnecessary. Given sprinklers' track record of saving lives and property, this concusion seems logical. However, closer examination of this position reveals that it is not as cut and dried as it may appear on the surface. There are many reasons to be skeptical of putting all the eggs in one basket when it comes to something as critical as fire and life safety. There are three basic components to a comprehensive fire protection program: Detection, Suppression and Compartmentation. The first two categories require activation, while the third category is generally passive. For instance, smoke alarms (which provide detection) and sprinklers (which provide suppression) must first be triggered in order to be effective. In contrast, fire walls, doors and ceilings compartmentalize and contain smoke and flames without any activation process. They offer around-the-clock protection in their passive state by acting as a physical barrier to fire and smoke. When active systems become the primary mode of fire protection, there is always the danger of mechanical failure, human error or poor maintenance interfering with the way the systems function. As seen with the recent massive recalls, sprinklers are no exception. The NFPA Journal has cited numerous additional causes that have rendered sprinklers inoperable in real world fires -- valves painted over, systems shut down during construction, fire burning through PVC supply pipe, fire fighters diverting water, etc. [I]. The point is not to negate the value of sprinklers, but instead to recognize their limitations. Relying exclusively on a single method of fire protection may create a false sense of security that is unwarranted. This becomes even more apparent when you examine the relationship between sprinklers and fire-rated building materials. For instance, misunderstandings abound regarding the interaction between sprinkler systems and glass. Deluge systems and non-rated glass are sometimes recommended as a combination instead of fire-rated glass and traditional sprinklers. Yet such configurations can pose a hazard. The problem centers around an issue known as thermal shock. Most glass cannot tolerate drastic variations in temperature on its surface. When one area is hot and another cooler, the glass doesn't know whether to expand or contract, and therefore it typically shatters and falls from the opening. If you have ever seen water sprayed on the glass doors of a fireplace when a fire is going, you may have seen this principle at work. Even without water involved, ordinary window glass breaks at about 250 ~ F, and tempered glass at about 500~ F. In contrast, fire-rated glass is often capable of withstanding temperatures above 1600 ~ F.



*...~'~ 1,500
mllmmm B mm B

1,000 m












Time(in minutes)

Figure 1 - Standard time.temperature curve for furnace testing of glass for fire rating.
1/4" (3 mm) *F *F Max *F Max

Thick Materials

Normal Service 1500

Thermal Shock 1400



Fire-Rated Glass Ceramic


Non-Rated Tempered Float Glass Non-Rated Float Glass







Figure 2 - Comparison of thermal properties for monolithic glass types.
When fire-rated glass is used, sprinklers pose no threat. However, when attempts are made to get by with non-rated glass, the situation becomes tricky. It is quite possible that the sprinklers could actually cause the glass to vacate the opening during a fire, leaving a breach for flames and smoke to spread.

The researchers discovered that the glass could survive the test if two conditions were met: First. Looking at the results of both tests. the fire had to start far away from the glazing assembly. Officials at Factory Mutual Research Corp. the sprinkler needed to activate very soon after the fire started. the glass fell out of the frame in less than 5 minutes. . the tests left too many questions unanswered to be considered definitive. Adjusting water flow during the test prolonged the endurance of the glass. and sprinklers were also installed for the test. outpacing the sprinklers' ability to cool the glass surface. when a smaller fire is concentrated close to the glass surface. They used two different sizes of fire (250 kW and 40 kW). the sprinklers activated early and the glass remained intact. A propane burner was ignited across the room from the tempered glass. Laboratory tests conducted over the last several years demonstrate the complexity of the sprinkler-and-glass relationship. When the larger fire was started. Any dry spots on the surface of hot glass cause heat stress and can be a primary cause of glass fracture. observed the test to determine if a non-fire-rated window assembly and sprinkler "system" could provide equal protection to that of a firerated assembly." However. The overall room temperature rises rapidly enough to activate sprinklers before the glass becomes "stressed.13 m (7 ft) from the glass surface. roughly 2. Lawrence Livermore National Laboratory (LLNL) conducted a similar experiment to find out what would happen when a fire started near the surface of non-fire-rated (tempered) glass [4]. Apparently. tempered glass may perform as needed. it may not activate sprinklers early enough. One such test was conducted in 1995 [3]. As in the tests mentioned previously.44 m (8 ft. Deluge sprinklers were installed on the bum side of the glass. the smoke will be free to escape into other areas of the building [2]. In a real-world fire. However. they can generate large volumes of deadly smoke. Second.156 THE USE OF GLASS IN BUILDINGS This is a crucial issue. But in both tests conducted with the smaller fire.). Nearly a decade prior to that test. generating sufficient stress to shatter the nonrated glass. When the heat source was brought closer to the non-fire-rated glass.even before the sprinklers activated. the glass fractured and fell out of the test assemblies in less than 4 minutes . we can conclude that when fires are relatively large and distant. One commonly proposed method of dealing with this problem is to utilize special deluge sprinklers that will bathe the glass with water before such stress can occur. officials adjusted the water flow rate from the sprinklers during the test when "dry spots" appeared on the fire-exposed face of the glass. The system (which combined specially designed sprinklers and tempered or heat-strengthened glass) was exposed to fire. This was the concept scrutinized in the "Hospital for Sick Children" test conducted at Canada's National Fire Laboratory in Toronto [5]. the non-rated glass was able to survive the test under these conditions. in this case approximately 2. When sprinklers suppress a fire. and were carefully positioned to ensure uniform water coverage. If glass windows have shattered. no one would be on hand to monitor water flow. the close proximity of the flames caused the temperature of the glass to rise too quickly. non-rated. with the hope that the glass would be able to stay intact. For instance.

