# ASME PTC 19.

1-2005

(Revision of ASME PTC 19.1-1998)

Test Uncertainty

--`,,``,````````,`,,```,`,,-`-`,,`,,`,`,,`---

A N A M E R I C A N N AT I O N A L STA N DA R D

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

--`,,``,````````,`,,```,`,,-`-`,,`,,`,`,,`---

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

Date of Issuance: October 13, 2006

The 2005 edition of ASME PTC 19.1 will be revised when the Society approves the issuance of the next edition. There will be no Addenda issued to ASME PTC 19.1-2005. ASME issues written replies to inquiries as code cases and interpretations of technical aspects of this document. Code cases and interpretations are published on the ASME website under the Committee Pages at http://www.asme.org/codes/ as they are issued.

ASME is the registered trademark of The American Society of Mechanical Engineers. This code or standard was developed under procedures accredited as meeting the criteria for American National Standards. The Standards Committee that approved the code or standard was balanced to assure that individuals from competent and concerned interests have had an opportunity to participate. The proposed code or standard was made available for public review and comment that provides an opportunity for additional public input from industry, academia, regulatory agencies, and the public-at-large. ASME does not “approve,” “rate,” or “endorse” any item, construction, proprietary device, or activity. ASME does not take any position with respect to the validity of any patent rights asserted in connection with any items mentioned in this document, and does not undertake to insure anyone utilizing a standard against liability for infringement of any applicable letters patent, nor assumes any such liability. Users of a code or standard are expressly advised that determination of the validity of any such patent rights, and the risk of infringement of such rights, is entirely their own responsibility. Participation by federal agency representative(s) or person(s) affiliated with industry is not to be interpreted as government or industry endorsement of this code or standard. ASME accepts responsibility for only those interpretations of this document issued in accordance with the established ASME procedures and policies, which preclude the issuance of interpretations by individuals.

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

--`,,``,````````,`,,```,`,,-`-`,,`,,`,`,,`---

No part of this document may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written permission of the publisher. The American Society of Mechanical Engineers Three Park Avenue, New York, NY 10016-5990 Copyright © 2006 by THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS All rights reserved Printed in U.S.A.

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

. . . . Defining the Measurement Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .CONTENTS
Notice . . . . . . . . . . . . . . . . . . . . . . . . . . . General . . . Combined Standard Uncertainty and Expanded Uncertainty of a Result . . . . . . . . . . . . . .`. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fossilization of Calibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .`. .. . . . . . . Comparative Versus Absolute Testing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 08/17/2010 06:03:02 MDT
. . . . Categorization of Uncertainties . . . . . . . Uncertainty of a Measurement . . . . . . . . . . . . . . . . . . . . .```. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pretest and Posttest Uncertainty Analyses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Random Standard Uncertainty of a Result . Examples of Uncertainty Propagation .`. . . Glossary. . . . . Measurement Uncertainty . . .. . . . . . . . . . . . . . . . . . . . . . . . . Random Standard Uncertainty of the Mean . . . . . . . . . . . . . . . . Harmonization With International Standards . . . . . . . . . . . . . . . . . . Object and Scope . . . . . . . . . . . . . . Selection of the Appropriate “True Value” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nomenclature and Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
iii
--`. . . . . . . . . . . . . . .. . . . . . . . Systematic Standard Uncertainty of a Result. . . . . . . . . . . Uncertainty of a Result . . . . . . . . Committee Roster . . . . . . . . . Correlated Systematic Standard Uncertainties. . . . . . Nomenclature . . . . Combined Standard and Expanded Uncertainty of a Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .`---
vii viii ix 1 1 1 1 2 2 2 3 3 3 5 5 5 5 11 13 13 13 13 15 16 17 17 18 19 19 22 22 23 23 24 24 24 28 28 31 35 36
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. . . . . . . . . . .``. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Spatial Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Systematic Standard Uncertainty of a Measurement . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .`. . . . Identification of Error Sources . . . . . . . . . . . . Section 1 1-1 1-2 1-3 Section 2 2-1 2-2 Section 3 3-1 3-2 Section 4 4-1 4-2 4-3 4-4 Section 5 5-1 5-2 5-3 5-4 5-5 Section 6 6-1 6-2 6-3 6-4 Section 7 7-1 7-2 7-3 7-4 7-5 7-6 Section 8 8-1 8-2 8-3 8-4 Introduction . . . . . . . . . . Assumptions . . . . . Overview . . . . . . . . . . . . . . . . . . Object. . . . . . . . . . . . Propagation of Measurement Uncertainties Into a Result. . . . . . . . . . . . . . . . . . . . .````````. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nonsymmetric Systematic Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fundamental Concepts . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Scope . . . . . . . . Sensitivity . . . . . . . . . . . . . . . . . . .`. . . . Classification of Uncertainty Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Measurement Error . . . . . . . . . . . . Foreword . . . . . . . . . . . . Additional Uncertainty Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .-`-`. . . . . . .

. General Considerations. . . . Periodic Comparative Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Generic Measurement Calibration Hierarchy . . . . . . . . . . . . . . . . . . . . . . Comparison of Test Results Using the Initial Field Test as the Control . . . . . . . . . . . . . . . . . . . . Pump Design Curve With Factory and Field Test Data Shown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1-3 10-5. .2-2 8-1
Analysis of Redundant Means . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Table of Data (Example 7-6. . . . . . . . . . . . . . . . .2 Tables 6-4-1 6-4-2 7-6. . . ؋ 4 in. . . . . . . . . . . . . .1 10-4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1) . . . . . . . . . . . . .2) . . . . . . . . . . . . Traverse Points (Example 10-1) . . . . . . Distribution of Measured Values (Normal Distribution) . . . . . . . . . . . . . . . . . . . . . . . . . .3 7-6. . .
iv
36 38 41 41 41 43 43 47 50 51 62 69 71 6 7 8 11 14 16 27 32 34 37 44 48 57 61 63 64 64 67
20 21 25 26 26 27 29
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. . . . . . . . . . . . . . . . Flow Rate Uncertainty Including Nonsymmetrical Systematic Standard Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . Schematic Relation Between Parameters Characterizing Nonsymmetric Uncertainty. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Uncertainty Interval . . . . . . . . . . . . . . . . . . . . . . .1) . Schematic of a 6 in. . . . . . . . . . . . . . . . . . . . . . . . . . . Step-by-Step Calculation Procedure. . 08/17/2010 06:03:02 MDT
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Venturi. Difference Between “Within” and “Between” Sources of Data Scatter . Typical Pressure and Temperature Locations for Compressor Efficiency Determination. . . . . . . . . . . Summary of Data (Example 7-6. . . . . . . . .3 5-3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relation Between Parameters Characterizing Nonsymmetric Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1-1 10-5. . . . . . . . . . . . . . . . . . . . . Bibliography . . . . . . . . . . .1 8-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Installed Arrangement . . . .1) . . . . . . . . . Summary of Data (Example 7-6. . References . . . . . . . Compressor Performance Uncertainty . . . . . Illustration of Measurement Errors . . . . . . Flow Rate Uncertainty . . . .1 4-3. .2 8-2. . . . . . . . . . . . . . . Calculation Procedure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Systematic Uncertainty of Average Circulating Water Bath Temperature Measurements (Example 6-4. . . . . .1 10-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Examples . . . . . . . . . . . . . Measurement Error Components .1) . . . . . . . . . . . . . . . . . . . .1 5-4. . . . . . . . . . . . . . . . . . Circulating Water Bath Temperature Measurements (Example 6-4. . . Comparison of Test Results With Independent Control Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2) . . . . . .2 8-5. . . . . . . . . . . . . . . Flow Measurement Using Pitot Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . .8-5 8-6 Section 9 9-1 9-2 Section 10 10-1 10-2 10-3 10-4 10-5 Section 11 Section 12 Figures 4-2-1 4-2-2 4-3. . . . Three Posttest Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 10-4.2-1 7-6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1-2 7-6. . . Regression Uncertainty . Burst Pressures (Example 8-1-1) . . . . Pareto Chart of Systematic and Random Uncertainty Component Contributions to Combined Standard Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 10-5. . . . . . The h-s Diagram of the Actual and Isentropic Processes of an Adiabatic Compressor. Table of Data (Example 7-6. . . . . . . . . . . . . . . .1 10-1. . . . .1-1 7-6. . . . . . . . . . . . . . . . . . . . . . . . . .1-2 10-5. .

-`-`. . .1. . . . . . . .3-1 10-1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Standard Deviations (Example 10-1) . .2-3
ˆ Determined Systematic Standard Uncertainty Components for Y from Regression Equation. . . . . . . . . . . Calculated Result . . . Elemental Random Standard Uncertainties Associated With Error Sources Identified in Para. . . . . . . . . . . . . .2-1 10-4. . .1-5 10-2. Relative Contributions of Uncertainties of Independent Parameters (Example 10-2: Calibrated Case) . . . . . . . .1-1 10-2. . . . . . . .`. . . . . . . . . . .5 9-2-1 9-2-2 10-1. . . . . Summary of Data . . . .. . Relative Uncertainty of Measurement (Example 10-2: Uncalibrated Case) . .2 .6 10-1. . . . . . . . . . . . . . . . . . . .`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. . . . . . Inlet and Exit Temperature Elemental Systematic Standard Uncertainties . . . . . . .1-3 10-2. . . . . . . . . . . . . . . . . . . . . . . .```. . . . . . . .1-1 10-4. . . . . . . . . Uncalibrated Case (Example 10-2) . . . . . . . . . . . . Summary of Results for Each Test .1-3 10-2. . . . . . . . . . Average Values (Example 10-1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .````````. . .`. . . . . . . . . . . . . . . . . . . .`. . . . . . . . . . . . . . . . . . . . . . . . .2-2 10-4. . . .1. . . . . . . . . . . . 08/17/2010 06:03:02 MDT
. . . . . . . . . . . . . . . . Summary of Test Results . . . . . . . Summary of Average Velocity Calculation (Example 10-1) . . . . . . . . . . . . .. . . . Absolute Sensitivity Coefficients in Example 10-2 (Calculated Numerically) . . . . . . . . . . . . . . . . . . . . . . . . . Absolute Sensitivity Coefficients in Example 10-2 (Calculated Analytically) . . . . . . . . . . . . . . . . . . . . . . . . .1-4 10-2. . . . . Absolute Contributions of Uncertainties of Independent Parameters (Example 10-3: Uncalibrated.. .1-4 10-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-2 10-1. .1-1 10-2.7 10-5. . . . . . . . . .1. Uncertainty of Result (Example 10-1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-2 10-5. . . . . . . . . . . . . . . .1-5 10-2. . . . . . . . . . . . . . Summary: Comparison Between Calibrated and Uncalibrated Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary: Uncertainties in Absolute Terms . . . . . . . . . . . . . . . . . . . Summary: Uncertainties in Absolute Terms (Example 10-3: Uncalibrated. .4. . . . . Summary: Uncertainties in Absolute Terms (Example 10-2: Uncalibrated Case) . . . . . . . Summary: Uncertainties in Relative Terms for the Calibrated Case . . . . . . . . . . . . . . . . . . . . . . . . . . . Independent Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1-6 10-2. . Pump Design Data (Tc p 20°C) . . . . . . Absolute Contributions of Uncertainties of Independent Parameters (Example 10-2: Uncalibrated Case) . . . . . . . . . . . . . . .1. . . . . . . . . .`. . . .
v
40 42 42 44 45 45 46 48 48 50 51 52 52 52 53 53 53 54 54 54
55 56 56 57 57 58 59 62 63 63 66 66 66
--`. . . . . . . . 10-4. . . . . . . .8-6. . . . . . .3.1-2 10-2. . . . . . . . . . . . . . . . . . . . . . . . . . . . .2-1 10-5.1-3 10-4. . . . . . . . . . . . . . . .2 10-1. . . . . . . . . . . . . . . . . . . Summary: Uncertainties in Relative Terms for the Uncalibrated Case . . . . . . . . . . . . . . . . . . Standard Deviation of Average Velocity (Example 10-1) . . . . . . . . . .1-2 10-2. . Relative Contributions of Uncertainties of Independent Parameters (Example 10-2: Uncalibrated Case) . . . . .9 10-2. . . . . . Inlet and Exit Pressure Elemental Systematic Standard Uncertainties . .1. . . . . . . . . . . . . . .``. Table of Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1-2 10-4. . . . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . .1. . . . . . . . . . . . . . . . . . . . . . . .`. . . . . . . . . . . . .1-1 10-5. . . .1-2 10-5. . . . . . .3. . . . . . . . . . . . . . . . . . . . . Uncertainty Propagation for Comparison With Independent Control . . . . . . . . . . . . . . . . . . . . . Evaluation of Analysis Error . . . Nonsymmetrical Systematic Uncertainty Case) . Nonsymmetrical Systematic Uncertainty Case) .1-7 10-3-1
10-3-2 10-4. . . . . . . . . . . . . Relative Uncertainties of Independent Parameters (Example 10-2: Calibrated Case) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .``.. . . D The Central Limit Theorem . . . . . . . . . .
vi
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`. . . . . . . . . . . . . . . . . . . . . . . . .````````. .
68 68 73 84 87 92
Nonmandatory Appendices A Statistical Considerations . . .`---
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.`. . . . . ..-`-`. . . Sensitivity Coefficient Estimates for Comparative Analysis . . . .`. . . ... . . . . . . . . . . . .. . . . . . . . . .```. . . . . . . . . . . . . . .`. . . .3-1 10-5. B Uncertainty Analysis Models . . . . . . . . . . . . . . . . . . . . . . . .`. . . . .10-5. . . .`. . . 08/17/2010 06:03:02 MDT
. . . . . . . . . . . . . . . . .. .3-2
Uncertainty Propagation for Comparative Uncertainty . . . C Propagation of Uncertainty Through Taylor Series . . . . . . . . . . . . . .

vii
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. it is recommended that the parties to a commercial test agree before starting the test and preferably before signing the contract on the method to be used for comparing the test results to the contractual guarantees. Therefore. yield the best available indication of the actual performance of the tested equipment. When tests are run in accordance with a Code. instrumentation. calculation methods. They were developed by balanced committees representing all concerned interests and specify procedures. the test results themselves. General Instructions. 08/17/2010 06:03:02 MDT
. The following information is based on that document and is included here for emphasis and for the convenience of the user of the Supplement. and uncertainty analysis. equipment-operating requirements. It is beyond the scope of any Code to determine or interpret how such comparisons shall be made.NOTICE
All Performance Test Codes must adhere to the requirements of ASME PTC 1. ASME Performance Test Codes do not specify means to compare those results to contractual guarantees. without adjustment for uncertainty. ASME Performance Test Codes provide test procedures which yield results of the highest level of accuracy consistent with the best engineering knowledge and practice currently available. It is expected that the Code user is fully cognizant of Sections 1 and 3 of ASME PTC 1 and has read them prior to applying this Supplement.

and was approved as an American National Standard by the ANSI Board of Standards Review on November 3. The two-fold objective was to improve the usefulness to the reader regarding clarity. “Test Uncertainty. The effort to update the 1998 revision began immediately upon completion of that document. conciseness. 08/17/2010 06:03:02 MDT
. and “random” (estimate of the limits of the error observed from the scatter of the test data). symbols. ASME PTC 19.FOREWORD
In March 1979 the Performance Test Codes Supervisory Committee activated the PTC 19. and technical treatment of the evolving subject matter. Work on the revision of the original 1985 edition began in 1991.11998. This 2005 revision is notable for the following significant departures from the 1998 text: (a) Nomenclature adopted for this revision is more consistent with the ISO Guide. 2005.” and which was intended—along with its subsequent editions—to provide a means of eventual standardization of nomenclature.” The determination of an uncertainty at some level of confidence is based on the root-sum-square of the systematic and random standard uncertainties multiplied times the appropriate expansion factor for the desired level of confidence (usually “2” for 95%).
viii
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.1-1985.1 Committee to revise a 1969 draft of a document entitled PTC 19. This same approach was used in the 1998 revision but the characterization of uncertainties at the standard-uncertainty level (“standard deviation”) was not as explicitly stated. “Measurement Uncertainty. The new nomenclature is expected to render PTC 19. as well as harmonization with the ISO “Guide to the Expression of Uncertainty in Measurement. (b) There is greater discussion of the determination of systematic uncertainties. and methodology of measurement uncertainty in ASME Performance Test Codes.” the new title reflecting the appropriate orientation of the document.1 Committee proceeded to develop a Performance Test Code Instruments and Apparatus Supplement which was published in 1985 as PTC 19.12005 more acceptable at the international level.” The PTC 19. 2005. Uncertainties remain conceptualized as “systematic” (estimate of the effects of fixed error not observed in the data). (c) There is new text on a simplified approach to determine the uncertainty of straightline regression.1 “General Considerations.1-2005 was approved by the PTC Standards Committee on September 13.” That revision was published as PTC 19. The new aspect is that both types of uncertainty are defined at the standard-deviation level as “standard uncertainties.

`.```. G.`.1 ON TEST UNCERTAINTY
(The following is the roster of the Committee at the time of the approval of this Supplement.`. A. Colorado State University J.. Steele. Secretary
COMMITTEE PERSONNEL
J. Osolsobe. F. K. 08/17/2010 06:03:02 MDT
. Ron Dieck Associates. Inc..`. Environmental Systems Corp.````````.-`-`. Figliola. M.. Bechtel National Corp.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. Intel Corp.. Rabensteine. S. Coutts. WSMS R. J. Vice Chair G.`.PERFORMANCE TEST CODE COMMITTEE 19. Steele. Clemson University H. W. H.``. Mississippi State University
ix
--`. G. Chair W. Dieck.)
OFFICERS
R. A. Soltani.. H. Pratt & Whitney D.. Dieck. Iyer. Maveety. Bernardin. R.