Recognizing this fact. NER516 requires that when using this system. Since the source of potential fires cannot always be identified ahead of time. These test limitations were noted in an article entitled "Fire Resistant Wall Assemblies with Glazing" in the Society of Fire Protection Engineers Bulletin. Recognizing the importance of heat source location. Summary Clearly. one sprinkler manufacturer states in its literature: "Blinds or curtains must not be between sprinkler and glass [9]." Once a facility is occupied. the National Evaluation Report No. once tenant improvements are made at a later date to a building. "all combustible materials shall be kept 2 in. glass failure could possibly occur [5]. in addition to restrictions on curtains and blinds [7]. experience with sprinklers.. the sprinklers activated early. sprinklers are not adequate in and of themselves . In two of the tests. file drawers. the sprinkler and tempered glass systems require deluge sprinklers on both sides 0fthe glass. coat racks.Smm) from the front face of the glass [8]. Again. the commonly seen window designs with intermediate horizontal mullions are not allowed.. thus causing the glass to fail early on during a fire. they recommended construction of a 36-inch-high pony wall. Curtains. initial advisories regarding window coverings are likely to-be ignored. this presents an awkward challenge.they must be part of a more ... The sprinkler system used was a "quick response" type that activated two to three times faster than standard sprinklers.large pieces of glass fell to the floor" after an average of just 4 minutes [6]. papers. Two additional conditions are worth mentioning as well." With sprinkler heads located up to 30 cm (12 in. the ledge created by a pony wall frequently becomes a convenient place to stack flammable books. When these materials are placed between the glass and sprinkler. since the mullions would interfere with the water's ability to evenly bathe the glass surface. Second.RAZWlCK ON SPRINKLER SYSTEMS AND GLASS 157 Sprinkler activation time was also critical to the success oftbe test. (50. The UL test report states that in all three cases. Should sprinkler activation be delayed so that the temperature of tempered glass is in the range of 250 ~ C (approximately 482 ~ F). This solution has its own problems. the water is unable to cool the glass. a pony wall does not prevent flammable objects such as desks. UL conducted four tests in which the heat source was placed close to the glass. Also. The glass failed three out of four times.. First. Yet as the National Fire Protection Association (NFPA) acknowledges in their evaluation of U. To prevent this. etc. To accomplish this. enforcement of such instructions becomes unworkable. "The location and response time of the sprinkler must be such that activation will occur before the glazing reaches critical temperature levels. but the thermal shock proved too much for the glass. causing it to break." Underwriters Laboratories (UL) tested one manufacturer's deluge type "system" that combined specially designed sprinklers and non-fire-rated glass. They concluded that the window sprinkler system would work as long as flammable materials could be kept away from the glass surface. The Canadian Construction Materials Center reviewed UL's test results. etc. sprinklers have improved the standard of fire safety. blinds or other window coverings can also affect the performance of sprinklers. from being placed near the glass.) away from the windows. ".S. however.

0Z5Q2. 80. Report No." National Fire Protection Association. May 1995. [8] National Evaluation Report for Central Window SprinklerTM Model WSTM. J. D. K and Boehrner. G. Project 94NK27353. [2] Mawhinney. URL: http://www. D. [3] Test report J. Quincy. NFPA International. R. T. April 2001.I. References [1] Journal Subject List. Care & Maintenance Figure G. There are ample glass products on the market today that offer outstanding fire-ratings and the ability to withstand thermal shock.AM (4510). July 1986. "Effect of Automatic Sprinkler Protection on Smoke Control Systems. there is no reason to compromise.158 THE USE OF GLASS IN BUILDINGS comprehensive fire protection program [10]. p. G. "U. 87-3. PA. August 1995. CCMC 12752-R. 6-2. and Tamura. No. 6." SFPE Bulletin No. MA. "Fire Resistant Wall Assemblies with Glazing. From a life safety standpoint." ASHRAE Transactions: Research 3785 (RP-677). pp.S. NER-516." Fire Journal. Society ofF[re Protection Engineers. Experience with Sprinklers. [6] File Ex683. Model W S TM Specific Application Window SprinklersTM Data Sheet No.M A February2002. Factory Mutual Research. [10] Rohr. July 1987. "Fire Endurance of Sprinklered Glass Walls. Northbrook Illinois. Vol. 43-45. 4. . National Evaluation Service. Combining sprinklers with fire-rated glazing offers the better solution. 2001. 50.orl~. Lansdale.nfpa. March 2001. VA. J. Falls Church. [7] Canadian Construction Materials Centre Evaluation Report. Quincy Mass. Norwood. NFPA Journal.. D. p. Tyco Fire Products. [5] Richardson. 1996. [4] Beason.0. Underwriters Laboratories. K. [9] Central Sprinkler. June 24. J.

Box 41023. Lubbock. J.H. and Conrath. E. laminated glass. TX 79409. The procedure finds its basis in an empirical relationship between air blast pressure. U. 2002. Ed. blast-resistant glazing should maintain closure of its fenestration. "Design Procedure for Blast-Resistant Laminated Glass. Relatively small explosions can cause significant window glass breakage. Army Corps of Engineers. The procedure is relatively simple to use. and 60-second equivalent design loading for laminated glass and window glass constructions fabricated with laminated glass. Keywords: air blast pressure. under air blast pressure loading. Department of Civil Engineering. Block. H. or formal guidelines to aid in designing blast-resistant glazing. air blast pressure will fracture windows. Omaha. loss of balance. In addition. if any. However. requiting window glass replacement and substantial cleanup. necessitating replacement. Architects and engineers have . Scott Norville ~and Edward J. insulating glass. air blast pressure entering buildings can cause severe damage to ears that can result in diminished hearing ability. significantly reducing air blast pressure-related injuries and cleanup costs. and headaches [3]. Even with blast-resistant glazing. Center Road. blast-resistant z Professor and Director. Texas Tech University. the shards flying and failing from fractured window glass injure and kill persons [1-6]. 159 Copyright9 by ASTM International www. blast resistance The Purpose of Blast-Resistant Glazing Design When explosions occur in populated areas. At the same time. Protective Design Center. Blast-resistant glazing should minimize and possibly eliminate flying and falling glass shards in any explosion. publicly available tools. 2 Structural Engineer. building owners should strongly consider using blast-resistant glazing in windows and curtain walls. Abstract: When an explosive threat exists. ASTM International. Nebraska 68144-3869. S. It gives a design example and shows comparisons with experimental results. procedures. positive phase impulse. glazing design.S.. V. Glass Research and Testing Laboratory. 12565 W. STP 1434. air blast pressure typically fractures windows. C o u r a t h 2 Design Procedure for Blast-Resistant Laminated Glass Reference: Norville.." The Use of Glass in Buildings. The paper also addresses framing considerations for the blast-resistant glazing. even in the absence of building collapse. This paper presents a design procedure for blast-resistant laminated glass. In the worst scenarios. PA.astm. West Conshobocken. causing catastrophic results.

In the vast majority of its applications. They design and install blast-resistant glazing based upon this predicted design threat. blast-resistant glazing must not contribute to the hazards associated with the blast. Under this theory. i. Explosive threats in today's uncertain world require architects and engineers to produce designs that afford protection from air blast pressure.e. any breakage occurring in a window glass lite. of a window glass lite. The more probable scenario for blast-resistant glazing arises when no explosion occurs during its in-service life. providing a barrier between environments inside and outside a building while allowing light to enter the building and building occupants to observe the outside world..e. in terms of the . They can define the potential blast threat in terms of (1) an amount of explosive and (2) a standoff distance from a building. In this case. Standard Practice for Determining Load Resistance of Glass in Buildings (E 130000) uses the failure prediction methodology as a basis for design procedures. construction. Consequently. and wind. any uniform loading having finite time duration that acts on an annealed window glass lite induces a non-zero probability of breakage. Within traditional window glass design. its own weight. window glass must resist loading from snow. without maintenance beyond that which standard glazing requires. constitutes failure. and its own weight. strength. To accomplish this. window glass must usually resist only wind loading. Blast-resistant glazing accomplishes this by remaining in its frame following fracture and eliminating or greatly reducing the number and sizes of flying and falling glass shards. Blast-resistant glazing must perform these everyday glazing functions economically. snow. blast-resistant glazing must economically perform the functions of standard glazing.e. properly designed blast-resistant glazing should (1) minimize flying and falling glass shards and their associated lacerative hazard and (2) maintain closure of most of the glazed openings. the blast-resistant glazing must perform the functions of standard glazing.. blast-resistant glazing will never experience loading from air blast pressure. i. It relates a uniform load having constant magnitude over specified time duration to a probability of breakage. and thickness designation(s) to resist uniform loads from wind. they must make a prediction of a potential blast threat to a building. Consequently. To afford this protection. This document defines the load resistance. For design of sloped glazing. Should an explosion larger than the design threat occur. the primary goal of blast-resistant glazing design consists of protecting people inside or near a building subjected to an air blast pressure loading.160 THE USE OF GLASS IN BUILDINGS glazing should remain in its openings.. as appropriate. In summary. window glass design consists of determining the appropriate window glass type. a crack or fracture. i. Window Glass Design and Blast-Resistant Glazing Design For design of vertical glazing. The failure prediction methodology addresses all factors known to affect window glass strength [8]. The failure prediction methodology [ 7-9] provides the theoretical model that describes load resistance of window glass for US design procedures. thus reducing the urgency for immediate replacement. provided it does not cause building collapse. Designers assume these loads act in a quasi-static manner.