Bannister R. Keyser S. Nuspl A. Yost. C. Gerber P. Hays F. Albert R. G. Gerhart T. L. M. H. Siegmund J. E. Dooley A. Milton S. H. Silvaggio W. Weinman. G. McHale J. C. Yost
HONORARY MEMBERS
W. Korellis P. Heil R. Westcott W. D. Johnson D. McHale M. G. Secretary
COMMITTEE PERSONNEL
P. Sommerlad
x
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. 08/17/2010 06:03:02 MDT
. Jorgensen G. M. G. L. R. Friedman G. Friedman. W. O. Mittendorf R.PERFORMANCE TEST CODES STANDARDS COMMITTEE
OFFICERS
J. Wood J. A. Steele J. Allen J. J. Campbell M. P. R. Burns W. M. A. W. C. A. Vice Chair S. Chair J. Plumley R. J. P. C. J. P. Rabensteine J. Light
MEMBERS EMERITI
R. J. Egli J. Priestley J. R. R.

harmony with international guidelines and standards. it is applicable for all known measurement and test uncertainty analyses. The term “precision” also is not used in this Supplement. i. this harmonization means the elimination of such ambiguous terms as bias. The most common confidence level used in this Supplement is 95% although methods for employing alternate confidences are also given. They are “systematic” and “random. The paramater values and uncertainty levels used throughout the examples are for illustrative purposes only and are not intended to be typical of standard tests. Throughout the Supplement. The confidence level of 95% is applied to “expanded” uncertainty. precision. The term “bias” is not used in this Supplement. “standard” uncertainties are always equivalent to a single standard deviation of the average. and precision index. The latter describes the limits to which a systematic error may be expected to go with some confidence. great care was taken to assure continued harmony with the International Organization for Standardization (ISO) Guide to the Expression of Uncertainty in Measurement (GUM) [1]. 08/17/2010 06:03:02 MDT
1-2 HARMONIZATION WITH INTERNATIONAL STANDARDS It is recognized that this Supplement and promulgated international uncertainty standards and/ or guides must be in harmony.1-2005
Section 1 Introduction
1-1 GENERAL This Supplement has significant additions and Sections that have been rewritten to both add to the available technology for uncertainty analysis and to make it easier for the practicing engineer.” Random errors (whose effects are estimated with “Random Standard Uncertainties”) cause scatter in test data. using test uncertainty analysis. While this Supplement is in harmony with the ISO GUM. The ISO GUM utilizes two major classifications for errors and uncertainties. Instead.. Type B uncertainties do not have data to calculate a standard deviation and must be estimated by other means.TEST UNCERTAINTY
ASME PTC 19. The former describes an error source that causes scatter in test data. Systematic errors (whose effects are estimated with “Systematic Standard Uncertainties”) do not. This Supplement utilizes two major classifications for errors and uncertainties.
1
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. the combined terms of “systematic error” and “systematic uncertainty” are used. The latter describes the limits to which a random error may be expected to reach with some confidence.” 1-3 APPLICATIONS This Supplement is intended to serve as a reference to the various other ASME Instruments and Apparatus Supplements (PTC 19 Series) and to ASME Performance Test Codes and Standards in general. They are “Type A” and “Type B. In rewriting this Supplement. the intent is to provide a Supplement that can be utilized easily by engineers and scientists whose interest is the objective assessment of data quality. For the practicing engineer.e. In addition. In this Supplement. The former describes an error source whose effect is systematic or constant for the duration of a test. was included in this Supplement for improved harmony with international guidelines and standards. bias limit. and the estimation of their limits. too. which is the uncertainty. In addition. Instead the combined terms of “random error” and “random uncertainty” are used. this Supplement emphasizes the effects of errors rather than the basis of the information utilized in the estimation of their limits. This term. Throughout. Harmonization of this Supplement with the ISO GUM is achieved by encouraging subscripts with each uncertainty estimate to denote the ISO Type. the effects of errors. careful attention was paid to discriminating between errors.” Type A uncertainties have data with which to calculate a standard deviation. using subscripts of either “A” or “B. the term “standard” uncertainty has been introduced to improve
.

and for propagation of those uncertainties into the uncertainty of a test result. describe. The end result of an uncertainty analysis is a numerical estimate of the test uncertainty with an appropriate confidence level. Depending on the application. (d) reduces the risk of making erroneous decisions. (b) fosters an understanding of potential error sources in a measurement system and the effects of those potential error sources on test results.ASME PTC 19. 08/17/2010 06:03:02 MDT
. and the effects of those uncertainties on derived test results. The various statistical terms involved are defined in the Nomenclature (subsection 3-1) or Glossary (subsection 3-2). Analysis of test measurement and result uncertainty is useful because it (a) facilitates communication regarding measurement and test results. (c) guides the decision-making process for selecting appropriate and cost-effective measurement systems and methodologies. and illustrate the various terms and methods used to provide meaningful estimates of the uncertainty in test parameters and methods.
2
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. uncertainty sources may be classified either by the presumed effect (systematic or random) on the measurement or test result. or by the process in which they may be quantified (Type A or Type B).1-2005
TEST UNCERTAINTY
Section 2 Object and Scope
2-1 OBJECT The object of this Supplement is to define. 2-2 SCOPE The scope of this Supplement is to specify procedures for evaluation of uncertainties in test parameters and methods. and (e) documents uncertainty for assessing compliance with agreements.

. 08/17/2010 06:03:02 MDT
␦p (unknown) total error. t95. difference between the assigned value of a parameter or a test result and the true value ⑀p (unknown) true random error. b−p upper and lower values of nonsymmetrical systematic standard uncertainty Np number of measurements or sample points or observations available (sample size) Rp result sRp random standard uncertainty of a result sXp standard deviation of a data sample. confidence level: the probability that the true value falls within the specified limits.e.1-2005
Section 3 Nomenclature and Glossary
3-1 NOMENCLATURE b Xp systematic standard uncertainty component of a parameter b Xkp systematic standard uncertainty associated with the kth elemental error source bRp systematic standard uncertainty component of a result bXYp covariance of the systematic errors in X and Y b+. random component of ␦ p absolute sensitivity ′p relative sensitivity p (unknown) true average of a population p number of degrees of freedom p (unknown) true standard deviation of a population 2p (unknown) true variance of a population Indices Ip total number of variables ip counter for variables jp counter for individual measurements Kp total number of sources of elemental errors and uncertainties kp counter for sources of elemental errors and uncertainties Lp total number of correlated sources of systematic error lp counter for correlated sources of systematic error Mp total number of multiple results mp counter for multiple results Np total number of measurements
3-2 GLOSSARY calibration hierarchy: the chain of calibrations that links or traces a measuring instrument to a primary standard. up combined standard uncertainty Up expanded uncertainty U+.. U−p upper and lower values of the nonsymmetrical expanded uncertainty Xp individual observation in a data sample of a parameter Xp sample mean. fixed or constant component of ␦
3
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. estimate of the standard deviation of the population x sXp random standard uncertainty of the mean of N measurements SEEp standard error of estimate of a leastsquares regression or curve fit tp Student’s t value at a specified confidence level with degrees of freedom. calibration: the process of comparing the response of an instrument to a standard instrument over some measurement range. average of a set of N individual observations of a parameter p (unknown) true systematic error.TEST UNCERTAINTY
ASME PTC 19. i.

with a defined level of confidence. stress. such as temperature.-`-`.. sample size (N): the number of individual values in a sample. elemental random error source: an identifiable source of random error that is a subcomponent of total random error.
parameter: quantity that could be measured or taken from best available information. mean (X): the arithmetic average of N readings. (usually 95%). population standard deviation (): a value that quantifies the dispersion of a population.`---
result (R): a value calculated from a number of parameters.
total error (␦): the true.. or specific heat..`. random error (⑀): the portion of total error that varies randomly in repeated measurements of the true value throughout a test process. pressure. Type B uncertainty: uncertainties are classified as Type B when data is not used to calculate a standard deviation. true value: the error-free value of a parameter or test result. unknown difference between the assigned value of a parameter or test result and the true value.
influence coefficient: see sensitivity. elemental systematic standard uncertainty (b Xk ): an estimate of standard deviation of an elemental systematic error source.
4
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. used in determining a result.. sample standard deviation (sx): a value that quantifies the dispersion of a sample of measurements as given by eq.ASME PTC 19. standard error of estimate (SEE): the measure of dispersion of the dependent variable about a least squares regression or curve.`. population mean (): average of the set of all population values of a parameter.1-2005
TEST UNCERTAINTY
degrees of freedom ( ): the number of independent observations used to calculate a standard deviation.
Student’s t: a value used to estimate the uncertainty for a given confidence level.`.`.. 08/17/2010 06:03:02 MDT
. population: the set of all possible values of a parameter.. Type A uncertainty: uncertainties are classified as Type A when data is used to calculate a standard deviation for use in estimating the uncertainty.`. requiring the uncertainty to be estimated by other methods.2). systematic standard uncertainty (b X ): a value that quantifies the dispersion of a systematic error associated with the mean.```. uncertainty interval: an interval expressed about a parameter or test result that is expected to contain the true value with a prescribed level of confidence. X and s X are statistics. (4-3.``.
expanded uncertainty (UX or UR ): an estimate of the plus-or-minus limits of total error. sensitivity: the instantaneous rate of the change in a result due to a change in a parameter. The value used is called the assigned value.
--`. random standard uncertainty of the sample mean (s X ): a value that quantifies the dispersion of a sample mean as given by eq. statistic: any numerical quantity derived from the sample data.````````.3). elemental random standard uncertainty (s Xk ): an estimate of the standard deviation of the mean of an elemental random error source. systematic error (): the portion of total error that remains constant in repeated measurements of the true value throughout a test process.
elemental systematic error source: an identifiable source of systematic error that is a subcomponent of the total systematic error. traceability: see calibration hierarchy. (4-3.

`. 08/17/2010 06:03:02 MDT
. 4-3 MEASUREMENT UNCERTAINTY There is an inherent uncertainty in the use of measurements to represent the true value. Other confidence levels may be used. The total random error in a measurement is usually the sum of the
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.. These terms are illustrated in Fig. including the measurement process and the data reduction process. (e) All appropriate engineering corrections are applied to the test data as part of the data reduction and/or results analysis process. The total systematic error in a measurement is usually the sum of the contributions of several elemental systematic errors. if required. 4-2-2). 4-3.-`-`. etc. 4-2. measurement methods.````````. environmental conditions..1 Random Standard Uncertainty Any single measurement of a parameter is influenced by several different elemental random error sources. (See Nonmandatory Appendix B. If an infinite number of measurements of a parameter were to be taken following the defined test process.1 Random Error Random error. the resulting population of measurements could be described statistically in terms of the population mean. is the portion of the total error that varies randomly in repeated measurements throughout the conduct of a test..`---
contributions of several elemental random error sources. the values of these elemental random error sources change resulting in the random scatter evident in the successive measurements..``. 95% confidence levels have been used throughout this document in accordance with accepted practice. Accurate measurement requires minimizing both random and systematic errors (see Fig.`. which results in a difference between the measured value. 4-2-1). is defined.) 4-2 MEASUREMENT ERROR
--`. Elemental random errors may arise from uncontrolled test conditions and nonrepeatabilities in the measurement system. the population standard deviation. 4-2..1 for a population of measurements that is normally distributed. ␦.```. .TEST UNCERTAINTY
ASME PTC 19. . For expanded uncertainty. ⑀. etc. and the frequency distribution of the population. The difference between the measured value and the true value is the total error. is controlled for the duration of the test. Elemental systematic errors may arise from imperfect calibration corrections. Total error consists of two components: random error and systematic error (see Fig. In successive measurements of the parameter. X. with respect to the conditions of the item under test and the measurement system employed for the test. . (d) The measurement system is calibrated and all appropriate calibration corrections are applied to the resulting test data.`. is the portion of the total error that remains constant in repeated measurements throughout the conduct of a test. Since the true value is unknown. The total uncertainty in a measurement is the combination of uncertainty due to random error and uncertainty due to systematic error.2 Systematic Error Systematic error. data reduction techniques.`. data reduction techniques. total error cannot be known and therefore only its expected limits can be estimated. (c) The test process. For measurements with zero systematic
5
Every measurement has error.1-2005
Section 4 Fundamental Concepts
4-1 ASSUMPTIONS The assumptions inherent in test uncertainty analysis include the following: (a) The test objectives are specified. (b) The test process.`. measurement methods.. 4-3. and the true value.

```.. 08/17/2010 06:03:02 MDT
. can be used to define the probable interval about the sample mean that is expected to contain the population mean with a defined level of confidence...2). is given by
6
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
.. there is an inherent error in the use of the sample mean to estimate the population mean.. The random standard uncertainty of the sample mean is related to the sample standard deviation as follows:
sX p
sX
(4-3. Since only a finite number of measurements are acquired during a test.`. the interval ± will include approximately 68% of the population and the interval ± 2 will include approximately 95% of the population.3)
Xp
jp1
͚
N
Xj
(4-3. For a normal distribution. sX.1-2005
error (refer to para.1)
ΊN
N
For a normally distributed population and a large sample size (N > 30).``.````````. the population mean is equal to the true value of the parameter being measured and the population standard deviation is a measure of the scatter of the individual measurements about the population mean.. For a defined frequency distribution.-`-`.`---
TEST UNCERTAINTY
Fig.`. 4-2-1 Illustration of Measurement Errors
sX p
Ίj͚ p1
N
( Xj − X )2 N−1
(4-3. sX. The sample standard deviation.`. 4-2. the true population mean and population standard deviation are unknown but can be estimated from sample statistics. the random standard uncertainty of the sample mean. the interval X ± sX is expected to contain the true population mean with 68% confidence and the interval X ± 2sX is expected to contain the true population mean with 95% confidence [where the value 2 represents the Student’s t value for 95% confidence and degrees of
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.2)
Since the sample mean is only an estimate of the population mean. The sample mean.`. X. is given by
where Xj represents the value of each individual measurement in the sample and N is the number of measurements in the sample.ASME PTC 19.`.

Knowledge of the frequency distribution and standard deviation of this population permits describing the uncertainty in X due to this single sample elemental systematic error in terms of a confidence interval. is equivalent to the error imparted to each individual measurement. As  Xk is constant for the test. 4-3.1-2005
Fig. In general. but unknown.TEST UNCERTAINTY
ASME PTC 19. The elemental systematic standard uncertainty. Each of these elemental systematic error
7
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
sources contributes a constant. error. and (c) it typically reduces the value of the random standard uncertainty of the sample mean. b Xk. 4-3. increasing the number of measurements collected during a test and used in the preceding formulas is beneficial as (a) it improves the sample mean as an estimator of the true population mean. (b) it improves the sample standard deviation as an estimator of the true population standard deviation.2 Systematic Standard Uncertainty Every measurement of a parameter is influenced by several different elemental systematic error sources. 08/17/2010 06:03:02 MDT
. the error imparted to the average value of successive measurements.  Xk. While  Xk is unknown. is defined as a value
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. to the successive measurements of a parameter for the duration of the test (the subscript k is used to denote a specific elemental error source). it may be postulated to come from a population of possible error values from which a single sample (error value) is drawn and imparted to the average measurement for the test. X [as given by eq. 4-2-2 Measurement Error Components
freedom of greater than or equal to 30 where the degrees freedom for the random standard uncertainty is N−1 (see subsection 6-1)].1].

as failure to identify any significant source of systematic error will lead to an underestimation
8
--`.1-2005
TEST UNCERTAINTY
Fig. By definition. and (c) combining the elemental systematic standard uncertainties into an estimate of the total systematic standard uncertainty for the average measurement.  X. total systematic standard uncertainty is quantified by (a) identifying all elemental sources of systematic error for the measurement. it is customary to use engineering analyses and experience to estimate the limits of the
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. b X.```.. successive measurements of a parameter do not provide sufficient data for direct computation of a standard deviation as described in para.``.````````. 08/17/2010 06:03:02 MDT
. Attempting to identify all elemental sources of systematic error requires a thorough understanding of the test objectives and test process.`. elemental systematic standard uncertainties for each source are evaluated.2.ASME PTC 19. As  Xk is both constant and unknown during a test.1 Engineering Judgment..2. 4-3. In these situations. total systematic error is constant. When neither published information or special data is available. an elemental systematic standard uncertainty is a value that quantifies the dispersion of the population of possible  Xk values at the standard deviation level. Once all elemental sources of systematic error are identified.1.. (b) evaluating elemental systematic standard uncertainties as the standard deviations of the possible systematic error distributions.`---
of test uncertainty.2. 4-3.2.`. As with elemental systematic error. Attempting to identify all of the elemental sources of systematic error for a measurement is an important step of an uncertainty analysis. it is often necessary to rely upon engineering judgment to quantify the dispersion of errors associated with an elemental error source.`. Therefore. All of the elemental systematic errors associated with a measurement combine to yield the total systematic error in the measurement..`. published information..1 Identifying Elemental Sources of Systematic Error. unknown. the evaluation of an elemental systematic standard uncertainty requires that a standard deviation be evaluated from engineering judgment.2 Evaluating Elemental Systematic Standard Uncertainties. or special data.
4-3.`.. and may be postulated to come from a population of possible error values from which a single sample (error value) is drawn and imparted to the average measurement for the test. is defined as a value that quantifies the dispersion of the population of possible  X values at the standard deviation level. Total systematic standard uncertainty. For further discussion refer to subsection 5-4.
4-3. 4-3.1 Distribution of Measured Values (Normal Distribution)
that quantifies the dispersion of the population of possible  Xk values at the standard deviation level. Typically.-`-`.

turbine flow parameters.. a normal distribution and large degrees of freedom are often assumed.`.``.2. 08/17/2010 06:03:02 MDT
. If the published information is presented as a confidence interval (limits of error at a defined level of confidence).-`-`. in a gas turbine test. unknown  Xk values.3 Special Data..4) would apply (refer to Nonmandatory Appendix B for values of the Student’s t statistic at other confidence levels and degrees of freedom).1-2005
elemental systematic error at 95% confidence. For some elemental systematic error sources. The value of 2 in the equation is based on the assumption that the population of possible systematic errors is normally distributed. and other technical references may provide quantitative information regarding the dispersion of errors for an elemental systematic error source in terms of a confidence interval. airflow can be measured with an orifice or a bell mouth nozzle..`. For some elemental systematic error sources. special data may be obtained that manifests the dispersion of the population of possible. published information from calibration reports. then the elemental systematic standard uncertainty is estimated as the expanded uncertainty divided by the coverage factor. the elemental systematic standard uncertainty is estimated as follows:
BXk
2
b Xk p
(4-3. the BXk estimates are made such that this degrees of freedom will be large (≥30). an ISO expanded
9
uncertainty statement. or computed from compressor speed-flow rig data. For these cases. For a 95% confidence level and large degrees of freedom. an interval is estimated which is expected to contain 95% of the population of possible  Xk values.`. Not withstanding information to the contrary. and that the limits of error are symmetric (equally spread in both the positive and negative directions). For treatment of nonsymmetric systematic uncertainty see subsection 8-2. Also. it can be shown that this large degrees of freedom (≥30) corresponds to an uncertainty in the estimate of BXk of 13% or less. For situations in which the frequency distribution and degrees of freedom are unspecified.
4-3. This uncertainty in the estimate can be converted into a degrees of freedom for the systematic standard uncertainty as shown in Nonmandatory Appendix B.`.2 Published Information. If the published information is presented as a multiple of a standard deviation.2.5)
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.TEST UNCERTAINTY
ASME PTC 19. (4-3. refer to an appropriate statistics textbook. If the analyst thinks that the error distribution might be other than normal..```. or jet nozzle calibrations. or a multiple of a standard deviation. then a different factor would be used to convert the 95% confidence level estimate of the systematic error limits to an elemental systematic standard uncertainty (see Nonmandatory Appendix B). Based upon these assumptions. the elemental systematic standard uncertainty may be evaluated as follows:
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`. If the published information is presented as an ISO expanded uncertainty at a defined coverage factor (sometimes referred to as a “k factor”). that the estimation of the limits of the error is based upon large degrees of freedom. Possible sources of this special data include (a) interlaboratory or interfacility tests. the Student’s t statistic is used. the value of the Student’s t statistic is approximated as 2 and eq. then the elemental systematic standard uncertainty is estimated as the confidence interval divided by a statistic that is appropriate for the frequency distribution of the error population. such as uniform (rectangular).````````.4)
The variable BXk in the preceding equation represents the 95% confidence level estimate of the symmetric limits of error associated with the kth elemental error source. For a normal distribution. the analyst typically assumes that the population of possible  Xk values is normally distributed. there is some level of uncertainty associated with the estimate of BXk.`. for example. In certain situations. instrument specifications. Usually. For situations involving other frequency distributions. then the elemental systematic standard uncertainty is estimated as the multiple of the standard deviation divided by the multiplier. knowledge of the physics of the measurement system will lead the analyst to believe that the limits of error are nonsymmetric (likely to be larger in either the positive or negative direction). Using the recommendations in Nonmandatory Appendix B.2. The specific value of this statistic must be selected on the basis of the defined confidence level and degrees of freedom associated with the confidence interval. and (b) comparisons of independent measurements that depend on different principles or that have been made by independently calibrated instruments.2. In other words. 4-3.`---
b Xk p
Ί
1 N X
NX
jp1
k
͚
k
(X kj − X k )2
N Xk − 1
(4-3...

k (a) Several independent measurement methods that depend on different principles are used to measure the same parameter.6)
where b Xp the systematic standard uncertainty s Xp the random standard uncertainty of the mean
The expanded uncertainty of the measurement mean is the total uncertainty at a defined level of confidence. (4-3.5) is equal to the number of k independent laboratories employed. In some cases.2. For applications in which a 95% confidence level is appropriate. the expanded uncertainty is calculated as follows:
UX p 2uX
(4-3. b X.5) is equal to the number of independent measurement methods employed. (4-3.
4-3. which is the total uncertainty at the standard deviation level. elemental systematic standard uncertainties may arise from the same elemental error source and are therefore correlated. See subsection 8-1 for a detailed discussion. 08/17/2010 06:03:02 MDT
.3 Combining Elemental Systematic Standard Uncertainties. the total uncertainty in a measurement is the combination of uncertainty due to random error and uncertainty due to systematic error. then the value for N X used in eq.5) to evaluate the elemental systematic standard uncertainty associated with the error inherent to the various measurement methods. Provided all elemental systematic standard uncertainties are evaluated in terms of their influence on the parameter being measured and in the units of the parameter being measured. these elemental systematic standard uncertainties are combined per subsection 6-2. The following illustrate some k possible cases where N X may be greater than one. Thus. is calculated as follows:
uX p
Ί ( b X)2 + ( sX) 2
(4-3.ASME PTC 19. Otherwise. If the average measurement reported for the test is the average of the results from all of the measurement methods. 4-3. Once evaluated. all of the elemental systematic standard uncertainties influencing a measurement are combined into an estimate
10
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
of the total systematic standard uncertainty for the measurement. k (4-3. Expanded uncertainty is used to establish a confidence interval about the measurement mean which is expected to contain the true value. only a single sample from the population of possible  Xk values is included in the computation of the average measurement for the test (X) and hence N X p 1. these elemental systematic standard uncertainties are combined per subsection 7-4.7)
where the assumptions required for this simple equation are presented in subsection 6-4.5) to evaluate the elemental systematic standard uncertainty associated with the error inherent to the various laboratories.1-2005
TEST UNCERTAINTY
where NXkp the number of special data values used in the computation of b Xk N X p the number of independent samples from k the population of possible  Xk values that are averaged together in the computation of the average measurement for the test (X) Xkp the average of the set of special data
Xkjp the j th data point of the set of special data that manifests the dispersion of the population of possible  Xk values associated with the k th elemental error source
For most measurements (especially those made using a single instrument calibrated at a single laboratory and installed in a single location). the interval X ± UX is expected to contain the true value with 95% confidence (see Fig. then the value for N X used in eq.3 Combined Standard Uncertainty and Expanded Uncertainty As mentioned previously. (b) An instrument is sent to multiple laboratories to obtain calibration data for the instrument prior to using the instrument in a test.
4-3. The results from each of the independent laboratories (each determined as an offset to be applied to the instrument when measuring a specific input level) are used as input to eq. The combined standard uncertainty of the measurement mean.3). If the average measurement from the instrument reported for the test is based upon application of the average offset from all of the laboratories.
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. (4-3. The results from each of the measurement methods (each determined as an average value over the duration of the test) are used as input to eq.