PB = 0. in 12 charts one for each nominal thickness designation. (1/4 in.-. ._.. .-.25/I tE .) j ~ ' " ' ' ' ' ' " '~' " ' ' ~ /~/"3000 J . Plate Length(in. . termed nonfactored load... Figure 2 indicates ."" vE 2000 =. Glass' . Currently. 9 psf / l J o. . . . Current window glass theory does not precisely define the relationship between loading. the probability of breakage for heat treated monolithic window glass and window glass constructions with any glass types has nominal values less than or equal to I . E 1300-00 uses glass type factors that relate the strength of other monolithic window glass types (heat strengthened and fully tempered) and constructions (insulating glass and laminated glass) to that of monolithic annealed window glass. . at its first occurrence.NORVILLE AND CONRATH ON BLAST-RESISTANT GLASS 161 magnitude of the uniform loading which acts over a time duration of 60-seconds to produce a nominal probability of breakage of eight lites per thousand. ' ' ' ' ' ' .. air blast pressure loads window glass lites dynamically over very short time durations. . 40 i i N.008.q 80 60 I1. Traditional window glass design methodology assumes that loads act quasi-statically with durations measured in seconds or longer periods. . E 1300-00 presents load resistance for annealed glass. similar to that contained in E 1300-00. lOOO 20 o o I ooo 2000 3000 4000 5000 P l a t e L e n g t h (rnm) Figure 1-NonfactoredLoad Chart. mm . . ' ' ' '/" 120 Nonfactored Load (kPa) ~Pb = 0.008 at the first occurrence of the design loading. When designed using E 1300-00. time duration.-L. / i /" / I . ~// . .. When an explosion occurs. . Figure 2 shows the approximate relationship between stress duration and stress magnitude at which fracture occurs for annealed window glass [13]. US model building codes [10-12] have adopted the E 130000 methodology in some form to facilitate design to resist uniform loading.)'.oox .008 1 kPa = 2 0 . Figure 1 shows a nonfactored load chart for window glass having nominal 6-mm (1/4 in.) 0 140 20 40 60 80 100 120 140 160 180 200 ' ' ' ' ' 8 . 0 .I % 100 v o. and probabilities of breakage for window glass constructions and heat treated monolithic window glass.

the probability of breakage for typical monolithic window glass lites or window glass constructions approaches 1. the distribution and severity of the load-induced tensile stresses in a window glass lite subjected to loading from air blast pressure typically overcome any increase in resistance resulting from the relatively short duration of the loading. Because of the excitation of higher modes. 9 I == == m 10. In addition to their dynamic nature. the stress distribution for a dynamically loaded window glass lite differs significantly from the stress distribution under quasi-static loading of the same magnitude in that stresses having high magnitudes occur over large regions of a window glass lite [14].000 1 second 1 hour 1 day 1 week 1 month Duration of Stress Figure 2 - ApproximateRelationship Between Magnitude of Fracture Stress and Stress Duration. In short.162 THE USE OF GLASS IN BUILDINGS that under short duration loading the stress at which fracture is initiated. air blast pressure loadings tend to have much larger magnitudes than wind and snow loadings that typically govern window glass design. increases dramatically. the dynamic air blast pressure loading associated with an explosion excites higher mode shapes in a window glass lite causing much larger deflections and stresses than would a quasi-static loading having the same magnitude of pressure.000 5. Incorporating these factors. Many factors .0 even for relatively small air blast pressure loading [14]. Some designers attempt to devise "strong" monolithic window glass lites or window glass constructions to resist a design air blast pressure without fracture. the failure prediction methodology indicates that under air blast loadings. On the other hand. which somehow correlates with window glass load resistance.

In properly designed blast-resistant laminated glass. laminated glass makes an excellent blast-resistant glazing material. the interlayer material should not tear under loading from air blast pressure. laminated glass should remain in its frame after fracture. PET can be scratched easily although it does not degrade from ultraviolet exposure in these constructions nearly as rapidly as it does in retrofit window film applications. Insulating glass consists of two window glass lites with a sealed air space between them. the majority of glass shards adhere to the PVB interlayer. no matter how small. any window glass construction designed to withstand even an air blast pressure loading of low magnitude would be very thick and would most probably involve heat treated window glass. Furthermore. As its name implies. In addition. it would have a finite probability of breakage under air blast pressure loads. Laminators use polyvinyl butyral (PVB) most commonly as the interlayer material in fabricating laminated glass. Small shards may spall from laminated glass in a blast. or a combination or window glass types. a single laminated glass lite can consist of glass plies having different thicknesses. Such window glass constructions would have prohibitive costs. Also. Its post breakage behavior characteristics make laminated glass an excellent blastresistant glazing. the designer should find this approach desirable when considering the effect of air blast pressure on an entire building. insulating glass provides thermal insulation far superior to that of single . If the blast-resistant glazing must completely hold all shards. especially if the PVB tears. Because window glass constructions that fracture transfer much less load into the structural frame. This load transfer would require a frame design that could resist such loading without collapse. For these reasons.NORVlLLE AND CONRATH ON BLAST-RESISTANT GLASS 163 tend to make this approach undesirable. When laminated glass fractures. Following fracture. fully tempered. The authors recommend that blast-resistant glazing constructions using glass should fracture under air blast pressure loading. Blast-resistant glazing that performs these functions will minimize damage to building contents and maximize safety to building inhabitants. heat strengthened. The two window glass lites can be monolithic glass or laminated glass. if the design employs monolithic window glass. a window glass construction with sufficient strength to resist without fracture an air blast pressure loading would transfer a large portion of the loading into the structural frame. Manufacturers provide commercially available laminated glass constructions with a layer ofpoly(ethylene terephthalate) (PET) laminated to the inside glass ply using PVB. First. regardless of its load resistance. When monolithic glass fractures under air blast pressure dire consequences ensue. Finally. The designer must base blast-resistant glazing designs on maintaining closure of fenestrations and eliminating flying and falling glass shards to the greatest extent possible. The designer should strive to ensure that blast-resistant laminated glass behaves in this manner. Aesthetically. they should rely on post breakage behavior characteristics to eliminate flying and falling glass shards and maintain closure of fenestrations. annealed plies make the best. distortion-free laminated glass. Laminated Glass and Insulating Glass in Blast-Resistant Glazing Laminated glass consists of two or more plies of monolithic glass bonded together using an elastomeric interlayer. the designer should consider using a plastic film over the inner glass ply. The glass plies can consist of annealed.

and anchorage of the window frame to the structural system.164 THE USE OF GLASS IN BUILDINGS window glass lites. In other words. the attachment of the window glass construction to the framing. Figure 3-Insulating Glass with Outside Lite Fractured in the Oklahoma City Bombing. by the time the window has fractured the blast pressure is long gone. . small amounts of explosives produce blast waves that will have passed the windows before they fracture. The authors feel that insulating glass fabricated using two laminated glass lites provides one of the most economical and effective blast-resistant glazing constructions available. Design Recommendation for Blast-Resistant Glazing In designing a blast-resistant glazing system. insulating glass provides excellent sound insulation. The authors observe that most blasts in the US come from relatively small quantities of conventional explosives. the framing system. When fabricated using laminated glass. et al [2] observed that insulating glass fabricated with two lites of monolithic glass could provide some minimal additional protection over monolithic glass under air blast pressure loading in the Oklahoma City bombing (Figure 3). Norville. The positive phase duration of the air blast pressure occurs on the order of milliseconds. the architect or engineer must consider four factors: the windowglass type and construction. In window glass design discussions encompassed herein.

00 . 175 "o 150 O~ o "J 125 . magnitude of positive phase impulse. Designers should use this chart to obtain 60-second duration equivalent design loads for laminated glass and insulating glass fabricated using laminated glass. The designer enters the chart in Figure 4 by projecting a vertical line from the horizontal axes that represent standoff distance to the sloping line that represents the charge weight.(X~ ! e2. the designer defines the design threat explosion in terms of an equivalent weight of a hemispherical TNT charge and a standoffdismnce._o e.00 ~ 3.00 ~" 8..NORVILLE AND CONRATH ON BLAST-RESISTANTGLASS 165 Window Glass Type and Construction The authors feel that laminated glass and insulating glass fabricated with laminated glass comprise the most effective and economical blast-resistant glazing materials. The authors offer a simple chart (Figure 4) that provides an empirical relationship between the weight of a hemispherical TNT charge detonated on the ground surface and its standoff distance from a window glass lite to a 60-second duration static design load.. The values in the chart do not apply to any monolithic glass type. or to monolithic glass with retrofit window film. the authors observe that designers using this approach will obtain laminated glass thickness designations equal to or slightly higher than they would obtain using more esoteric procedures.70 80 45 I 50 -i ~ = 0 '/ t 4. 7. In developing this chart the authors considered magnitude of reflected air blast pressure.00 "o 6. insulating glass fabricated with monolithic glass lites only.- 7 8910 \ 20 ~ 1 30 40 50 60 70 80 100 120 I I I I I! 9.00 ~ 1. From the intersection of the vertical projection and the sloping line.00 ~ . the designer projects a .. In comparing designs obtained using values from this chart..50 ~ 40 50 75 100 125 150 200 250 300 400 8 4O 30 20 Figure 4- 3O Standoff Distance (ft) Chart Relating Equivalent TArTCharge Size and Standoff Distance to 60Second Equivalent Load. To use the chart. and experimental results [15] from blast tests involving laminated glass and insulating glass fabricated using laminated glass [16].o 90 t. Standoff Distance (m) 200 ~.00 ~ 5.m Or) 100 .