```.`. prior to the conduct of a test. A pretest uncertainty analysis is based on data and information that exist before the test. calibration techniques. prior to expending resources to conduct a test.`. a pretest uncertainty analysis facilitates communication between all parties to the test about the expected quality of the test. prior measurement uncertainty analyses. In addition to the data and information used to conduct the pretest uncertainty analysis. previous tests with similar instrumentation. and (3) increasing sample sizes by increasing sampling frequencies..TEST UNCERTAINTY
ASME PTC 19. and/or measurement of different parameters. and/or conducting repeated testing. 4-3.`. increasing test duration.. and/or measurement locations. after conducting a test..``..`---
Fig. testing under different conditions.
11
(2) selecting alternative measurement methods by varying test instrumentation.`.`. A pretest uncertainty analysis allows corrective action to be taken.. (b) The objective of a posttest analysis is to establish the uncertainty interval for a test result.1-2005
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`.-`-`. Possible corrective actions include (1) selecting alternative testing methods that rely upon different analysis procedures. either to decrease the expected uncertainty to a level consistent with the overall objectives of the test or to reduce the cost of the test while still attaining the objectives. 08/17/2010 06:03:02 MDT
. Additionally.3 Uncertainty Interval
4-4 PRETEST AND POSTTEST UNCERTAINTY ANALYSES (a) The objective of a pretest analysis is to establish the expected uncertainty interval for a test result. if necessary. pretest and
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. expert opinions. such as calibration histories. This can be essential to establishing agreement on any deviations from applicable test code requirements and can help reduce the risk that disagreements regarding the testing method will surface after conducting the test.. a posttest uncertainty analysis is based upon the additional data and information gathered for the test including all test measurements. special tests. and. installation methods.````````.

..`.
12
--`.-`-`. A posttest uncertainty analysis serves to (1) validate the quality of the test result by demonstrating compliance with test requirements.. etc.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.```.`.ASME PTC 19..`.
(2) facilitate communication of the quality of the test result to all parties to the test.. 08/17/2010 06:03:02 MDT
. and (3) facilitate interpretation of the quality of the test by those using the test result.`.``..`.1-2005
TEST UNCERTAINTY
posttest instrument calibration data.````````.

`. measurement methods.``. or both.1).. Calibrations are also used to provide traceability to known reference standards or physical constants. often overlooked. Consideration must be given to the selection of the appropriate “true value” of the measurement and the time interval for classifying errors as systematic or random. This simple step.-`-`. and data reduction techniques. when analyzing a thermocouple measurement in a gas stream. The starting point for the analysis could begin with the “true value” defined as the metal temperature of the thermocouple junction. or the mass flow weighted average of the gas temperature at the plane of the instrumentation. Any of the aforementioned “true values” may be appropriate. providing the ultimate reference for every standards laboratory. is essential to successfully develop and apply the uncertainty information.```.`.. Estimates to reflect the extent of these errors are
13
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. Examples of error sources include imperfect calibration corrections. The main purpose of the calibration process is to eliminate large.2 Uncertainty Due to Test Article and/or Instrumentation Installation Measurement uncertainty can also exist from interactions between (a) the test instrumentation and the test media or (b) between the test article and test facility. the gas stagnation temperature or junction temperature corrected for probe effects. 5-2 SELECTION OF THE APPROPRIATE “TRUE VALUE” Depending on the user’s perspective. known systematic errors and thus reduce the measurement uncertainty to some “acceptable” level.`.. uncontrolled test conditions.. Examples of these types of uncertainty are
--`.1-2005
Section 5 Defining the Measurement Process
5-1 OVERVIEW The first step in a measurement uncertainty analysis is to clearly define the basic measurement process. The selection of the “true value” for the uncertainty analysis must be consistent with the goal of the measurement [3]. These uncertainties in the measurement process can be grouped by source (a) calibration uncertainty (b) uncertainty due to test article and/or instrumentation installation (c) data acquisition uncertainty (d) data reduction uncertainty (e) uncertainty due to methods and other effects 5-3. This section provides an overview of how the measurement process should be defined.`---
.````````. Having decided on the “acceptable” level. the errors associated with measuring the true value must be identified.`. several starting points or “true values” can be selected. Each additional level in the calibration hierarchy adds uncertainty in the measurement process (see Fig. 08/17/2010 06:03:02 MDT
represented as uncertainties. 5-3 IDENTIFICATION OF ERROR SOURCES Once the true value has been defined. 5-3. environmental conditions. Requirements of military and commercial contracts have led to the establishment of extensive hierarchies of standards laboratories. a national standards laboratory is at the apex of these hierarchies. 5-3. several measurement objectives or goals and hence corresponding “true values” (measurements with ideal zero error) may exist simultaneously in a measurement process.`..1 Calibration Uncertainty Each measurement instrument may introduce random and systematic uncertainties..TEST UNCERTAINTY
ASME PTC 19. the calibration process achieves that goal by exchanging the large systematic uncertainty of an uncalibrated or poorly calibrated instrument for the smaller combination of systematic uncertainties of the standard instrument and the random uncertainties of the comparison. In some countries. For example.

By comparing known input values with their measured results. 08/17/2010 06:03:02 MDT
.3 Data Acquisition Uncertainty Uncertainty in data acquisition systems can arise from errors in the signal conditioning. A correction factor must be applied that corrects between uninstalled to installed aircraft engine performance. An example is the testing of an automotive engine. the sensors.ASME PTC 19. it may affect the measured total and static pressure and thus the calculated airflow. it is not always possible to do this. In these cases. and the effects at other altitude conditions may need to be extrapolated. convection. 5-3.
5-3. Depending on the design of the pitot static probe.. the recording devices.`. it is necessary to evaluate each of the elemental
--`. (2) Environmental effects on sensors/instrumentation may exist when the sensors experience environmental effects that are different from those observed during calibration.1 Generic Measurement Calibration Hierarchy
(a) Interactions Between the Test Instrumentation and Test Media: (1) Installation of sensors in the test media may cause intrusive disturbance effects.````````..`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. The effect of the oily air can influence the quoted rating of the unit..`.`. (b) Interactions Between the Test Article and Test Facility: (1) Test-facility limitations for certification testing affects product measurement uncertainty. An automotive test facility may only be able to test at specified altitudes and speeds. A second example is the testing of a gas turbine engine in an altitude facility. The facility simulates altitude by lowering the ambient pressure at the test article exhaust and raising the inlet pressure at the engine inlet. In appli14
cation the inlet pressure is elevated due to the ram drag effects of the aircraft.```. etc.1-2005
TEST UNCERTAINTY
Fig. The fuel consumption of an automotive engine changes with altitude and speed. However.-`-`. An example may be an air conditioner that was bench tested in a laboratory but used in an automotive mechanics shop.``.`.. (2) Facility limitations for testing may require extrapolations to other conditions. These may be such things as conduction.. estimates of the data acquisition system uncertainty can be obtained.`.. An example could be the measurement of airflow in an air conditioning duct. The best method to minimize the effects of many of these uncertainty sources is to perform overall system calibrations. and radiation on a sensor when installed in a gas turbine.

If the nature of an elemental error is fixed over the duration of the defined measurement process. beyond those contained in calibration. systematic and random). Those evaluated with statistical methods are classified as Type A.`.. These uncertainty sources. This categorization approach supports the identification. when comparing results among various laboratories. 5-3. then the error contributes to the systematic uncertainty. data acquisition.`. 5-4 CATEGORIZATION OF UNCERTAINTIES This Standard delineates uncertainties by the effect of the error (i. 08/17/2010 06:03:02 MDT
.`.````````. are classified as Type B.. Because measurement uncertainties are categorized by the effect of the error. The significance of this is discussed in para. For example. 5-4. may significantly affect the uncertainty of the final results. and managing of test uncertainties. understanding.1-2005
uncertainties and to combine them to predict the overall uncertainty. The stability of a measurement method is a generic concept related to the closeness of agreement between test results. which are assumed fixed over the life of the calibration interval. varying experimental setups. (b) using approximating engineering relationships or violating their assumptions. For example.`. Variability in independent test results obtained under different test conditions. categorizes the uncertainties based on the method used to estimate uncertainty. Depending on the selection of the defined measurement process. installation sources. the computational resolution uncertainty sources are often negligible. which are evaluated by other means.TEST UNCERTAINTY
ASME PTC 19.```. if the test objective is to measure the average gas mileage of model “XYZ” cars. With the recent advances in computer systems.3 Test Objective The classification and number of error sources are often affected by the test objective.4 Data Reduction Uncertainty Computations on raw data are done to produce output (data) in engineering units. calibration corrections. the objective of the test may affect the categorization as discussed in para.-`-`. The effects of a time interval may also be important when considering the stability and control of a test process.2. the time interval and duration of the measurement process can be important considerations and so must be clearly
15
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
stated. In addition. Other examples of data reduction uncertainty include (a) the assumptions or constants contained in the calculation routines. the variability among or between cars of the same model must be considered. Random error obtained in a test from a given car would not include car-to-car
--`. 5-4. Typical uncertainty sources in this category stem from curve fits and computational resolution.1 Alternate Categorization Approach An alternate approach.. 5-4.3. The time interval must be clearly specified to classify an error. while those.`. which is used in the ISO GUM. can be considered variable if the process consists of a time interval encompassing several different calibrations.``.e. For example. however. then the source contributes to the random uncertainty.5 Uncertainty Due to Methods and Other Effects Uncertainties due to methods are defined as those additional uncertainty sources that originate from the techniques or methods inherent in the measurement process. and it may not always be the same interval as the test duration. 5-3. and data reduction.. or configuration changes allow for additional between-test random errors. it may be appropriate to classify an error as random rather than as systematic even though that error may have been constant for the duration of any single test. 5-4. If the error source tends to cause scatter in repeated observations of the defined measurement process. curve-fit error uncertainty can be significant.2 Time Interval Effects Errors that may be fixed over a short time period may be variable over a longer time period. These additional uncertainties may be of either a systematic or random nature depending on their effect on the measurement... there may be no simple correspondence between random or systematic and Type A or Type B. 5-4. Process stability is estimated from observations of scatter within a data set and is treated as a random error..`---
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. and (c) using an empirically derived correlation such as empirical fluid properties.

The first test is run with the standard or baseline configuration. To observe the random error associated with car-to-car variability. Of course.`. with the smallest measurement uncertainty possible. if the data of interest is gas mileage of a given single car.```.`. the experiment would need to be run again using a random selection of different cars within the same model (see Fig.3 Difference Between “Within” and “Between” Sources of Data Scatter
variations and thus would not represent all random error sources. road conditions. the effects of correlated uncertainties (see Section 8) may cause the total uncertainty of the difference between the test results to be less than the uncertainty of each separate test result. The second test is then run in the same facility with the design change and hopefully with instruments. An example of back-to-back uncertainty analysis is shown in Example 8. The same short-term and long-term effects must be applied for other variables affecting gas mileage (temperature. 08/17/2010 06:03:02 MDT
..1 in subsection 8-1. etc.````````.`.`. altitude. This variation would be more representative of the total random error associated with determining gas mileage for the fleet of model “XYZ” cars.. Depending on whether common instrumentation. the net effect of a design change. then the estimated variation with testing the representative given car is an appropriate estimate for the random error. The total variation in the test result is greater than that observed from a test of a single given car.
16
--`.``. setups.1-2005
TEST UNCERTAINTY
Fig..ASME PTC 19.. driver variations. 5-4. 5-4. and calibrations are used between comparative tests.. The difference between the results of these tests is an indication of the effect of the design change..`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. humidity.`. setups.3).).-`-`.
5-5 COMPARATIVE VERSUS ABSOLUTE TESTING The objective of a comparative test (also known as a back-to-back test) is to determine. and calibrations identical to those used in the first test.

If the sample standard deviation of the variable being measured is also expected to be representative of other possible random variations in the measurement. 08/17/2010 06:03:02 MDT
. When a statistic is calculated from the sample. the degrees of freedom is the sample size (N). (43. or minutes. In a sample of measurements. e.. then this information can be used to streamline the test procedure by reducing the number of measurements taken in the later phases of the test. Once the random variation of the test variables is understood. X is calculated by eq. as with a computer-based data acquisition system. (4-3. these larger data sets are taken in the early phases of an experimental program..3). For example. taking multiple measurements as a function of time while holding all other conditions constant would identify the random variation associated with the measurement system and the unsteadiness of the test condition. If a much larger set of previous measurements for the same test conditions is available. This “short time frame averaged” value should then be handled in the same manner as a single measurement. and each has N − 1 degrees of freedom ( )
pN−1
(6-1.g.2)
where
1 NP
N
P
XP p
͚ XP jp1
j
(6-1. the degrees of freedom associated with the statistic is reduced by one for every estimated parameter used in calculating the statistic. the appropriate random standard uncertainty of the mean (sX) is given by eq.1-2005
Section 6 Uncertainty of a Measurement
6-1 RANDOM STANDARD UNCERTAINTY OF THE MEAN 6-1. This type of estimate is an ISO Type A estimate.TEST UNCERTAINTY
ASME PTC 19.2) and (4-3. Information about the possible variations in a single measurement must be obtained from previous measurements of the variable taken over the time frame and conditions that cover the variations
17
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
in the variable. the time frame over which the measurements are taken may be on the order of milliseconds or less while the random variations in the process may be on the order of seconds. respectively. In this latter case. (4-3. etc. the sample standard deviation for the variable can be calculated as
1 s X p s X Pp NP − 1
΄
͚ (XP j p1
N
P
j
− XP )
2
΅
1⁄
2
(6-1. from a sample of size N.3).3)
The appropriate random standard uncertainty of the mean for the current measurement (X) is then
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. or even days. the measurement of a variable may be only a single measurement or an average of measurements taken over a short time frame. repeatability of test conditions.1 General Case For X that is determined as the average of N measurements. Another situation where previous values of a variable would be useful is when a small sample size (N) is used to calculate the mean value (X) of a measurement. then it could be used to calculate a more appropriate standard deviation for the current measurement [4]. For example.1)
because X (based on the same sample of data) is used in the calculation of both quantities.
6-1. The sample standard deviation (sX ) and the random standard uncertainty of the mean (sX ) are calculated from eqs.1). then these additional error sources will have to be varied while the multiple data measurements are taken to determine the standard deviation. Typically. When NP previous values (XPj) are known for the quantity being measured.2 Using Previous Values of sX In some test situations. variation in test configuration.