The authors note that blast tests on curtain walls indicate that more flexible frames supporting window glass constructions tend to provide better glass performance under blast loading [17]. helping to maintain it in its frame.) PVB will usually suffice for smaller explosions.58 mm (0.5 acted on the window glass lite [16]. i. but less than 231 kg (500 lb). although this might prove overly conservative. or fully tempered) and to determine the required thickness designation(s) to resist the 60second duration equivalent design load. The designer can also use dynamic analysis techniques with a design blast loading to determine the loading that the lite would transfer to the frame if it never fractured. Thicker PVB will result in larger forces that the window frame must accommodate the laminated glass plies should be annealed or heat strengthened glass. Attachment of Window Glass Construction to Framing For blast-resistant glazing. The larger shards adhere well to the PVB thus reducing flying and falling glass shards and giving the laminated glass some stiffness following fracture.030 in. The chart in Figure 4 coupled with standard window giass design procedures provides a means to arrive at a laminated glass design suitable to resist the specified design threat explosion. For charge sizes other than those shown. although 0. annealed and heat strengthened glass plies tend to produce larger shards than do fully tempered glass plies. the designer should avoid "dry glazing.762 mm (0. The authors interpret this as the frame loading that would result ifa 60-second equivalent loading associated with a probability of breakage of 0. the designer can interpolate between the sloping lines. Upon fracture. the designer should design the window glass framing system and its anchorage to the surrounding structure to resist the maximum loading that the window glass would transfer to its supporting frame at fracture. The designer then uses procedures in E 1300-00 to select the laminated window glass type(s) (annealed. then the designer should use wind load to design the laminated glass or insulating glass fabricated using laminated glass. the 60-second equivalent design loading associated with a given bomb goes down rapidly with increasing standoff distance. The PVB in laminated glass should have 1. Laminated glass and insulating glass fabricated with laminated glass and designed using values from this chart will more than satisfy GSA Level 3B Criteria in that it fractures "safely. If blast-resistant glazing is designed using values obtained from this chart.166 THE USE OF GLASS IN BUILDINGS horizontal line to the vertical axes and reads the equivalent 60-second duration equivalent design load. Procedures in E 1300-00 present load resistance values based on maximum nominal frame deflections of L/175..060 in.e. If the wind load for a given design exceeds the magnitude of the 60-second duration static design loading determined from the chart in Figure 4. heat strengthened. The chart in Figure 4 also indicates that the best protection from a bomb is standoff distance. Framing System Under air blast pressure loading. glass will fracture and require replacement. Standard glazing bites with gaskets .) thickness in blastresistant glazing [16]." producing minimum hazard." in which gaskets alone hold the blast-resistant glazing in its frame.

Comparison with an Experiment Figure 5 shows a nominal 6 mm thick laminated glass test specimen with rectangular dimensions of 1190 x 1640 mm after being subjected to a blast loading that corresponds to the design loading in this example. the designer projects a horizontal line to the vertical axis. For example.76 mm thick PVB interlayer. For this size opening the load resistance of a 10 mm laminated glass lite is 2. The use of very deep bites with gaskets might restrain the blast-resistant glazing but could lead to other problems. The laminated glass specimen has a 0. where the width is measured parallel to the plane of the glass. to a point corresponding to the 25 kg. the authors recommend using a PVB interlayer with 1. the authors reeornmend a 6 mm (1/4 in.39 kPa. determined by interpolation. Glazing tape has more flexibility than structural silicone and the designer should use a width of glazing tape 2 to 3 times the thickness of the blast-resistant glazing material with which it is in contact.) air space. This thickness will usually be less than the thickness of the entire blast-resistant glazing construction. The bite depth should not exceed standard depths any more than necessary to facilitate the width of the structural silicone bead or the glazing tape. producing minimal hazard. When using structural silicone sealant. Blast-resistant glazing should attach to the frame using either structural silicone sealant or adhesive glazing tape. the width of the bead forming the structural connection should equal the thickness designation of the blast-resistant glazing material with which it is in contact. As mentioned above.05 kPa.52 mm thickness and using a 10 mm structural silicone bead to attach the glass to the frame. Under the same blast conditions. The design approach presented herein would require a larger thickness designation.NORVILLE AND CONRATH ON BLAST-RESISTANTGLASS 167 will not restrain fractured laminated glass under air blast pressure loading and the entire lite could fly from the frame. This intersection indicates that the lite should resist a 60-second equivalent design loading should of approximately 2. a laminated glass lite designed with this procedure would have broken in a similar manner. Clearly the specimen fractured in a safe manner. this width should result in some tearing of the silicone bead before the PVB interlayer tears but should maintain the laminated glass in its frame. This mode of failure will tend to eliminate flying and falling glass shards while maintaining the blast-resistant glazing in its frame. In the event of an explosion. the standoff distance. if the blast-resistant glazing construction consists of an insulating glass unit with two nominal 6 mm (1/4 in. . From this point. Design Example Compared With Test Result Design Procedure Consider selecting the thickness of a single laminated glass lite required to glaze a fenestration having rectangular dimensions of 1190 x 1640 mm. the designer projects a line down from 25 meters. Going through procedures in E 1300 -00 to design the laminated glass gives a thickness designation of 10 mm.) structural silicone bead.) lites and a 12 mm (1/2 in. especially insulating glass units. To use the chart in Figure 4. The design blast loading results from a 25 kg TNT charge situated 25 m from the window.

positive phase impulse. and 60-second equivalent loading. Use of this approach results in laminated glass designs . They note that in formulating the empirical relationship between air blast pressure. the authors looked at results f~om numerous blast tests with laminated glass and laminated glass constructions.168 THE USE OF GLASS IN BUILDINGS Figure 5-Specimenfrom Blast Test. Conclusions The authors present this design approach to give guidance to the window glass and engineering community on how to size laminated glass and laminated window glass constructions to resist blast loadings.

Vol. Italy. 22(3). Shariat. L." Bulletin of the American Ceramic Society. Revised.. '" Glass Research and Testing Laboratory. Lubbock. "Misty Picture Data: Window Glass Experiment. L.." Journal of Wind Engineering and Industrial Aerodynamics. 1982. H." American Journal of Surgery.. laminated glass designed according to this approach will still provide a significant level of protection. 1999. Conrath. and Blocker. "The Texas City Disaster: a Survey of 3000 Casualties. W. 1995. V. .. [4] Blocker. E. W. 1980: an Analysis of the Effects and Injuries Sustained.. L. "Journal of Performance of Constructed Facilities. W.. 1996. TX. L. L. Lubbock. Texas Tech University. G.. S. Lubbock. Lubbock.NORVILLE AND CONRATH ON BLAST-RESISTANTGLASS 169 that will produce minimal hazard in the occurrence of an air blast pressure loading of design size or smaller. [2] Norville. "A Failure Prediction Model for Window Glass. [3] Norville. Texas Tech University. 64(11). "Survey of Window Glass Broken by the Oklahoma City Bomb on April 19. 1990. J. 1949. "Glass-Related Injuries in the Oklahoma City Bombing. Smith.." Falls Church. E.. S.. "NTIS Accession Number PB81-148421. L. [7] Beason... H. For explosions with higher than design air blast pressures or positive impulses. [6] GRTL. S. 215-220. S. and Bregenwald. H. 36. "International Building Code. M. [9] Beason. Texas Tech University. Texas Tech University.pp.. W. S. TX. 756-771."Development of a New Glass Thickness Selection Procedure." Birmingham Alabama. "Standard Building Code. N. pp. Jr. Virginia. 1995. 13(2). and Norville.. 1987. 1995. Swofford. References [1] Norville. [5] Brismar B. Mallonee. Elsevier Science Publishers. [11] Southern Building Code Congress International (SBCCI). H.. 2000. Institute for Disaster Research. pp. J. J. "The Strength of Weathered Window Glass. TX. 1999. 1980. American Society of Civil Engineers.. 78. "Survey of Window Glass Broken by the Oklahoma City Bomb on April 19. [10] International Code Council (ICC)." Glass Research and Testing Laboratory. K. and King.. T. S. Smith. K." Journal of Trauma. pp. M. Harvill. H. "The Terrorist Bomb Explosion in Bologna. S.. [8] Norville.. and Minor.1135-1144. 1984. TX. Final Data Report. and King." Glass Research and Testing Laboratory. 1467-1470.