3.3 Using Elemental Random Error Sources Another method of estimating the random standard uncertainty of the mean for a measurement is from information about the elemental random error sources in the entire measurement process. When all error sources have large sample sizes. (6-1.ASME PTC 19. then the method in para. In this case.5)
͚
K
(s Xk) 4 k
(6-1. 08/17/2010 06:03:02 MDT
.1 or 6-1. However.7)].1-2005
TEST UNCERTAINTY
sX p
sX
ΊN
(6-1. 4-3.
6-1. It should be noted that while b X is an estimate of the dispersion of the systematic errors in a measurement. previous information might not be available to estimate the sample standard deviation as discussed in para. the degrees of freedom () associated with the combined random standard uncertainty is calculated using the Welch-Satterthwaite formula [5] above [eq. The true systematic error () is unknown.2 depending upon which is appropriate.4)
p
(s Xk) 2 k͚ p1
kp1
K
2
where N is the number of current measurements averaged to determine X. the calculation of is unnecessary.2 as a value that quantifies the dispersion of the systematic error associated with the mean. If in each of the N measurements of the variable X the output of an elemental component is averaged Nk times to obtain Xk.6)
6-2 SYSTEMATIC STANDARD UNCERTAINTY OF A MEASUREMENT The systematic standard uncertainty b X of a measurement was defined in para. 6-1.
sX p
ΊN
΄ ͚ (s
K kp 1
) X
k
2
΅
1⁄
2
(6-1.6)]. The systematic standard uncertainty of the measurement is the root-sum-square of the elemental systematic standard uncertainties b Xk for all sources. Each of the elemental random standard uncertainties of the mean sXk is calculated using the methods described in para.2 would apply. If all the random standard uncertainties are expressed in terms of their contribution to the measurement. and each is assumed to be an ISO Type A estimate. an estimate of the sample standard deviation (sX) would be made using engineering judgment and the best available information.5) as appropriate. then the method in para.1) or (6-1. This type of uncertainty estimate would be an ISO Type B estimate. 6-1. The degrees of freedom for the estimated random standard uncertainty of the mean (sX) is dependent on the information used to determine each of the elemental random standard uncertainties of the mean and is calculated as
18
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
K
1⁄
2
bX p
(6-2. 6-1.
( b X ) 2΅ ΄ k͚ p1
k
where K is the total number of random error (or uncertainty) sources. If instead previous information is used to obtain sXk. when combining elemental random standard uncertainties of the mean by the root-sum-square method [see eq. (6-1. (6-1. 6-1.1)
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.4 Using Estimates of Sample Standard Deviation In a pretest uncertainty analysis.1 would be used.2 or 6-1.7)
This estimate of the random standard uncertainty is an ISO Type A estimate since it is obtained from data. then the random standard uncertainty for the measurement mean is the root-sum-square of the elemental random standard uncertainties of the mean from all sources divided by the square root of the number of current readings (N) averaged to determine X
1
where k is the appropriate degrees of freedom for sXk and is obtained from eq. but b X is evaluated so that it represents an estimate of the standard deviation of the distribution for the possible  values. The case where the data sample is only a single measurement is handled above with N p 1. 6-1. the systematic error that is present in a specific measurement is a fixed single value of . The number of degrees of freedom for this random standard uncertainty of the mean sX is
p NP − 1
(6-1. for small samples.

3).TEST UNCERTAINTY
ASME PTC 19. In this Supplement. 08/17/2010 06:03:02 MDT
. rather than by their source.1) all of the elemental systematic standard uncertainties are expressed in terms of their contribution to the measurement.`. such as the calibration process.Ap elemental systematic standard uncertainty calculated from data. (6-1. one should be very careful to identity all sources of fixed error in the measurement.. with this quantity defined as the combined standard uncertainty (uX ) [1].3). 4-3.. the larger or dominant ones will control the systematic uncertainty in the measurement. and fixed errors of method.. There can be many sources of systematic error in a measurement. with a defined level of confidence.````````. environmental effects.`. or (6-1. the ISO Guide [1] classifies uncertainties by source as either Type A or Type B.1-2005
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`.``. however.2)
where b Xp the systematic standard uncertainty [eq.. For each systematic error source in the measurement.-`-`. Also. as in a calibration process b Xk. then the method given in subsection 8-2 should be used to determine the systematic standard uncertainty of the measurement.`---
where Kp the total number of systematic error sources and each bXkp an estimate of the standard deviation of the kth elemental error source
Note that in eq. such as radiation effects in a temperature measurement. uncertainties are classified by their effect on the measurement. This Supplement recommends the following nomenclature for dual classification: b Xk.4). Such classifications would be useful in international test programs.
19
There may be situations when it is convenient to classify elemental uncertainties by both effect and source. (6-2.Ap elemental random standard uncertainty calculated from data s Xk.`. can cause systematic errors of method.`. transducer errors. Type A uncertainties are the calculated standard deviations obtained from data sets. Type B uncertainties are also given as standard deviation level estimates. This effect classification is chosen since most test operators are concerned with how errors in the test will affect the measurements.`..Bp elemental random standard uncertainty estimated from best available information
6-4 COMBINED STANDARD AND EXPANDED UNCERTAINTY OF A MEASUREMENT For simplicity of presentation.1)] s Xp the random standard uncertainty of the mean [eq. Sometimes previous data are available to make estimates of uncertainties that remain fixed during a test (and are therefore ISO Type A estimates). the elemental systematic standard uncertainty must be estimated from the best available information. Type B uncertainties are those that are estimated or approximated rather than being calculated from data.1)
represents a band about X within which the true value is expected to lie with a given level of confidence (see Fig. Usually these estimates are made using engineering judgment (and are therefore ISO Type B estimates).Bp elemental systematic standard uncertainty estimated from the best available information s Xk. (6-2. (4-3. 6-3 CLASSIFICATION OF UNCERTAINTY SOURCES As discussed in subsection 1-1. a single value is often preferred to express the estimate of the error between the mean value (X) and the true value. Because of the resulting effect of combining the elemental uncertainties in a root-sum-square manner.. There usually will be some elemental systematic standard uncertainties that will be dominant. instrument systematic errors. either random or systematic. The uncertainty interval is composed of both the systematic and random uncertainty components. If any of the elemental systematic uncertainties are nonsymmetrical.
uX p
Ί ( b X )2 + ( sX ) 2
(6-4. The interval
X ± UX
(6-4.6) as appropriate]
In order to express the uncertainty at a specified confidence level. the combined standard uncer-
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. The general form of the expression for determining the uncertainty of a measurement is the rootsum-square of the systematic and random standard uncertainties of the measurement.```.

75 84.3)
(a) Uncertainty Due to Random Error.14 85.11 84. various confidence levels may be appropriate. min
11 12 13 14 15
16 17 18 19 20
Measured Temp.72 85.50 85.07 84.3) as follows:
sX
where uX p the combined standard uncertainty [eq.89 85.20 85.39 84. °C
85.74 85.50 85.11 84. of the temperature measurements is determined using eq.2) as follows:
sX p
Ίj͚ p1
N
( Xj − X )2 p 0. A digital thermometer was used to measure the average temperature of a circulating water bath that is being used in an experiment. For other confidence levels see Nonmandatory Appendix B.22
85. min
21 22 23 24 25
26 27 28 29 30
Measured Temp.80 84.2)]
Example 6-4.1.23 85.43 84.56
tainty must be multiplied by an expansion factor taken as the appropriate Student’s t value for the required confidence level (see Nonmandatory Appendix B). The uncertainty for 95% confidence and large degrees of freedom (t p 2) is calculated by
UX p 2uX
(6-4. A t value of 1.. Depending on the application.72 85.39 84.ASME PTC 19.77 85.24
84. (4-3. This expansion factor of 2 is used for most engineering applications.28 85.28°C N−1
(3) The random standard uncertainty of the sample mean is determined using eq.
20
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.1-2005
TEST UNCERTAINTY
Table 6-4-1 Circulating Water Bath Temperature Measurements (Example 6-4. (4-3.04°C jp1
(2) The sample standard deviation is determined using eq.22
85. The uncertainty due to the random error of the average temperature measurement is evaluated as follows.79 85.18
Elapsed Time. The degrees of freedom used is a combined degrees of freedom based on the separate degrees of freedom for the random standard uncertainty and the elemental systematic standard uncertainties (see Nonmandatory Appendix B). 08/17/2010 06:03:02 MDT
sX p
ΊN
p 0..58 85. Temperature measurements were collected every minute resulting in a total of 31 data points as presented in Table 6-4-1. (1) The sample mean.00 85..
. (6-4.35 84. °C
85.96 (usually taken as 2) corresponds to large degrees of freedom and defines an interval with a level of confidence of approximately 95%. °C
85. (4-3. The experiment lasted a total of 30 min.05
Elapsed Time.12 85.05 84.1) as follows:
N
Xp
1 N
͚ Xj p 85. min
0 1 2 3 4
5 6 7 8 9 10
Measured Temp. The Student’s t is chosen on the basis of the level of confidence desired and the degrees of freedom.05°C
(b) Uncertainty Due to Systematic Error. or average value. The uncertainty due to the systematic error of the average circulating water bath temperature measurement is evaluated by (1) identifying all elemental sources of systematic error for the measurement.1)
Elapsed Time.

08/17/2010 06:03:02 MDT
.) on digital thermometer Effects of conduction and radiation heat transfer between the digital thermometer and the surroundings Uniformity of circulating water bath (spatial uncertainty) Total systematic uncertainty
Systematic Standard Uncertainty.05 0. and combining elemental systematic uncertainties.17°C
Therefore.1-2005
Table 6-4-2 Systematic Uncertainty of Average Circulating Water Bath Temperature Measurements (Example 6-4.07
(2) evaluating elemental systematic standard uncertainties as the standard deviations of the possible systematic error distributions. and (3) combining the elemental systematic standard uncertainties into an estimate of the total systematic standard uncertainty for the measurement. (6-4.005
0.2 and subsections 6-2.04 ± 0. For the purposes of this example. The expanded uncertainty of the average circulating water bath temperature measurement is evaluated using eqs.0005
0.05 0. 8-1.3) as follows:
U X p 2 Ί ( bX ) 2 + (s X ) 2 p 0.1)
Description of Systematic Uncertainty Source
Calibration of digital thermometer Environmental influences (ambient temperature. and 8-2 for further discussion of the process for identifying. humidity.TEST UNCERTAINTY
ASME PTC 19. 7-4.17°C
21
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. 4-3. evaluating.
(c) Expanded Uncertainty. °C
0. the true average temperature of the circulating water during the experiment is expected to lie within the following interval with 95% confidence: X ± UX p 85. a summary of this evaluation is presented in Table 6-4-2. etc. Refer to para.2) and (6-4.

a test design that averages multiple test results on one rig or for only one day may produce optimistic random uncertainty estimates compared to testing several rigs. The statistics found by combining these duplicated data sets allow for a reasonable estimate of the variations possible in the result that might be due to the control of test operating conditions. Whereas these influences might normally be considered systematic errors during repeated tests. The overall reported result will usually be combined to provide the mean of the multiple results. or test location. When tests are duplicated under similar but somehow changed operating conditions.ASME PTC 19.1)
where the subscript I signifies the total number of parameters involved in R. Ni will be 1 for a single data point or assigned value of a parameter. That is. R. In this case. is expressed in terms of the average or assigned values of the independent parameters (Xi ) that enter into the result. Historic data are invaluable for studying these effects. some of the parameters may be based on single measurements and others may be the
22
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
mean values based on Ni repetitions. A statistical technique called analysis of variance (ANOVA) is useful for partitioning the total variance by source. The list of test variation causes are many and may include the above plus environmental and test crew variations. Careful consideration should be given to designing the test series to average as many causes of variation as possible within cost constraints. The test design should be tailored to the specific situation.1. more basic parameters. If the pretest
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. R. The statistics from repeated tests allow for quantifying the expected variation in a parameter or in a result derived from parameters. such as temperature and pressure. these generate multiple data sets for the measured parameters. 7-1. This case is called a single test result. a set of parameters (Xi ) is measured and a single result.
R p f ( X1. are not usually measured directly. 08/17/2010 06:03:02 MDT
. . including using unmeasured properties. and the average values of the independent parameters are obtained as
1 Ni
Xi p
͚ Xi jp1
Ni
j
(7-1. if experience indicates that timeto-time and test apparatus-to-apparatus variations are significant. R. The estimation of the sample mean and standard deviation based on Ni measurements is calculated as described in para. the duplicated tests can randomize these systematic errors providing error estimates from the statistical variations in the combined data pool [6]. Ni can be different for each Xi.1 Single and Multiple Tests In some experimental situations. . such as in an estimate of efficiency. are either measured or assigned and the required result is calculated as a function of these parameters. . such as tabulated coefficients. Instead. Repeated tests are those run under the same conditions to estimate a parameter or a set of parameters (Xi ). For example. The effect of the propagation can be approximated by the Taylor series method (see Nonmandatory Appendix C). X I )
(7-1. 4-3. To estimate the uncertainty of a calculated result. is calculated. instrumentation. the result. X 2. or use of different test rigs..1-2005
TEST UNCERTAINTY
Section 7 Uncertainty of a Result
7-1 PROPAGATION OF MEASUREMENT UNCERTAINTIES INTO A RESULT Calculated results.2)
where Ni is the number of repeated measurements of Xi. each monitored several times over a period of several days. Uncertainties in these measurements or assigned values of the parameters are propagated to the result through the functional relationship between the result and the parameters.

3) and (7-2. Two approaches to estimating the sensitivity coefficient of a parameter are discussed in the following text.`.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.2) or (7-2.```. . then the absolute (dimensional) sensitivity coefficient (i) of the parameter Xi may be obtained by partial differentiation.`. . determined according to the methods presented in subsection 6-1. . The average result from more than one test is given by
and i′ by
⌬R R
⌬X i Xi
p
X i ⌬R R ⌬X i
i′ p
(7-2.`. X2.g.4). (7-2. respectively. .-`-`..4)
The result is calculated using Xi to obtain R [7].1-2005
uncertainty analysis (see subsection 4-4) identifies unacceptably large error sources.e. the value of ⌬Xi used should be as small as practical (i. 08/17/2010 06:03:02 MDT
.. XI ). 7-2 SENSITIVITY Sensitivity is the instantaneous rate of change in a result to a change in a parameter. the uncertainty of the average test result may be reduced from that for one test because of the reduction in the random uncertainty of the average. In this case.. However..
7-3. 7-2.``.2)
7-2. The derivatives and values for ⌬R are best estimated by use of central differencing methods.
23
R p
͚ Rm mp1
M
M
(7-3.1) or (7-2. large enough to keep truncation errors from influencing the calculations).3)
The symbols i and i′ are the absolute and relative sensitivity coefficients.1)
The relative random standard uncertainty of a result is
Analogously. . special tests to measure the effects of these sources should be considered. and its parameters (X1.3)
--`. .````````. if R p f (X1. the relative (nondimensional) sensitivity coefficient (i′) is
∂R R
∂X i
Xi p
X i ∂R R ∂X i
sR p R
i′ p
΄͚
ipi
′ i
sX i
Xi
΅
2
1⁄
2
(7-3. and sX i is the random standard uncertainty of the measured parameter average (Xi ). X2.. repeated tests).2 Multiple Tests: Repeated Tests When more than one test is conducted with the same instrument package (i. R. i is given by
i′ p ⌬R
⌬Xi
(7-2.2 Numerically Finite increments in a parameter also may be used to evaluate sensitivity using the data reduction calculation procedure.TEST UNCERTAINTY
ASME PTC 19. To best approximate the sensitivity that would be obtained analytically.
7-3 RANDOM STANDARD UNCERTAINTY OF A RESULT 7-3.. of eqs. systematic uncertainty will remain the same as for a single test...1 Single Test The absolute random standard uncertainty of a single test result may be determined from the propagation equation (see Nonmandatory Appendix C) as
sR p
΄͚
I
I
ipi
͑ i s X i͒
2
΅
1⁄
2
(7-3. then
i p
∂R
∂X i
(7-2. 8]..`. XI).e..`.
Thus..1)
Numerical differentiation is covered in various references [e.2)
(7-2.1 Analytically When there is a known mathematical relationship between the result.

(7-3. and alternate uncertainty equations is addressed in Nonmandatory Appendix B. The statistical equations used to assess variations between tests are similar to those discussed in para. With high-speed computing capabilities. 7-3. Monte Carlo methods have become popular for determining the test result uncertainty using the test input variable values and their associated uncertainties.1 The Magnitude of Uncertainty in the Test The magnitude of uncertainty in a test can be quantified by using the data reduction equations
24
΄͚
ip1
I
i′
bXi
i
X ΅
2
1⁄
2
(7-4. t95 p 2). respectively.1) for a single test result or from eq.2 assume the systematic uncertainties for the different parameters are independent. The degrees of freedom associated with sR is determined by p M − 1. 08/17/2010 06:03:02 MDT
. (7-3. Equations 7-4..`. error distributions.`. (7-4.5) for a multiple test result.``.5)
(7-5. 7-6 EXAMPLES OF UNCERTAINTY PROPAGATION 7-6.1)
ΊM
This random standard uncertainty of the result also has R pM − 1 degrees of freedom.`. (4-3. the average result and average standard deviations are given by combining the information of each test..1)
The relative systematic standard uncertainty of a result is
bR p R
The special cases of correlated systematic uncertainties and nonsymmetric systematic uncertainties are covered in subsections 8-1 and 8-2.. 7-5 COMBINED STANDARD UNCERTAINTY AND EXPANDED UNCERTAINTY OF A RESULT The general form of the expression for determining the combined standard uncertainty of a result is the root-sum-square of both the systematic and the random standard uncertainty of the result.1). The following simple expression for the combined standard uncertainty of a result applies in many cases:
u R p ͓ ( bR) 2 + ( S R)2͔
1⁄ 2
sR p
͚ (Rm − R )2 mp1
M−1
M
1⁄
2
(7-3. Treatment of uncertainty intervals with alternate confidence levels.ASME PTC 19. This factor can be modified as appropriate for other confidence levels and small degrees of freedom as discussed in Nonmandatory Appendix B.95 p 2 uR (7-5.-`-`.````````.e.95 (7-5. The expanded uncertainty in the result at approximately 95% confidence is given by
U R. The random standard uncertainty of the result is estimated directly from the sample standard deviation of the mean result from multiple tests and is
sR p
sR
(7-3.. the estimate of the standard deviation of the distribution of the results is
6-2).`.```..`---
where bR is obtained from eq.3 Multiple Tests: Combined Tests When tests are duplicated under similar but somehow changed operating conditions.2)
where the use of the factor of 2 assumes sufficiently large degrees of freedom for the 95% confidence level (i.1 and 7-4.1) and sR is obtained from either eq.3)
I
1⁄
2
bR p
(7-4.`. The interval within which the true result should lie with 95% confidence is given as
R ± U R.1-2005
TEST UNCERTAINTY
where M signifies the number of tests available.2)
The symbol b Xi is the systematic standard uncertainty of the measured parameter (see subsection
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. Following eq. 7-4 SYSTEMATIC STANDARD UNCERTAINTY OF A RESULT The absolute systematic standard uncertainty of a result may be determined from the propagation equation (see Nonmandatory Appendix C) as
(i bX ) ΅ ΄ i͚ p1
2
i
--`. 7-3.4)
where sR includes random errors within tests and variation between tests..2 but discussion of their interpretation is beyond the intention of this document and is presented in detail in reference [6]..

1 6. The guarded hot plate is used because it is relatively inexpensive and effective in providing the boundary conditions necessary to ensure one dimensional heat flow through the material.``. and L is the thickness of the material.1). we must know or estimate the measurement errors associated with the power.12E-04 2. P is the electrical power dissipated by the hot plate.`. Example 7-6.14 W/mK at 95% confidence. is determined as a function of the parameters using their nominal values and eq. Therefore.0012 0.1-1 lends insight into the magnitude of the contributions to the overall uncertainty.01 ± 0. 7-6. bX SX i i b X 2 i sX 2
i
i
i
i
Parameter Information (in Parameter Units)
Symbol
P L A ⌬T
Description
Power Length Crosssectional area Temperature difference
Units
W m m2 °C
Nominal Value
18.2 Ranking of Uncertainty Components Uncertainty analysis can be used to identify the largest contributors to the overall uncertainty.1-1 lists the independent parameters for this example and Table 7-6.14E-04 1.-`-`. temperature.1)
Independent Parameters
Uncertainty Contribution of Parameters to the Result (in Result Units Squared)
Absolute Absolute Absolute Systematic Absolute Systematic Random Standard Random Standard Standard Absolute Uncertainty Uncertainty Uncertainty. From the previous analysis.2: A test is to be conducted to determine the pressure loss coefficient across a
25
kp
P L A ⌬T
(7-6.30E-04 0.354
2.9E-05 1.. then appropriate corrective action to reduce the uncertainty can be taken.`. This enables designers and engineers to make pretest hardware and/or experimental methodology improvements.0013 0.26E-04
and estimates for each of the test parameters. Inspection of Table 7-6.4 0. ⌬ T is the temperature difference across the hot and
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. For steady state conditions.```.1-2005
Table 7-6. the conductivity was determined to be k p 3.092 2.4E-05 0.1).. R.1: Tests are often conducted using the guarded hot plate technique to determine the thermal conductivity of a material.1)
The result. Table 7-6.018 8.5
Ni
50 32 32 50
0.50E-04 1.`. To estimate the expected conductivity uncertainty.1). it is important to select the power setting such that measurements are at least at the midscale of the power supply range.`. and geometry. which itself is just the specific form of eq. Uncertainty. Sensitivity.`---
cold surfaces. (7-1.65 0.3 −167 −0.TEST UNCERTAINTY
ASME PTC 19.47E-03 6. 08/17/2010 06:03:02 MDT
.5
Standard Deviation
0. the one-dimensional form of Fourier’s law can be used to characterize the thermal conductivity of the test specimen. A is the cross-sectional area.1-2 the final results...
--`.1-1 Table of Data (Example 7-6.````````. (7-6.25E-04 6. Contribution. which in this example is given by the thermal conductivity k.`.4E-05 6. Contribution.025 0. In eq.1
0.. Example 7-6. The uncertainty in conductivity will be directly dependent on the method used to measure each parameter in the computation. and the pretest calculations show the best results will be on the order of 15%. (7-6. respectively.25E-03
2.071
0.67E-04 5. If for example the uncertainty in the results needs to be 10% or less. The uncertainty in power required to generate a constant heat flux will depend on the type of material being tested. This is important and useful because it provides insight into whether a particular test or methodology is feasible based on the acceptable order of magnitude in uncertainty..163 120.

Data from the loss coefficient test are summarized in Tables 7-6. 95
0.6 K NP p 50 NT p 50
where ⌬P is the pressure drop across the valve.3) 26
Following the procedure outlined in subsection 7-1. and the data are given as:
P p 101.47E-06 6.041
Symbol
K
Description
Conductivity
Units
W/mK
Calculated Value.137 at 95%
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.10E-04 2.12E-04 . i
0. where the uncertainty contribution from the gas constant was assumed negligible (9).0 6. The density uncertainty is treated as simply a systematic contribution in the loss coefficient calculation.013
1. Substituting into eq. In this example we will consider a round tube so A p D2/4.
Ni
50 32 32 .068
Expanded Uncertainty of the Result.
Absolute Sensitivity.027 kg/ m3 at 95% confidence.25E-05 2. the density and expanded uncertainty in density were determined to be p 1.. R
3.27E-04 1. p P/RT.0 0.1-2 Summary of Data (Example 7-6. (7-6.. i S X 2
i
5.. From the previous analysis.1)
Absolute Systematic Standard Uncertainty.2)
The Mach number for this flow is well below 0.ASME PTC 19.192
Standard Deviation
10. and the density is assumed to follow the ideal gas equation of state.
valve at a specified operating condition..2-2.41 2. where A is the cross-sectional area of the tube. i b X 2
i
Absolute Random Standard Uncertainty Contribution. a blower for cycling the working fluid (air).3.14
Table 7-6.35E-06 1.0 1.1-2005
TEST UNCERTAINTY
Table 7-6.0 −1. bR
0.433 ± 0.0 −247.20E-03 .202
Absolute Systematic Standard Uncertainty Contribution. the loss coefficient was determined to be K p 1. sR
0.2-1 Table of Data (Example 7-6..01
Absolute Random Standard Uncertainty.030 0. Measurements of the ambient air were made. 08/17/2010 06:03:02 MDT
. The apparatus consists of a closed loop tube containing a differential pressure gauge for measuring the pressure drop across the valve.2) gives
Kp
2 D4⌬ P 8 Q 2
(7-6.3 kPa T p 296. uR
0.055
Combined Standard Uncertainty.2 K sP p 8. SX
i
Symbol Description
⌬P D Q Pressure drop Diameter Volumetric flow rate Air density
Units
Pa m m3/s kg/m3
Nominal Value
230.. is the density of air and V is the average air velocity.0E-04 0.45E-04
7. From the definition of mass flow rate. UR.70E-04 1.
9.0 kPa sT p 1.85E-05 2.74E-03 .64E-05 1. bX
i
Uncertainty Contribution of Parameters to the Result (in Result Units Squared)
Absolute Random Standard Uncertainty. the velocity can be replaced with the volumetric flowrate Q through the expression Q p AV. The equation characterizing the loss coefficient is
Kp ⌬P 1 2 V 2
(7-6..0062 191.2-1 and 7-6..2)
Parameter Information
Parameter Information (in Parameter Units)
Absolute Systematic Standard Uncertainty.0116 1.192 ± 0. and a flow meter.

The analysis shows that random errors in the volumetric flow rate measurement and systematic errors in the pressure drop measurement are the largest contributors to the overall uncertainty in loss coefficient.137
Fig.053
Combined Standard Uncertainty of the Result. 95
0.
27
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. are shown and the pareto is ranked (left to right)
in terms of the combined standard uncertainty contribution of each parameter to the result. 08/17/2010 06:03:02 MDT
.2-2 Summary of Data (Example 7-6.TEST UNCERTAINTY
ASME PTC 19.1-2005
Table 7-6. 7-6. A pareto chart (see subsection A-4) of the systematic and random uncertainty contributions to the loss coefficient is given in Fig.2.068
Expanded Uncertainty of the Result. bR
0. as given by ͑isXi ͒2(uK )2.2)
Calculated Results
Absolute Systematic Standard Uncertainty.433
Absolute Random Standard Uncertainty. R
1.2 Pareto Chart of Systematic and Random Uncertainty Component Contributions to Combined Standard Uncertainty confidence. 76. The relative contributions of the systematic uncertainty. uR
0. and of the random uncertainty.043
Symbol
K
Description
Loss coefficient
Units
Dimensionless
Calculated Value. sR
0. UR. as given by ͑ibXi ͒2(uK )2.

1). (8-1. X 2 .2) must be properly interpreted. Xi and Xk. For instance.. are from the same
28
bip systematic standard uncertainty in parameter i bikp covariance between the systematic standard uncertainties for the ith and kth parameters. 08/17/2010 06:03:02 MDT
.ASME PTC 19. term represents the sum of the products of the portions of bXi and bXk that arise from the same source and are therefore perfectly correlated [10]. Each bXiXk. bXiXk. and X 3 .5)
Ip number of parameters i and kp indexes indicating the ith and kth parameters p sensitivity coefficients
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.`.`---
The first three terms under the square root in eq. The covariance terms in eq.2)] assume that the systematic standard uncertainties of the measured parameters are all independent of each other. X 3 )
(8-1.`. The terms bXiXk are the estimates of the covariances of the systematic errors in Xi and Xk (see Nonmandatory Appendix B). and the last three terms are those that account for the correlation among the systematic standard uncertainties in X 1 .2) is
and the absolute systematic standard uncertainty of the result is given as
b R p ͓͑ 1 b X 1͒ 2 + ͑ 2 b X 2͒ 2 + ͑ 3 b X 3͒2
+2 1 2 bX 1 X 2 + 2 1 3 bX 1 X 3
+ 2 2 3 bX 2 X 3 ͔
1⁄ 2
͚ (i bi)2 + 2 i͚ ͚ ik bik ip1 p 1 k p i +1
I
I-1
I
(8-1. (8-1.1) and (7-4. some of the systematic errors are said to be correlated and these nonindependent errors must be considered in the determination of the systematic standard uncertainty of the result [9]. or thus their measurement errors are no longer independent.1)
bR p
source making them correlated. if elemental systematic standard uncertainties 1 and 2 for parameters 2 and 3 were from a common source.4)
where
(8-1. The result is calculated as
R p f (X 1 .2) are the same as those obtained by using eq. There are many situations where systematic errors for some of the parameters in a result are not independent.```. another form of eq. calculated as follows:
b ik p
͚ bi bk lp1
l
L
l
(8-1. Therefore. In these cases..-`-`. These terms must be included when systematic standard uncertainties for separate parameters. (7-4.2) can be expanded to any number of parameters by including the term for each pair of parameters that has correlated systematic standard uncertainties. (8-1.`.``. Consider an example where the result (R) is determined from three parameters (X 1 . X 3 ) that have correlated systematic errors. (7-4. X 2 . X 2 .. would be determined as
b X2 X 3 p b X 2 + b X2 b X 3
1
2
(8-1.1-2005
TEST UNCERTAINTY
Section 8 Additional Uncertainty Considerations
8-1 CORRELATED SYSTEMATIC STANDARD UNCERTAINTIES The expressions for the systematic standard uncertainty of the result in subsection 7-4 [eqs.````````. then bX2X3.`.`.. are the product of the units of Xi and Xk. The units of the correlation terms (covariances). Examples would include using the same apparatus to measure different parameters or calibrating different parameters against the same standard.. (8-1.3)
2
The example in eq.2)
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`..

7
Systematic Standard Uncertainty. then for the balance check to be satisfied the absolute value of z must be less than or equal to the uncertainty in z
| z | ≤ 2 bz
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.`. Program 1 (no correlated systematic uncertainties): 52.0 .0
Improved Design.025)(10 Pa) (0. 42.````````.. the same pressure transducer was used for both tests.0238 (106 Pa)−1 Rp
6 −1 6 2 b2 Rp [(−0. the systematic standard uncertainty was the same and was correlated for the two test measurements. Consider a burst test for an improved container design. bP.5)(10 Pa)]2 + 2(−0.30 40.0250 (106 Pa)−1 Pn 6 −1 6 2 b2 Rp [(−0. The improvement in the design can be expressed as the fraction
Rp Pn Pb
This example demonstrates the strength of the back-to-back testing technique using the same instrumentation. therefore..5)(106 Pa)(0.0238)(106 Pa)−1 (0.5)(10 Pa)] 6 −1 6 + [(0. Example 8-1-2 (adapted from [9]): Consider the piping arrangement shown with the four flowmeters:
where Pnp the burst pressure of the new design Pbp the burst pressure of the original or base design Table 8-1 provides burst tests for two different programs. There were no correlated systematic standard uncertainties common between these two transducers.5
R p 0.0238)(10 Pa) (0. 106 Pa
Program 1 Meter #1 Meter #2 Program 2 Meter #3 40.`.0 p 1.0036
where Lp number of common (correlated) error sources lp an index Example 8-1-1: The use of back-to-back tests is an excellent method to reduce the systematic standard uncertainty when comparing two or more designs.`.`.0325 (106 Pa)−1 Pb Rp
29
--`.5)(106 Pa) bRp 0. Pn..2)(10 Pa)] 6 −1 6 2 + [(0.30 42.0310 (106 Pa)−1 np 0. 106 Pa
0..0 bp −0. Even though the pressure transducer in Program 2 had a systematic standard uncertainty of more than twice those of the transducers in Program 1.7 p 1.`---
From conservation of mass.1-2005
Table 8-1 Burst Pressures (Example 8-1-1)
Base Design.`.0 54.. In the second program.2)(10 Pa)] bRp 0. This method is a special case of correlated systematic standard uncertainties.``.2 0..... 08/17/2010 06:03:02 MDT
.0310)(10 Pa) (0. the systematic standard uncertainty of the result for Program 2 was less than half of that for Program 1. a balance check would yield
z p m 4 − m1 − m 2 − m 3 p 0
If the errors in the flow rate measurements are predominantly systematic.2 0.-`-`. P b.```. 52.0 −R bp p −0.0310)(106 Pa)−1 (0. In the first test program.TEST UNCERTAINTY
ASME PTC 19.0082
np
Program 2 (correlated systematic uncertainties): 54.0325)(10 Pa) (0.. different pressure transducers were used in the tests on the two designs. 106 Pa
.