H. Lubbock.. and Lawver. . Test Planning. No. Vol. S.. [16] Norville. Omaha." Glass Research and Testing Laboratory. 1999. New York." Protective Design Center.. [13] Minor. "Basic Glass Strength Factors. Blast Prediction. Weidlinger Associates. and Post-Shot Analysis. D. S. J. pp... Rubin. 69(9). 3. 1999. 215 N. Inc. [17] Smilowitz. 80-86. J. [14] Norville. Termant. 2001. "Las Vegas Courthouse Curtain Wall Project. "The BOCA National Building Code/1999. E. "Window Design and Analysis Software... 2000. NY. "Considerations for Blast Resistant Glazing Design. American Society of Civil Engineers. H. 17th Street. "Dynamic Failure Prediction for Annealed Window Glass Lites. 7. 1990 [ 15] WinDAS. TX. WANY 99-01." Country Club Hills. and Conrath.." Report No. Inc." Journal of Architectural Engineering. E. D. (BOCA). Omaha District Corps of Engineers. Illinois. December. 1990. R. Texas Tech University. NE. D..170 THE USE OF GLASS IN BUILDINGS [12] Building Officials and Code Administrators International.." Glass Digest.

W.. . Ryan. Thomas. Iva~ 3 Razwick. Henri. 49 Smith. 57 Johnson. 79. C. Mostafa M. Jerry. Roger E. 105 M Major. George. Christopher J. 159 O O'Day. 8 C P Conrath. 131 Lichtenberger.. H. John E. 57 Berube. Jeffrey C. Anthony. 35 V Van Duser.. 79 Jagota. Edward J.. 66. Michael... 121 Bennison. Ted W. 131 Redner. 35 Kaskel. Allan. Alex. 35 Lingnell. 20 Barry.. 3 Morse. 121 Shah. 121 R E1-Shami. 131 Sciaudone. George R. 66. Bruce S. 90 N Norville. Paul E. 26 S Jackson. Stephen M. 131 Pelletier. William. Daniel J. 147 Betten.. Werner. 79 K Schmidt. 57 W Wise. 159 E Pearson. Valerie L.. 79 H Hennings. 20 Beason.. Alex S. Lynn... 2002 Author Index B Mazula.STP1434-EB/Dec. Mark A... Scott. Anand. 105 Beers.astm. 57 T Torok. Bipin V.. 147 Block. Stephen J... 131 Pilcher. Mark K. 49 171 Copyright9 by ASTMInternational www..

121 Insulating glass. 66. security. 8. 159 Airport control towers. 20 Hydrophobic. 121 Hurricane resistance. 3 Glass storage. 3 Glass lites. 121. 57 173 . 131 thermal stress evaluation. 105 Fire-rated glass. 57. 90 Dew-point measurement. 131 ASTM E 1996. 49 P Photocatalytic. 79 Anchored-film glass retention systems. 105 ASTM C 1036. 90 ASTM E 1886. 49 NFRC 500 Procedure. 90 Computer technology. 66. 3 ASTM C 1048.STP1434-EB/Dec. 90 blast resistance. 131 Annealed glass. 8 Damaged glass. 20 Impact resistant glass. 131 Glass inspection. 121 D Heat-strengthened glass. 90 Condensation resistance rating. 8 Glass retention. 20 Polyvinyl butyral interlayers. 121 blast-resistant. 3 Glass strength. 35 E Laminated glass. 147 C Coatings. breakage during fire. 79 Glass quality. 57 Glazing. 49 Creep resistance. 35 blast resistance. 79 H Blast resistance. 66 Longevity. 79. 131 B G Glass. 20 Computer software. 159 Load resistance. 2002 Subject Index A Air blast pressure. 8 Burglary. 26 ASTM C 1279. 131 Glass specification. 79 Fallout protection. 35 M Missile impact test. 131 Hydrophilic. 26 High winds. 121 N Edge strength failure prediction model. 131 Finite element analysis. 8 NFRC 100. 66 F Failure prediction. 131 National standards. 105 Edge stress. 147 Glazing design. 57 Cyclic pressure cycles. 66. 159 Interlayers. 79. 57 International standards. 131 Fenestration. 26 ASTM E 1300. 66. 159 Building codes. 79. 3 Design standards. 159 wind load resistance.

121 Temperature differential. 8 Scratched glass. 79 . 26 Tensile stress. 131 Temperature performance. 105 Three-sided support. 131 Standards Design Group. 131 Thermal stress cracking. 57 Storm shutters.1. 3 Security. 8 Sprinkler systems. 20 SentryGlas Plus. 121. 49 Wind resistance. 105 THERM 2. 79 measurement. 66 W Windbome debris. 90 Stiff interlayer. 57 Serivce life. 121 Safety glazing. 57 Tempered glass.174 THE USE OF GLASS IN BUILDINGS Q Quality control. 57 Stiffness.1.. 147 Self-cleaning glass. 131 WINDOW 4. 66. Inc. 13 i evaluation procedure. 3 T Retrofitting. 26 R Stress maximum. 26 Surface damage. 35 Skylights. hurricane resistant glass. 49 Thermal shock.