`.5) + (0. (8-1.5 kg/h for the three small meters and ±4. and 3 are calibrated against the same standard.3 kg/h
1⁄ 2
and the balance check will be satisfied if
| z | ≤ 2 bz p 10. the systematic standard uncertainties are
b m1 p bm2 p bm3 p 1.. the systematic standard uncertainty for z is larger than if all the meters had been calibrated against different standards (Case 1).6 kg/h
2 2
1⁄ 2
or
b z p ͓͑ bm1 ͒ 2 + ͑ bm2 ͒ 2 + ͑ bm3 ͒ 2 + ͑ b m4 ͒ 2 + 2 b m 1m 2 + 2 b m 1m 3 + 2 b m 2m 3 ͔ b z p 6.. 3. and 3 Calibrated Against the Same Standard and Flowmeter 4 Calibrated Against a Different Standard. 2.1).5) + (0. the systematic standard uncertainty for the three small flowmeters that originates from the calibration standard is correlated. In this case.5 kg/h for the large meter. Case 3: Flowmeters 1. In this case. Note that for this example the derivatives for eq. (8-1.`. all of the systematic standard uncertainties are uncorrelated with the systematic standard uncertainty for the three small flowmeters determined as
b mi ( i p 1. the systematic standard uncertainty for z becomes
b z p ͓͑ m1 bm1 ͒ 2 + ͑ m2 b m2 ͒ 2 + ͑ m3 b m3 ͒ 2 + ͑ m4 b m4 ͒ 2 + 2 m 1 m 2 b m 1 m2 + 2 m 1 m 3 b m 1 m3 + 2 m 2 m 3 b m 2m 3 ͔
1⁄ 2
Case 1: Each Flowmeter Calibrated Against a Different Standard.2) are
m1 p m2 p m3 p −1
because it is due to the scatter in the calibration line. The calibration standard systematic standard uncertainty for each flowmeter is ±1.5 kg/h)(1.6 kg/h b m4 p 4.`.5 kg/h
--`. The systematic standard uncertainty from their curve fits is not. 2.`. the systematic standard uncertainty for z is
b z p ͓͑ m1 bm1 ͒ 2 + ͑ m2 b m2 ͒ 2 + ͑ m3 b m3 ͒ 2 + ͑ m4 b m4 ͒ 2 ͔
1⁄ 2
or
b z p ͓͑ bm1 ͒2 + ͑ bm2 ͒ 2 + ͑ bm3 ͒ 2 + ͑ b m4 ͒ 2 ͔ p 5.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.5 kg/h
and
b m1m2 p b m1m3 p b m2 m3 p (1. m2.. If flowmeters 1.5) ͔ p 1. and 4 Calibrated Against the Same Standard.ASME PTC 19.9 kg/h
Using eq.5 kg/h.```. 3) p ͓ (1..`.1-2005
TEST UNCERTAINTY
Consider the case where the dominant systematic errors are from the calibration standard and from the calibration curve fit. 2.````````. The final uncertainty is obtained as follows:
b m1 p bm2 p bm3 p 1..4) for four parameters with three of them having correlated systematic standard uncertainties.5 kg/h)
and
m4 p 1
Using eq.``..
30
and
b m4 p 4. and flowmeter 4 is calibrated against a different standard. 2. (7-4.6 kg/h
Case 2: Flowmeters 1.6 kg/h
Note that in this case the signs for all the correlated terms are positive because all of the derivatives of z with respect to m1.4 kg/h
1⁄ 2
and the systematic standard uncertainty for the large flowmeter calculated as
b m4 p ͓(4. 08/17/2010 06:03:02 MDT
.-`-`.5) ͔ p 4. In this case. and m3 are negative.5 kg/h
2 2
1⁄ 2
and the balance check will be satisfied if
| z | ≤ 2 bz p 12. The curve fit systematic standard uncertainty for each meter is ±0.

````````. the following procedure should be employed for constructing a nonsymmetric uncertainty interval for the unknown true value of the quantity being measured (see Fig.1 should be used.1-2005
with
b m1m2 p b m1m3 p b m2m3 p (1.0 kg/h
1⁄ 2
and the balance check will be satisfied if
| z | ≤ 2 bz p 2... calibrating all the flowmeters against the same standard will yield the minimum systematic standard uncertainty for z. as the difference between the center of the interval specified in (a) and the measured value.5 kg/h) (4.-`-`. For this case.```. correlated systematic standard uncertainties can either decrease. all with correlated systematic standard uncertainties. 8-2.`.1 Nonsymmetric Systematic Uncertainty Interval for a True Value If the distribution of the systematic error associated with a measured variable is symmetrically distributed but not centered at zero.`.. in the absence of random errors. as
B 2
bX p
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.`.0 kg/h
qp
(c) Define the quantity B as follows:
Note the signs for each of the correlated terms.
and
b m1m4 p b m2m4 p b m3m4 p (1. q.``. In this case.. increase. This interval accounts for systematic errors only.5 kg/h)
8-2 NONSYMMETRIC SYSTEMATIC UNCERTAINTY In some experiments. X + B+ ) relative to the measured value X within which one may expect the true value to fall with 95% confidence.4) for four parameters.`---
Bp
( X + B+ ) − ( X − B−) B + + B− p 2 2
Thus B is equal to one-half the length of the interval specified in (a). (8-1. depending on the form of the data reduction equation and on which parameters have correlated systematic errors.
31
--`.`.. (b) Define the offset. In general. It is interesting to note that the systematic standard uncertainty in z is less than the smallest estimate of systematic standard uncertainty for any of the flowmeters. the systematic standard uncertainty for z is
b z p ͓͑ m1 bm1 ͒ 2 + ͑ m2 b m2 ͒ 2 + ͑ m3 b m3 ͒ 2 + ͑ m4 b m4 ͒ 2 + 2 m1m2 bm1m2 + 2 m 1 m 3 b m 1 m3 + 2 m 1 m 4 b m 1 m4 + 2 m 2 m 3 b m 2 m3 + 2 m 2 m 4 b m 2 m4 + 2 m 3 m 4 b m 3 m4 ͔
1⁄ 2
8-2. but if not then the method of para. 8-2.`. This paragraph presents a method for determining nonsymmetric uncertainty intervals in these cases [11].5 kg/h)(1..1): (a) Specify an interval (X − B−. then the overall uncertainty interval for the unknown true value will not be centered on the measured value of the variable. (d) Calculate b X. Thus
( X + B+ ) + ( X − B−) B+ − B − −Xp 2 2
or
b z p ͓͑ bm1 ͒ 2 + ͑ bm2 ͒ 2 + ͑ bm3 ͒ 2 + ͑ b m4 ͒ 2 + 2b m1m2 + 2b m1m3 − 2 b m1m4 + 2 b m 2m 3 − 2 b m 2m 4 − 2 b m 3m 4 ͔ b z p 1. or have no effect on the systematic standard uncertainty of the result. If this can be done then it should be. physical models can be used to essentially replace the asymmetric uncertainties with symmetric uncertainties in additional experimental variables.TEST UNCERTAINTY
ASME PTC 19.5 kg/h)
Using eq. 08/17/2010 06:03:02 MDT
. the systematic standard uncertainty for the measurement.

``. but wishes to account for them in an uncertainty analysis. the user finds that X p 534. the difference between the center of the interval specified in (a) and the value measured with the thermocouple.`.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.. using the standard formula uX p Ίb 2 + s 2 . the user of the thermocouple believes that the true gas temperature falls between the average measured with the thermocouple.7°C)/2 − 534. the combined standard uncertainty for the measurement.7°C.7°C.7°C and sX p 2. So B− p 0°C and B+ p 10°C.1: Suppose a thermocouple is being used to measure the temperature of a gas stream. 8-2.```.ASME PTC 19. 544.`.4)
Example 8-2. X p 534. using
2 2 U 95 p 2 uX p 2 Ί b X + sX
(8-2.. The user does not have enough information to properly correct the thermocouple reading for these effects.`.`...) (g) Calculate an approximate 95% confidence interval for the true value using
͓X + q͔ ± U95
(8-2.7°C p 5°C
--`.
32
q p (544. i. then a nonsymmetric confidence interval accounting for this nonsymmetric systematic uncertainty may be computed as follows: (a) Specify an interval (corresponding to 95% confidence) for the systematic error in question. X X (f) Calculate U95. From a sample of more than 30 readings using the thermocouple.7°C + 534.1)
This calculation is based on the assumption that the degrees of freedom for the combined standard uncertainty are large.e. (For small degrees of freedom see Nonmandatory Appendix B. (See Nonmandatory Appendix B for other distributional models.-`-`... the expanded uncertainty for the measurement. (b) Determine q. If the user believes that the true gas temperature may be as much as 10°C higher than X due to radiation effects.4°C.1-2005
TEST UNCERTAINTY
Fig.`.````````. In this case.1 Schematic Relation Between Parameters Characterizing Nonsymmetric Uncertainty
This is based on the assumption that a Gaussian distribution is an appropriate model for the systematic error. but the user of the thermocouple believes there may be a tendency for the thermocouple to provide a temperature reading that is lower than the actual gas temperature due to a radiative heat transfer mechanism.) (e) Calculate u X. and a value that is 10°C higher than X. 08/17/2010 06:03:02 MDT
..2)
(h) Express the final result as an asymmetric 95% confidence interval for the true value with the lower limit given by
X lower limit p X + q − U 95 p X − U−
(8-2.3)
and the upper limit given by
X upper limit p X + q + U 95 p X + U+
(8-2. In this case
where U− p U95 − q and U+ p U95 + q.

X n + qn͒ ± U95. ..7°C + 11.6°C
(d) Calculate b X. the systematic standard uncertainty for the measurement. . this 95% confidence interval is given by
(c) Determine the sensitivity coefficient.6)
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.9°C + 5°C p 11. using the standard formula
uX p
2 2 + sX p Ί (2.TEST UNCERTAINTY
ASME PTC 19. . .) (f) Calculate an approximate 95% confidence interval for the derived result using
r ͑X 1 + q1 . Xn + q n͒
− r͑ X 1 . then it should be evaluated at the point ͑X1 + q1 . i... X 2 + q 2 . 8-2. .4°C) 2 p 3. ..9°C − 5°C p 1. .45°C Ίb X
(f) Calculate U95.2 Nonsymmetric Systematic Uncertainty Interval for a Derived Result A nonsymmetric systematic uncertainty in a measured variable may also result in a nonsymmetric uncertainty interval for a derived result. Xn ͒. and qi for each average Xi that contributes to the determination of the derived result.
This is based on the assumption that a Gaussian distribution is an appropriate model for the systematic error. X2 . X 2 + q2 . . For small degrees of freedom. . . using
U 95 p 2 u X p 6. (For small degrees of freedom see Nonmandatory Appendix B). (b) Determine the offset.8°C
33
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
(This is based on the assumption that the degrees of freedom are large. X2 + q2. .9°C
This calculation is based on the assumption that the degrees of freedom for the combined standard uncertainty are large.9°C p 532. .9°C
(e) Calculate U95.9°C
The final result may be expressed as an asymmetric 95% confidence interval for the true value using the lower limit given by
X lower limit p X − U− p 534. X n͒ . the combined standard uncertainty for the derived result. qr . (e) Calculate uX.2): (a) Determine Xi . the expanded uncertainty for the measurement. .9°C p 546.. X 2 . the combined standard uncertainty for the measurement. X2 . (d) Calculate ur . see Nonmandatory Appendix B. i. which is defined as
q r p r ͑ X1 + q 1 . as
U 95. . as
bX p
B p 2.1-2005
(c) Calculate the quantity B as
Bp B + + B− 10°C + 0°C p p 5°C 2 2
and the upper limit given by
X upper limit p X + U+ p 534. . .7°C − 1. Xn + qn͒.e.5°C) 2 + (2. . Xn͒. . The following procedure may be employed for propagating the nonsymmetric uncertainties in a set of measured variables to a derived result (see Fig. .7°C + 5°C] ± 6. (g) Calculate an approximate 95% confidence interval for the true value using
͓X + q͔ ± U95
In this case.5°C 2
8-2. uXi . r p 2 u r
and
U + p U 95 + q p 6.9°C
(h) Calculate
U − p U 95 − q p 6. i p i ͑X1 . using the standard formula:
ur p
Ί͓͑1uX ͒2 + ͑2uX ͒2 + . . If a sensitivity coefficient depends on the values of any averages. r ͑X1 . + ͑nuX ͒2͔
1
2
n
(8-2. r the expanded uncertainty for the derived result at a 95% confidence level. . for each average X.5)
[534. that contributes to the derived result following standard procedure..r (8-2. 08/17/2010 06:03:02 MDT
.

. and R.2: Suppose the user of the thermocouple in the example in para.r + q r) p r + U r+ (8-2. In this case.7 K )]
1⁄
2
p [ kR ]
1⁄
2
(0. In this case. X 2 . .1-2005
TEST UNCERTAINTY
Fig. for the measured variable T.0604 K)
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. qc . .0175 K−
1⁄ 2
1⁄
2
(3.7K. In this case. .0175 K) −
1⁄
2
(d) Estimate the combined standard uncertainty for the derived result. uT . . uc.2 Relation Between Parameters Characterizing Nonsymmetric Uncertainty
(g) Express this confidence interval as an asymmetric 95% confidence interval for the derived result as follows:
r ͑X 1 . it should be evaluated at T + qT p 812. 8-2. 8-2.
u c p [{(kR )
1⁄ 2
where k.7 K )]
1⁄ 2
and the upper limit on this interval is given by
r upper limit p r ( X 1 . T p (X + 273. . and qT p 5K. X 2 .ASME PTC 19.7K. In this case. 08/17/2010 06:03:02 MDT
.8)
negligible uncertainty and T is the measured value of the absolute temperature.9)
− [kR (807. the ratio of specific heats. Example 8-2. X n)
+ (U 95. so that here
T p ( 1 ⁄2)[ kR/ (812. . T p 1 (1/2)[kR/T] ⁄2. .7)
where the lower limit on this interval is given by
r lower limit p r ( X 1 . and qT for the measured variable T. The uncertainty interval for c may be calculated as follows: (a) Determine T.r + qr.1 wishes to use this gas temperature to estimate the speed of sound for the gas using the following relation:
c p [ kRT]
1⁄ 2
(c) Determine the sensitivity coefficient. T p 807. are taken to be constant with
34
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
(0. . as follows:
q c p c ( T + q T) − c(T ) p [kR (812. X n͒ ± ͑U 95. Since this sensitivity coefficient depends on T. where X is the average value of the temperature of the gas using the thermocouple. uT p 3. the gas constant for the gas.45K.r − q r ) p r − U r− (8-2.7 K )]
1⁄ 2
p [ kR ]
1⁄
2
(0. T . (b) Determine the offset.0878 K )
1⁄
2
with Ur− p U95.r − qr and U + r p U95. X n)
− (U 95. .45 K )} 2]
1⁄ 2
1⁄
2
p [kR ]
(0. .2)K. . X 2 .r ± q r͒
(8-2.

an estimated standard deviation of the random error components of the calibration process scp sc ΊNc.`. .c ± q c)
In this case. For example. Calibration errors are frequently reclassified as a result of the Defined Measurement Process.-`-`.TEST UNCERTAINTY
ASME PTC 19. .12% below to 0.7 K ]
1⁄
2
± {[ kR]
1⁄
2
(0. calibrations performed only once during a test program. . the calibration process random error does not have an opportunity to introduce scatter in the test data set and therefore should be reclassified (fossilized) as systematic error during the final uncertainty analysis.```.1208 K 2)
1⁄
1⁄
± [kR ] 2 (0.1208 K)
1⁄
2
(This is based on the assumption that the degrees of freedom are large. as
U 95.c the expanded uncertainty for the derived result c.033 K 2)
1⁄
and whose upper limit is equal to
c upper limit p [ kR]
1⁄ 2
[28.1-2005
(e) Calculate U95.7 K ]
1⁄
2
± [ kR ]
1⁄
2
(0. Final classification is dependent on the Defined Measurement Process.c. As an example. this 95% confidence interval is given by
[ kR ]
1⁄ 2
calibrations (i... for test measurements involving a single calibration (i.73% above the value for the speed of sound assessed using the measured value of the temperature.. the fossilized elemental systematic standard uncertainty due to the random error of the calibration process A more general form of eq.`. However.. For small degrees of freedom. b2.1) for a single calibration is
* bc p ͓ (b c) 2 + ( s c)2͔
1⁄ 2
(8-3. (8-3. at a 95% confidence level.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. . for test measurements involving multiple
35
b* cp total systematic standard uncertainty of the single calibration process b1. 08/17/2010 06:03:02 MDT
.42K 2 ] − [kR ]
1⁄
(0. ( s c ) 2
(8-3.c p 2 uc p 2[kR ]
1⁄ 2
(0.e. see Nonmandatory Appendix B.. the calibration process random error will introduce scatter in the test data set and therefore should remain classified as a random error source or random standard uncertainty in the uncertainty analysis. In this case. the calibration random standard uncertainty should be treated as another elemental systematic standard uncertainty and combined with the other calibration systematic standard uncertainties to obtain the total systematic standard uncertainty for the calibration process as:
* bc p
[807.`. calibrations performed prior to each of the multiple tests or periodically throughout a test program).`.2086 K 2)
1⁄
In this example. If they do not. 8-3 FOSSILIZATION OF CALIBRATIONS Definition of the Measurement Process is a prerequisite for determining measurement uncertainty estimates. .1)
whose lower limit is equal to
c lower limit p [ kR]
1⁄ 2
where
1⁄ 2
[28. this 95% confidence interval is given by
[ kR]
1⁄ 2
[812.. the uncertainty interval for the speed of sound of the gas extends from 0. Elemental errors are classified as random if they add scatter to a result.2)
where bcp calibration process initial systematic standard uncertainty before reclassifying (fossilizing) the calibration random standard uncertainty into systematic standard uncertainty
--`. and all data samples processed using the same calibration constants).42K 2 ] + [kR ]
1⁄
1⁄
2
(0.0604 K)
1⁄
2
p [kR ]
1⁄
2
(0.p elemental systematic standard uncertainty components of the calibration process Ncp number of calibration points within a single calibration scp standard deviation of the calibration process. different Defined Measurement Processes for the same test will result in different estimates of measurement uncertainty.1208K 2)
1⁄
(g) Express the final result as an asymmetric 95% confidence interval using
c (T ) ± (U 95.````````..) (f) Compute a 95% confidence interval for the derived result using c(T + qT) ± U95.`. In this case.``. they are systematic.0878 K 2)}
1⁄
Ί ( b 1) 2 + ( b 2 ) 2 + ..e.

A simple illustration of the impact of spatial variation is the measurement of average velocity in fully developed flow of an incompressible fluid in a pipe.
36
b c* p ( b c) + ͑ sc / Ί N rc͒
΄
2
΅
1⁄
2
(8-3.. This is in contrast to the situation at high Reynolds numbers.2) should be reduced by dividing by ΊNrc. it raises a problem in determining performance level. Spatial variation effects are considered errors of method (see para. the calibration process random standard uncertainty is common to all data samples and thus manifests itself in the data set as a systematic standard uncertainty (becomes fossilized). 5-2.3)
where Nrcp number of repeated independent (single) calibrations averaged and used in obtaining a single set of calibration constants common to all samples within the test data set
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. While this point characteristic may be useful for other purposes.1-2005
TEST UNCERTAINTY
To illustrate the fossilization process. yields a point measurement of a parameter rather than an average measurement. An example of an averaged output is an averaging pressure probe often used to measure average velocity at a cross section of pipe. In many instances. such as installing multiple sensors and averaging the outputs. The calibration process random standard uncertainty is a function of the random standard uncertainties in both the master and test meters. The magnitude has been reduced by having averaged multiple calibrations. the random and systematic standard uncertainties of the calibration process should be combined and carried forward as a systematic standard uncertainty. the velocity at the center of the pipe is twice the average velocity. Thus. a portion of the random standard uncertainty still becomes fossilized into systematic standard uncertainty.g. the flow is laminar and has a parabolic velocity profile. In this case. For example. the calibration process random standard uncertainty will cause scatter in the individual test meter data samples and thus should remain classified as calibration random standard uncertainty in the uncertainty analysis. Therefore. when a data set from the test meter involves only a single calibration.2) would be
8-4 SPATIAL VARIATION [12] Measurement requirements for a performance test are often such that an average measurement of individual parameters is required. Thus a more general form of eq. In this case. multiple calibrations are performed but the results are averaged into a single set of calibration constants for use in processing all data samples (e. The term sc in eq. (8-3. Usually. pre. When a data set of interest from the test meter involves multiple calibrations.ASME PTC 19. At higher Reynolds numbers. An additional systematic standard uncertainty may need to be assigned to the mean result to account for the possible difference between the determined mean and the true mean. Most instrumentation. depending on the defined calibration. consider a master flowmeter installed in line with a test flowmeter for use in establishing a calibration correction for the test meter. however. or both. it often is necessary to install several measurement sensors at different spatial locations to account for spatial variations of the parameter being measured. At low Reynolds numbers. 8-5 ANALYSIS OF REDUNDANT MEANS When redundant instrumentation or calculation methods are available. test conditions vary between these two extremes and it is not possible to correct the readings in a simple manner. the profile approaches uniformity and any measurement will yield a reasonable estimate of the average velocity. the systematic standard uncertainty of the test meter is replaced by that of the master meter. As a result of the calibration process. (8-3. The uncertainty of this mean velocity would be calculated by considering it to be a determined result. Circumferential variations may also be present. making the point measurement inadequate. Fossilization of calibration random standard uncertainty can occur at any or all levels of calibration hierarchy. from a national laboratory to the test application.and post-calibrations). 08/17/2010 06:03:02 MDT
. One measurement will not give the true average value directly. However.5). where the flow is turbulent. other approaches are chosen. or measurement processes. Sometimes. the quantity measured varies in space. the individual results and their uncertainties should be compared with each other and with the pretest uncertainty analysis.

the true value must lie in the region where the uncertainty intervals overlap.1 need to be considered. respectively. Investigation to identify bad readings. as in Case 2. or the true value is not constant. The process is outlined in the following paragraphs. the larger the overlap. For both measurements and both uncertainty intervals to be correct. A-2 may be used to obtain a better estimate of the true value than either of the individual measurements. where a partial overlap of the uncertainty intervals exists. the two measurements are consistent with one another) is given by
Zp
X 1 − X2
2 2 ͓ + s2 2 + b 1 + b2͔
s2 1
1⁄
2
If |Z| > 2. 8-5. then the data indicate. Assuming that the degrees of freedom associated with the systematic and random uncertainty components are large.1-2005
Fig. Likewise. Case 3: Case 3. the random and systematic standard uncertainties associated with X1. As the difference between the two measurements increases. Case 2: When the uncertainty intervals completely overlap. Case 1: A problem clearly exists when there is no overlap between uncertainty intervals. the overlap region shrinks. the weighting method described in para. the more confidence we have in the validity of the
37
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
measurements and the estimate of the uncertainty intervals.1 Three Posttest Cases
When comparing redundant means (X1 and X2) and their uncertainty intervals. let s2 and b2 denote.e. one can be reasonably confident that there has been a proper accounting of all major uncertainty components. there is
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. at the 95% confidence level. is the most difficult to analyze. Otherwise. is necessary to resolve this discrepancy. that the two measurements are consistent with one another. overlooked or underestimated systematic uncertainty. Standard statistical hypothesis testing may be used to evaluate the significance of the difference observed. the random and systematic standard uncertainties associated with X2. etc. respectively. and that there are no correlated errors.. Either uncertainty intervals have been grossly underestimated. 8-5.. the three cases illustrated in Fig. 08/17/2010 06:03:02 MDT
. Let s1 and b1 denote. Consequently.TEST UNCERTAINTY
ASME PTC 19. the Z-statistic for testing the null hypothesis that the two measurements X1 and X2 have an expected difference of zero (i. The smaller uncertainty interval X2 ± U2 is wholly contained in the larger interval X1 ± U1 Since the individual measurements are valid.

see Refs.. the curve-fit expression is
ˆ p mX + c Y (8-6.6 will cover only straight-line regressions to estimate the relationship for Y versus X.3).1-2005
TEST UNCERTAINTY
no compelling evidence to declare the two measurements to be inconsistent with one another. or a simple linear regression.1) along with several values of X and the resulting straight line can be plotted ˆ vs. This mean. [9. and in the representation of the final test results. only a special case is considered for the systematic standard uncertainty. ˆ . 14. and correlated systematic standard uncertainties between X and Y. Subsection 8.1) represents the relationship between the mean value of Y and X. In other cases. in the data reduction program. 15. Yj).2 Least-Squares For a straight-line. This special case is where the systematic standard uncertainty for the Y values and/or the X values is a constant (i. (8-6. [9. 16.
Caution: Lack of evidence for demonstrating inconsistency does not “prove” consistency.1) 38
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
For a given value of X. Yj) data. The random standard uncertainty for the curvefit will be determined using standard least squares analysis [9. Leastsquares regression analysis is the most popular means of curve fitting. Once the slope and intercept are calculated from eqs. see Refs. where the assumption is made that there is no random standard uncertainty in the X values and the random standard uncertainty in the Y values is constant over the range of the curve-fit. or a simple (first-order) linear regression. percent of full scale) and there are no correlated elemental systematic standard uncertainties between the X and Y values. the random standard ˆ obtained from uncertainty associated with the Y the curve-fit [eq. variable systematic standard uncertainties in X and Y. Since the Y a mean value for the data set. 14]. is not the average of the Y data but the mean Y j Y response from the curve-fit for a given X. In many cases. 08/17/2010 06:03:02 MDT
.1)] is
1 ˆ p SEE sY + N
΄
(X − X )2
jp1
͚
N
(X j − X) 2
΅
1⁄
2
(8-6. ˆ 8-6.3 Random Standard Uncertainty for Y Determined From Regression Equation The statistic that defines the standard deviation for a straight-line curve-fit is the standard error of estimate
SEE p
΄
͚ (Yj − mXj − c)2 jp1
N−2
N
΅
1⁄
2
(8-6. 13. the anticipated representation of the data is a straight line. In this section. or transformed.e. 17.4)
8-6. A more general approach to regression uncertainty is presented in [9] where the methodology applies for variable random standard uncertainties in X and Y. the curve should be a good representation of the data if the simple linear fit is appropriate. (8-6. 18].1 Linear Regression Analysis Curve fitting often is used in the calibration process. 16]. into linear coordinates [9. these constants can be substituted into eq. the data to be curve-fit often can be rectified.3)
The least-squares process essentially provides an average for the data so that the regression expression in eq. (8-6.
where for N data pairs (Xj.5)
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.2)
Xj
and the intercept c is determined from
j ) ͚ Y j − ͚ X j ͚ (X jYj) ͚ ( X2 jp1 jp1 jp1 jp1 N N N N
cp
N͚
N
jp1
(X2 j)
−
͚ X
jp1 j
N
2
(8-6. 17].ASME PTC 19.2) and (8-6. X curve is over the (Xj. For regression uncertainty when X and Y are functions of other variables. (8-6.
N ͚ ( Xj2) −
jp1
͚
jp1
N
2
(8-6. For higher-order linear regressions and other regression methodologies. the slope m is determined from
N ͚ XjY j − mp
jp1
N
N
Yj ͚ Xj j͚ jp1 p1
N
N
8-6 REGRESSION UNCERTAINTY 8-6.

therefore. Since all of the systematic standard uncertainties for Xj and Xnew are correlated. There is also the possibility that the Xnew values will have systematic standard uncertainties from the same sources as the Xj data causing these uncertainties to be correlated. If each of the Yj data points has the same systematic standard uncertainty. In subsection 8-6.. Cases are considered where Xnew has systematic standard uncertainty which is correlated with that in Xj and where Xnew has systematic standard uncertainty which is not correlated with that in Xj. in the Yj and Xj data. then the resulting elemental systematic standard ˆ from the curve-fit is [9] uncertainty for the mean Y
2 2 ˆ p ͓ ( mb X ) + ( mb X bY new) ͔ 3
1⁄ 2
(8-6. There also can be systematic standard uncertainty in the ˆ value.4 using eq. then the resulting elemental systematic standard uncertainty ˆ from the curve-fit is [9. the random standard uncerˆ obtained from eq. but Xnew has a different (no common systematic error sources) systematic standard uncertainty.9)
This case would occur if different instruments were used to measure the Xj values and Xnew.4 Systematic Standard Uncertainty for Y Determined From Regression Equation There can be systematic standard uncertainty.1) to obtain the total ˆ value from the curve-fit. If each of the Xj data points has the same systematic standard uncertainty. bX.`. 8-6. and are. and Xnew has the same systematic standard uncertainty (from the same sources).7)
͚ Xj jp1
N
(8-6. All of the bYj uncertainties are assumed to be completely correlated with each other. of the elemental systematic standard uncertainties for each of the Yj data points are from the same sources.. and Xnew has no systematic standard uncertainty. 14.`.4 Systematic Standard Uncertainty in Xj Data With Uncorrelated Systematic Standard Uncertainty in Xnew. bY1. The effect on the curve-fit is to shift it to the right or left depending on the sign of the errors (the signs and magnitudes of the errors are unknown). bYj and bXj respectively.. 17]
ˆ p mb x bY 2
(8-6. (8-6. The assumption is made that there are no common uncertainty sources between Yj and Xj (no correlation between the bYj and bXj systematic standard uncertainties).3 Systematic Standard Uncertainty in Xj Data With Correlated Systematic Standard Uncertainty in Xnew.TEST UNCERTAINTY
ASME PTC 19. This shift has no effect on the value ˆ obtained from the curve since the shift in of Y Xnew is the same as the shift in Xj. 8-6. It is very likely that most.
--`. correlated.6)
If there is no random standard uncertainty in the Xj data or the new X values used in the regression equation.`. bX. then the resulting elemental systematic standard uncertainty ˆ from the curve-fit is determined for the mean Y as [9.1 Systematic Standard Uncertainty in Yj Data.```.4.````````.4... The same is true for the Xj data points. only constant systematic standard uncertainties for (Xj. bXnew. For uncertainty for the Y random standard uncertainty in the Xj or X values. and probably all.2 Systematic Standard Uncertainty in Xj Data With No Systematic Standard Uncertainty in Xnew. bX. 17] for the mean Y
39
8-6. This case would occur if the same instruments are used to measure Xnew as were used to measure Xj. If each of the Xj data points has the same systematic standard uncertainty.`. If each of the Xi data points has the same systematic standard uncertainty. the general approach in [9] should be used. (7-5.-`-`.``..4.4. the systematic standard errors are all the same. and the systematic standard uncertainty for Xnew is bXnew. Yj) and Xnew are considered.1-2005
where
Xp 1 N
ˆ p bY bY 1 1
(8-6.5) is combined with tainty sY the systematic standard uncertainty (discussed in para. 8-6. X value used in the curve-fit to find a Y This curve-fit X will be called Xnew to distinguish it from the Xj data points.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. and all of the bXj uncertainties are assumed to be completely correlated with each other.8)
This case would occur when the regression equation from a set of test data is used later in a design or analysis process where Xnew might be taken as a value that has no uncertainty. 8-6. 08/17/2010 06:03:02 MDT
.`. ˆ 8-6. then the resulting elemental ˆ systematic standard uncertainty for the mean Y from the curve-fit is zero [9].

5).1). (8-6.15)
ˆ From Regression 8-6.
40
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`. and the degrees of freedom for bY datory Appendix B).`. For systematic standard uncertainty in Yj only or systematic standard uncertainty in Yj with correlated systematic standard uncertainty between Xj and Xnew
ˆ p bY ˆ bY 1
For no systematic standard uncertainty in the Yj data and correlated systematic standard uncertainty between Xj and Xnew
ˆ p 0 bY
(8-6.10)
(8-6.4. (7-5. 08/17/2010 06:03:02 MDT
.12)
bY ˆ p bY
1
1
For no systematic standard uncertainty in the Yj data. The uncertainty ˆ in eq. the systematic stanˆ is dard uncertainty for the curve-fit value of Y
ˆ is based Note that the degrees of freedom for Y ˆ .`---
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.-`-`.e. (8-6.4. because of the expression for sY As noted earlier...1-2005
TEST UNCERTAINTY
Table 8-6.ASME PTC 19. 8-6..13)
0
͓͑mbX͒2 + ͑mbXnew͒2͔
1⁄ 2
For no systematic standard uncertainty in the Yj data.16) will vary with X (i.`.11)
For systematic standard uncertainty in the Yj data and the Xj data and uncorrelated systematic standard uncertainty in Xnew.5 Systematic Standard Uncertainty for Y The systematic standard uncertainty for the mean ˆ from the curve-fit will be the appropriate rootY ˆ elemental systematic stansum-square of the bY i dard uncertainties defined above and summarized in Table 8-6.````````.```.`.15) only applies if there is no random standard uncertainty in X and if the systematic standard uncertainties are percent of full-scale values. the uncertainty expression in eq.5 Uncertainty for Y Equation ˆ obtained from The total uncertainty in the Y the simple linear regression expression..`.14)
ˆ. (8-6. eq.. is given by eq. systematic standard uncertainty in the Xj data.1) for the case where the ˆ are sufficiently large so degrees of freedom for Y that t ≈ 2
ˆ + sY ˆ͔ ˆ p 2 ͓b Y UY 2 2
1⁄ 2
(8-6.``. and uncorrelated systematic standard uncertainty in Xnew.. the systematic standard uncertainty ˆ is for the curve-fit value of Y
ˆ p bY ˆ bY 3
(8-6.16)
For systematic standard uncertainty in the Yj data and the Xj data and no systematic standard uncertainty in Xnew
ˆ p ͓bY ˆ + bY ˆ ͔ bY 1 2 2 2
1⁄ 2
(8-6.5.4.5 Systematic Standard Uncertainty ˆ Determined from Regression Components for Y Equation
Systematic standard uncertainty in Yj data Systematic standard uncertainty in Xj data with no systematic standard uncertainty in Xnew Systematic standard uncertainty in Xj data with correlated systematic standard uncertainty in Xnew Systematic standard uncertainty in Xj data with uncorrelated systematic standard bY uncertainty in Xnew ˆ p
3
ˆ p ͓bY ˆ + bY ˆ ͔ bY 1 3
2
2
1⁄
2
(8-6. Xnew) band UY ˆ from eq. The use of the factor t ≈ 2 will be appropriate in most cases.`. (8-6. systematic standard uncertainty in the Xj data.. which is N − on the degrees of freedom for sY ˆ (see Nonman2. and no systematic standard uncertainty in Xnew
ˆ p bY ˆ bY 2
bY ˆ p mbX
2
(8-6.

and for other analyses. one or more of the systematic uncertainties may be nonsymmetric (see subsection 8-2). It may also be used to compare alternative instrumentation systems and test designs and to determine corrective action if the predicted uncertainty is unacceptably large. (2) List all independent measurement parameters and their nominal levels.e. according to the functional relationship defined in step (a)(4). the degrees of freedom may be small for some of the uncertainties (see Nonmandatory Appendix B). For most uncertainty analyses. The second is a table that summarizes the uncertainty information at the result level. For some analyses. i. (c) Calculate the Systematic Uncertainty and Random Uncertainty (Standard Deviation of the Mean) for Each Parameter (see subsections 6-1 and 6-2). the predicted uncertainty should be smaller than the required uncertainty. according to the procedure that follows. (e) Calculate Uncertainty (see subsection 7-5). provides data for validity checks. and provides a statistical basis for comparing test results. The posttest analysis (see subsection 4-4) validates the pretest analysis. standard deviation of the mean. (4) Define the functional relationship between the independent measurement parameters and the test result. b i (f) sample random standard uncertainty..i (g) sensitivity. either by differentiation or by numerical analysis. (i b i)2
. be conducted before and after each test. all measured parameters will have symmetric systematic uncertainties and large degrees of freedom. 9-2 CALCULATION PROCEDURE (a) Define Measurement Process (see section 5).1-2005
Section 9 Step-by-Step Calculation Procedure
9-1 GENERAL CONSIDERATIONS It is recommended that an uncertainty analysis. Be sure to check for uncertainties in measurement system components that affect two or more measurements simultaneously (correlated uncertainties). The first is a detailed report that displays all the information used in the calculation of the nominal value and uncertainty of the result. 08/17/2010 06:03:02 MDT
(1) The systematic uncertainty and random uncertainty (sample standard deviations of the means) of the independent parameters are propagated separately all the way to the final result. The pretest analysis (see subsection 4-4) is used to determine if the test result can be measured with sufficient accuracy. (f) Report. following the methods of Sections 4 through 8 of this Supplement. The detailed report table should include. (2) Propagation of the standard deviations of the means is done. (1) Review test objectives and test duration. by using the Taylor series method (see section 7). (3) List all calibrations and instrument setups that will affect each parameter. (1) The uncertainty analysis for each calculated result should be reported on two tables. (b) List Elemental Error Sources (see subsection 5-3). (d) Propagate the Systematic and Random Standard Deviations (see subsections 7-1 through 7-4). the following information for each parameter used in the calculation of the result: (a) symbol used in the calculations (b) description (c) units (d) nominal value (average of measurements). This requires a calculation of the sensitivity factors. Furthermore. (1) Combine the systematic and random uncertainties to obtain the total uncertainty.TEST UNCERTAINTY
ASME PTC 19. it may be used to evaluate the need for calibration. (1) Make a complete and exhaustive list of all possible test uncertainty sources for all parameters. s x. X (e) systematic standard uncertainty.
41
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. as a minimum. (h) systematic standard uncertainty contribution to the combined uncertainty of the result.

bX
i
Symbol
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F Differential pressure head across venturi (at 68°F)
Units
.`..4 1. (i sX)2
The summary report table should display the information associated with the result as detailed in Tables 9-2-1 and 9-2-2 which are based on Example 10-2.001 0.0 0.1-2005
TEST UNCERTAINTY
Table 9-2-1 Table of Data
Independent Parameters
Absolute Systematic Standard Uncertainty.00013 0. i SX 2
i
0..0005 0. UR
0.286
Absolute Expanded Uncertainty. in.-`-`.. sR
0.0000049 0.`---
42
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale..```.0 0.001 62..`.002 0.11 0.4
Absolute Random Standard Uncertainty.0 0.002 0. R
138. i
140 86.276 0.0 0.. 08/17/2010 06:03:02 MDT
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
.0766
Absolute Combined Standard Uncertainty.
--`.4
0. lbm/ft3 in. SX
i
Absolute Sensitivity.`.984 3.``.ASME PTC 19.276
Symbol
˙ m
Description
Mass flow rate
Units
lbm/s
Calculated Result. uR
0.0 0.`.0 0.`.00186 0.999 6. in.4 100
Absolute Random Standard Uncertainty.. H2O
Nominal Value
0.6919
Absolute Systematic Standard Uncertainty Contribution.2 −11..00375 0.0000049 0.````````.15
0.0766
Table 9-2-2 Summary of Data
Calculated Result
Absolute Systematic Standard Uncertainty. bR
0.572
(i) random standard uncertainty contribution to the combined uncertainty of the result. i b X 2
i
Absolute Random Standard Uncertainty Contribution.0108
0.

A total of 60 measurements are taken in succession at each traverse point once the Pitot tube is positioned. (c) the output of the differential pressure transmitter is measured and recorded by a computerized data acquisition system (DAS).3-1. (f) the Pitot tube. Several simplifying assumptions are made for this example: (a) the pipe diameter is large compared to the Pitot tube diameter such that blockage effects and wall interference effects can be neglected.2)
.
10-1. and making the
43
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. (4-3. This avoids the need to develop weighting factors for each sample area. (d) the DAS computes velocity for each measurement by taking the square root of the output of the differential pressure transmitter. differential pressure transmitter.1). the velocity is measured at 40 unique traverse points (10 traverse positions along 4 equally spaced radii) corresponding to the centroid of equal areas as shown in Fig. 08/17/2010 06:03:02 MDT
appropriate calibration corrections and unit conversions. (e) the DAS automatically takes 60 readings at each traverse point and computes and records average values and standard deviations for the data collected at each traverse point. the average velocity along each radius. Vi .1-2005
Section 10 Examples
10-1 FLOW MEASUREMENT USING PITOT TUBES 10-1. One technique for measuring the integrated-average velocity of the fluid is to traverse the cross-sectional flow area with a Pitot tube. For this example.3 Velocity Results The DAS is also programmed to output the sample standard deviation of the 60 measurements at each traverse point based on eq. The sample standard deviation at each point is summarized in Table 10-1. and (g) the flow rate and the velocity profile remain constant for the duration of the test.
10-1. Traverse points are typically specified at the centroid of equal areas so that the integratedaverage velocity may be estimated as the average of the measured values for all traverse points. and DAS are calibrated together as a system.1 Define the Measurement Process The flow rate of an incompressible fluid in a pipe may be determined by multiplying the integrated-average velocity of the fluid by the cross-sectional flow area of the pipe.1.2 Data Summary The computerized data acquisition system is used to compute average values from the 60 measurements at each traverse point using eq. (4-3. (b) the velocity pressure developed across the pitot tube is measured by a differential pressure transmitter. The traverse points are located at the centroid of equal areas so that the integrated-average velocity in the pipe is approximated by the average of the velocities determined at the traverse points. is approximated as
Vi (ft/sec) ≈
͚ (⁄ jp1
1
10
10
) X if (ft/sec)
(10-1. The resulting average values at each traverse point Xij are summarized in Table 10-1. 10-1.TEST UNCERTAINTY
ASME PTC 19.2).2. First.1)
Then. Measurements at each traverse point can be used to determine local fluid velocity. the average velocities calculated from these point velocities are treated as results (Section 7). the average velocity in the pipe V is approximated as
V (ft/sec) ≈
͚ ( ⁄ ) Vi (ft/sec) ip 1
1 4
4
(10-1. making the appropriate corrections for fluid density. The point velocity values at individual traverse points are treated as measurements (Section 6).

55 5.65 5. i ؍1 X ij (ft/sec)
5.21 5.68 5.65 5.1-2005
TEST UNCERTAINTY
Fig.`.72
Radius 2. 10-1.80
Radius 3.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.70
Radius 4.74 5..61 5.42 5.74 5.46 5. i ؍4 X ij (ft/sec)
5.`.47 5.````````.65 5.37 5.73 5.70 5.63 5.27 5.2 Average Values (Example 10-1)
Traverse Point.58 5.-`-`.65 5.53 5. i ؍3 X ij (ft/sec)
5..`.```.50 5.76 5.31 5.79 5..25 5.ASME PTC 19.65 5.75 5.55 5.00 5.69 5..62 5..77 5.31 5.1 Traverse Points (Example 10-1)
Table 10-1. j Radius 1.76 5..63 5.``.67
j j j j j j j j j j
p p p p p p p p p p
1 2 3 4 5 6 7 8 9 10
44
--`.16 5. 08/17/2010 06:03:02 MDT
.`. i ؍2 X ij (ft/sec)
5.`.

208 0.3-2.31 1..67 5.3-1. j
1 2 3 4 5 6 7 8 9 10
ij
Radius 2.169 0.137 0.
10-1.34 1. Nij p 60.3-1. The results of these calculations are summarized in Table 10-1.
--`.238 0. 08/17/2010 06:03:02 MDT
.04 1.36 1.81
(ft/sec)
0.141 0.187
(ft/sec)
1.65 1.29 1.54
(ft/sec)
0.93 1.167 0.61 1.43 1.199
Table 10-1.45
(ft/sec)
0..21 1.232 0.89 1.225 0.185 0.80 2.148 0.65 1.``.176 0.173 0.152 0.`.213 0. The sources of uncertainty which are considered systematic in this simplified example are the uncertainty of the calibration of the instruments used to measure and record velocity at each traverse point and the uncertainty of the integrated-average velocity due to spatial variation.62 5.26 1.263 0.169 0.`.26 1.31 1.276 0.`---
The resulting values for SXij are also presented in Table 10-1.105
(ft/sec)
1. i ؍1
SX
Traverse Point. the degrees of freedom at each traverse point is
ij p Nij − 1 p 59
(10-1.36 1.14 1.18 1.45 5.-`-`.4 List Elemental Uncertainty Sources The sources of uncertainty which are considered random in this simplified example are those causing variation in the 60 repeated measurements of velocity at each traverse point.185 0.129
(ft/sec)
1.09 1.21 1.`. i ؍2
SX
ij
Radius 3.141 0.TEST UNCERTAINTY
ASME PTC 19.120 0.176 0.163 0.00
(ft/sec)
0.105 0.244 0.1) as
10-1.81 1.4)
where
1 ∂ Vi p ∂ X ij 10
The random standard uncertainty of the average velocity in the pipe s V is then calculated as
sV p
΄j͚ p 1 ∂V
4
∂V
i
s Vi
΅
2
1⁄
2
(10-1..06 1.1-2005
Table 10-1.208 0. i ؍4
SX
ij
SX
ij
SX
ij
SX
ij
SX
ij
(ft/sec)
1.06 1.133 0.3-1 Standard Deviations (Example 10-1)
Radius 1.6 Propagate Random Standard Uncertainty The random standard uncertainty for each average velocity along a radius SVi is calculated from eq.2 is calculated from eq.57
where values for SXij are shown in Table 10-1.. (7-3.3) as
45
SVi p
͓͑∂Vi / ∂Xij͒͑sX ͔͒2· Άj͚ p1
ij
10
1⁄
2
(10-1.5)
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.`.41 1.`.74 1. respectively. (4-3.15 0.129 0..78 1.09 0.163 0.157 0.5 Calculate Random Standard Uncertainty The random standard uncertainty of the mean value at each traverse point presented in Table 10-1..208 0.3)
The subscripts i and j are used in the previous equations to designate radius and traverse positions.43 1.00 1. 10-1.156 0.84 1. Since there are 60 measurements at each traverse point.213 0.182 0.61 1.156 0.````````.230 0.03 0.03 1.3-2 Summary of Average Velocity Calculation (Example 10-1)
Parameter
V1 V2 V3 V4 V
S Xij p
S Xij
Ί N ij
Value (ft/sec)
5.61 2. i ؍3
SX
ij
Radius 4.52 5.```.133 0.

2 Spatial Variation.0523 0. Therefore. (10-1. Averaging the velocities at the centroid of equal area points is a numerical approximation of this integrated average. ft/sec
sV 1 sV 2 sV 3 sV 4 sV
Value
0.``. While not shown here.-`-`. The resulting value is shown in the following equation.0618 0.
10-1. as computed using eq.6. estimated as the limits of the elemental systematic error at 95% confidence. it is unknown whether the average of the traverses of the four radii sufficiently characterizes the integrated average around the pipe due to circumferential variations. the uncertainty due to spatial variation must be inferred from available data. the degrees of freedom associated with each s Vi as well as SV can be calculated using eq. of the average pipe velocity measurement is estimated as
bVc p (0. The published degrees of freedom for this value is 30. and DAS are calibrated together as a system.`. and (d) evaluation of the variation in test measurements as a function of spatial position. This uncertainty can be estimated in a variety of ways.0111 ft/sec
Since no additional pipe taps are available for testing.0440 0.03/2)(5. (4-3.6 Standard Deviation of Average Velocity (Example 10-1)
Parameter. Using eq. 08/17/2010 06:03:02 MDT
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
. it will be estimated by evaluation of the variation in the radial averages. digital pressure transmitter. For this example.7 Calculate Systematic Standard Uncertainties 10-1. including (a) special tests which provide independent knowledge of the velocity profile.6)
The degrees of freedom of bVC is assumed to be large (≥30). Therefore. (B-1. (c) published reports which document the uncertainty of similar measurement techniques in similar measurement situations. Since the velocity profile remains fixed for the duration of the test.1 Calibration.ASME PTC 19.
10-1. Special tests had previously been conducted to determine if taking additional traverse points (20) along each of the 4 radii would result in a significant change in the measurement of the overall average velocity in the pipe. and DAS. Velocity is measured at each point with the same Pitot tube. This uncertainty of method is sometimes referred to as uncertainty due to spatial variation. there is an inherent uncertainty in the method used to approximate the integrated-average velocity.```.`.`.4).. the spatial variation is a source of systematic uncertainty for the test result.`. Assuming that the four measured
46
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. the elemental systematic standard uncertainty associated with the calibration of the instruments.. the analyst must recognize that the numerical average may not equal the integrated-average value over the entire cross-sectional flow area. the velocity profile is distorted due to the presence of flow disturbances upstream of the measurement location.0836 (ft/sec)
(10-1.0287
where
∂V
∂Vi
p 1 4
The results are summarized in Table 10-1.7).`---
of spatial position. (b) special tests which compare the measurement technique to other techniques which yield the desired integrated-average. This is due to incomplete sampling of a profile which varies as a function
--`.
b VSR p 0.0687 0. The Pitot tube..7.57) p 0.````````.`.7.. The standard deviation published from these past tests was assumed to be representative of the systematic standard uncertainty (for the present test) resulting from spatial variation in the radial direction. For this example.2). the calibration uncertainty of the instruments. is 3% of measured velocity. Even if exact values of velocity are known at each of the traverse points. digital pressure transmitter.1-2005
TEST UNCERTAINTY
Table 10-1. For this example. The results of these past tests were used to compute a standard deviation representing the dispersion of errors (differences) in average pipe velocities computed using 20 traverse points along each of the 4 radii versus those obtained using 10 traverse points.. The true value being measured is the integrated average of the velocity over the cross-sectional flow area..

(102. (B-1.8 Propagate Systematic Standard Uncertainties The systematic standard uncertainties due to calibration and spatial variation are combined using eq.1 General Description
bVSC p
Ί
1 N V
i
͚ ip1
NV
i
( V i − V ) p 0.204 (ft/sec)
Use of the 2 in the above equation is appropriate as it can be shown using eq. 10-2 FLOW RATE UNCERTAINTY [12] 10-2. A pretest analysis is required to determine if an uncalibrated venturi could be used to satisfy the test objective. the systematic standard uncertainty in the average velocity measurement due to spatial variation in the circumferential direction is estimated using eq.1)
10-1. The degrees of freedom associated with ˆ − 1 p 3. (d) List Elemental Systematic Uncertainty Sources. The systematic values were evaluated and are listed in Table 102.099702 Cd 2 Ί h
(10-2. if not.1-2005
radial averages come from a population of values which are normally distributed.
˙ p m 0.7)
where NVip the number of independent radial average velocities used in the computation of bVSC N V p the number of radial average velocities i averaged together in the computation of V. A summary presentation of the uncertainty
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.0977 (ft/sec)
2
(10-1.2 Uncalibrated Venturi Case (b). Calculate the Systematic and Random Uncertainties.0494)2] p 0. The standard deviation of the mean of the
47
The expanded uncertainty of the resulting average pipe velocity measurement is determined using eq.0111) 2 + (0.9)
The definitions and values for each of the measurements used in the calculation of the mass flow rate are displayed in the first four columns of Table 10-2.8)
In this example.0836) 2 + (0.TEST UNCERTAINTY
ASME PTC 19. Estimate Elemental Uncertainties. definitions. 10-2. etc. are displayed for each of the variables required for the analysis. 08/17/2010 06:03:02 MDT
.5%. The table associated with the uncalibrated case is shown here first.. Flow rate can be calculated by making the measurements required to define the independent variables found in eq. The clearest way to present the results of the steps in the uncertainty analysis is to develop a table in which the names. and. as Table 10-2. venturi (see Fig. uncertainties.7) that the combined degrees of freedom of the result is ≥30. The systematic uncertainties for each input parameter are displayed on an absolute basis. Both of these values are equal to 4. The table has been developed in accordance with the step-bystep procedure of Section 9. 10-2. the test objective is to determine the flow of water using a 6 x 4 in. (7-5. (7-4.1) within an uncertainty of 0..e.1)[21].1-1. (c).9 Uncertainty of Result The combined standard uncertainty of the resulting average pipe velocity measurement is determined using eq. sensitivities.1-1. The example looks at each case in both absolute and relative (i. values.1-1.1) as follows:
b V p [(b VC ) 2 +(b VSR ) 2 + ( b VSC ) 2]
1⁄ 2 1⁄
p [(0.2) as follows:
U V p 2 u V p 0. The steps in the development of the table are as follows: (a) Define Measurement Process.0494 ft/sec N Vi− 1
2
(10-1. This example outlines differences in the analysis for both the uncalibrated and calibrated cases. b VSC is equal to NV i
10-1. The degrees of freedom associated with each of the uncertainty estimates are assumed to be greater than 30. whether calibration of the venturi would achieve the desired objective. (4-3.102 (ft/sec) (10-1.1) as follows:
uV p [( b V )2 + ( s V ) 2]
1⁄ 2
Ί
d 1− D
4
p 0. percent) value formats. (7-5.5) as
analysis at the test result level is reported in Table 10-1-9.

999 6.2°F and 0.. H2O
Nominal Value
0.```.`.``.002 0. 08/17/2010 06:03:02 MDT
. respectively.`---
ASME PTC 19.`. bX
3.002 0. in. lbm/ft3 in. x 4 in. in.102 0.1 Schematic of a 6 in. sX
i
0 0 0 0.0E–03 0. Uncertainty.`. Uncertainty.001 62. 10-2...1-1 Uncalibrated Case (Example 10-2)
Independent Parameters
Absolute Systematic Standard Uncertainty.. Units Calculated Value bv sv uv Uv
5. Venturi
Table 10-2.57 0.1°F.9 Uncertainty of Result (Example 10-1)
Absolute Absolute Absolute Systematic Random Combined Absolute Standard Standard Standard Expanded Uncertainty.15
Symbol
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [25] Differential pressure head across venturi (at 68°F)
Units
.984 3..37 100
Random Standard Uncertainty of the Mean.0287 0.4
GENERAL NOTES: (a) The systematic and random estimates for density are based on water temperature measurements having systematic and random uncertainties of 0. Uncertainty.75E–03 5.0977 0...0E–04 1.
48
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. (b) The systematic uncertainty for the differential pressure head is assumed to be one-half the least count of the manometer scale.1-2005
TEST UNCERTAINTY
Table 10-1.Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`.`.204
Symbol
V
Description
Average ft/sec velocity in pipe
Fig.````````.-`-`.`..

1-3 display the same parameters as Tables 10-2. (e) Propagate the Systematic and Random Uncertainties. Estimate Elemental Uncertainties. As the formulas do not change.5% is marginally The test objective of m satisfied by using a calibrated venturi (Table 10-2. Tables 102. by finding the partial derivatives of the result with respect to each of the measured parameters.1. (7-2. respectively. These new values can be inserted into the tables (spreadsheets) used for the uncalibrated case. ˙ .13 shows the formulas for the partial derivatives for each of the measured parameters with respect to the calculated mass flow rate. the largest contributor is the systematic uncertainty in the discharge coefficient.1. (g) Report. 08/17/2010 06:03:02 MDT
. In addition.111). Multiplying by 100 would provide an expression of the uncertainty in terms of a percentage of the nominal value. Table 10-2.2. The parameters and results are displayed using a relative basis for the uncertainties and sensitivities in Tables 10-2. the same basic equation for the ˙ was repeated seven times (once calculation of m for each independent parameter.1. along with the formulas used in the spreadsheet. either numerically or analytically.1-7 would be included in the report.. Calculate Uncertainty. Calculate the Systematic and Random Uncertainties.1-5 shows the nominal value. and 10-2. ˙ ± 0.1% (i. 10-2. and total uncertainties for m described previously.1-4.1-2005
measurements taken is shown for each of the independent variables. 7-2. the sensitivity coefficient can also be determined analytically. (c).1-5. To convert the uncertainty of any parameter from absolute to relative terms. The uncertainty estimate for discharge coefficient comes from calibration data.3 Calibrated Venturi Case (a) Define the Measurement Process (b). The sensitivity of the result to each of the individual parameter uncertainties is calculated.1. The sensitivity coefficients were converted to relative terms using eq. A quick and easy way to numerically calculate the sensitivity coefficients of the independent parameters is to develop a table using a spreadsheet program on the personal computer. As noted in para.1-1. Tables 10-2.061%.1-2.
49
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
it is necessary only to divide the absolute uncertainty by the nominal value of the parameter.1. 0.1-2.1-5. where Qx is the quantity by which Xi is to be perturbed. 10-2. and the sensitivity coefficients found using these formulas. Table 102.1.1. In this case the systematic uncertainty of the discharge coefficient changes. with text noting that calibration of the venturi is required to meet
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.1-4 as Bc /C p 0. the systematic uncertainties of the throat and pipe diameters are eliminated (set to zero in Table 10-2. which is standard practice. and is listed as a systematic uncertainty in Table 102.1.1-4. (f) Propagate the Systematic and Random Uncertainties. 10-2. A perturbed m calculated for each independent parameter and ˙ is subtracted from each the baseline value of m ˙ perturbed m. in accordance with subsection 7-2.2).1 Relative Sensitivities and Uncertainties. and the systematic. X. with Xi being replaced with Xi + Xi Qx. so the units are the same as the units for the measured parameter. The most promising path to take to obtaining additional reductions in the uncertainty of the mass flow rate determination would be through an increase in the number of differential head readings to reduce s X. but with the sensitivities and uncertainties expressed in relative (percent) terms rather than in absolute terms.1-1. (e). Sensitivities and uncertainties may also be stated in relative terms.001). there is nothing more that needs to be done. Multiplying by 100 yields the percentage change in the result for a 1% change in the measured parameter. it can be seen at once which parameters contribute most to the uncertainty of the result.TEST UNCERTAINTY
ASME PTC 19.1. Dividing the difference by the amount by which Xi was perturbed (Xi + Xi Qx) provides the absolute sensitivity coefficient. (f) Calculate Uncertainty.1-4) because the entire flow section is calibrated as a unit. 10-2.1-4.122%.1-1 through 10-2. A summary of these uncertainties is given in Table 10-2. which corresponds to an absolute systematic standard uncertainty of bc p Bc / 2 p 0. Table 10-2. The total uncertainty of the result is then calculated by root-sum-squaring the systematic and random contributions. Table 10-2.1. In this case.1-2 shows the results obtained using such a spreadsheet.1-6. and 10-2. A forward differencing scheme is used with a value of Qx equal ˙ is therefore to 0.1. (d) List Elemental Uncertainty Sources. and 102.e. C. calculated as random. which has been fossilized. These values are also expressed in absolute terms. As can be seen by looking at the formulas shown in Table 10-2.1. By looking at the numbers in the last two columns of Table 10-2.1-7.

1-2005
TEST UNCERTAINTY
Table 10-2.6
Ί1 − D
d
CQx d 3.099702d 2
Ί1 − D
d
Ί( + Qx )h − m ˙
4
1.0095 and B+ p 0. H2O and B+ p 0.````````.`.5% uncertainty or less in the mass flow rate determination.1 in.`.099702d 2
d
Ίh
4 x
−11.. The only differences are the following:
50
(a) the discharge coefficient will have a nonsymmetrical absolute systematic uncertainty (95% confidence estimate) represented by B− p 0.099702(d + dQx ) d2
Ί1 −
D 6.4
˙ Formula for m
0.`.692
Result
138..5 in.
10-3 FLOW RATE UNCERTAINTY INCLUDING NONSYMMETRICAL SYSTEMATIC STANDARD UNCERTAINTY This example is nearly identical to that presented in subsection 10-2 for the uncalibrated venturi case..1%
Symbol. 08/17/2010 06:03:02 MDT
.3
0.-`-`.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. and (b) the differential pressure head across the venturi (at 68°F) will have a nonsymmetrical absolute systematic uncertainty (95% confidence estimate) represented by B− p 0.1-2 Absolute Sensitivity Coefficients in Example 10-2 (Calculated Numerically)
Quantity by Which to Perturb Independent Parameters (Qx ): 0.3
˙ −m
Ί1 − D + DQ
DQx
62.099702Cd 2
Ίh
4
Ί1 − D
d
the objective of 0..`..ASME PTC 19.984
Formulas for Absolute Sensitivity ˙ ٢m ٢Xi
0.```.``. H2O.`.099702d 2
Ί1 − D
d
hQx
Symbol
˙ m
Ί(h + hQx ) − m ˙
4
0. Xi
C
Nominal Values
0.37
0.099702 (C + CQx ) d 2
Absolute Sensitivity ˙ ٢m ٢Xi
140.11
Qx
h
100
0.001
d + dQx D
dQx
4
Ίh − m ˙
86. the discharge coefficient will have a nonsymmetrical
--`.. Using the methodology in subsection 8-2.0055.999
4
Ίh − m ˙
0.

X + q. 08/17/2010 06:03:02 MDT
.099702CΊ h͒ d 2 (−4)
4
D d d 2 Ί1 − D D1 − D
d
4
3
72. The absolute sensitivity..1-1.`---
flow rate.999
Ί1 − D
d
4
(2)0. This example highlights the following: (a) the identification and quantification of elemental sources of uncertainty.001
−
4(0.TEST UNCERTAINTY
ASME PTC 19..````````.37
Ί1 − D
0. is estimated by q p (b+ − b−)/ 2. H2O.1-3 Absolute Sensitivity Coefficients in Example 10-2 (Calculated Analytically)
Formulas for Absolute Sensitivity ˙ ٢m ٢Xi
0.`. for each variable is estimated by b p (b+ + b−)/ 2 and the offset.05 in. and the differential pressure head across the venturi will have a nonsymmetrical absolute systematic standard uncertainty of b− p −0.4
D
d
d
4
4 3⁄2
0.35 in.099702Cd 2
Ί/h
2
4
h
100
Ί
d 1− D
0.6
D
6. RX. qr.099702C
Ίhd −
͑0.099702)Cd 6
2D5 1 −
Ίh
−11. q. b..`.`.692
absolute systematic standard uncertainty of b− p 0.1-1.-`-`.00575 and b+ p 0.984
Ίh
4
140
Ί
d 3.. Following the method given in subsection 8-2.. the absolute systematic standard uncertainty.R − qr..099702Cd 2
Ίh/
2
1. and U + p U95. The results of this uncertainty analysis are presented in Tables 10-3-1 and 10-3-2 in which each symbol has the same description as in Table 10-2.`.``.`. The lower and upper limits on the uncertainty of the mass flow rate are given by U − p U95. respectively. and in which each corresponding number has the same units as those given in Table 10-2.
10-4 COMPRESSOR PERFORMANCE UNCERTAINTY The following example follows the step-by-step procedure outlined in Section 9. RX + q. R + qr.00175.1-2005
Table 10-2.```. .11
62. and the mass flow rate evaluated when all variables are set to their measured values. for each variable is estimated with all variables set to their offset values. Xi
Nominal Values
Absolute Sensitivity ˙ ٢m ٢Xi
C
0. is determined by the difference between the mass flow rate evaluated when all variables are set to their offset values.
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. The apparent offset in the mass
51
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`.099702d 2
d 1− D
Symbol. H2O and b+ p 0.

75E–03 5.`.``. in.ASME PTC 19.`. (b X i )2
i
Absolute Random Standard Uncertainty Contribution.2 −11.4
Absolute Systematic Standard Uncertainty.92 ؋ 10−6 7.95
1.1-2005
TEST UNCERTAINTY
Table 10-2.0E–04 1.1-5 Summary: Uncertainties in Absolute Terms (Example 10-2: Uncalibrated Case)
Absolute Random Standard Uncertainty.0125% 0.692
Absolute Systematic Standard Uncertainty Contribution.```.`.29 ؋ 10−4 4.`---
Table 10-2.86 ؋ 10−3 1. s′X ؍s X /Xi
i
i
0.400%
52
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. H2O
Nominal Value
0.37 100
Relative Random Standard Uncertainty...604
Total Absolute Uncertainty.0032% 0.1-1 Relative Uncertainty of Measurement (Example 10-2: Uncalibrated Case)
Independent Parameters
Relative Systematic Standard Uncertainty. 08/17/2010 06:03:02 MDT
.1-4 Absolute Contributions of Uncertainties of Independent Parameters (Example 10-2: Uncalibrated Case)
Independent Parameters
Parameter Information (in Parameter Units)
Absolute Systematic Standard Uncertainty.150%
0% 0% 0% 0.0167% 0. sR
0.93 ؋ 10−6 1.276
Symbol
˙ m
Description
Mass flow rate
Units
lbm/sec
Calculated Value
138.4
0. (s X i )2
i
0 0 0 0.001 62... b′X ؍b X /Xi
i
i
Symbol
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [24] Differential pressure head across venturi (at 68°F)
Units
..361% 0.0E–03 0.276 1.002 0..999 6. in.537
Combined Standard Uncertainty of the Result. bX
3.984 3.15
Uncertainty Contribution of Parameters to the Result (in Results Units Squared)
Absolute Random Standard Uncertainty.37 100
Absolute Sensitivity.11 0.. in.1. lbm/ft3 in. in.0032% 0. sX
i
Symbol Xi
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [25] Differential pressure head across venturi (at 68°F)
Units
. uR
0..002 0. i
140 86.-`-`. lbm/ft3 in.21
Table 10-2. bR
0.999 6. UR.4 1.08 ؋ 10−2
0 0 0 4..`.001 62.````````.984 3.66 ؋ 10−2
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`..`. H2O
Nominal Value
0.

50
Relative Random Uncertainty Contribution.1-2005
Table 10-2.999 6.37 100
Relative Sensitivity. ′ ؍sX / Xi sX
i
i
0.95 /R
0.061% 0% 0% 0. 08/17/2010 06:03:02 MDT
. sR /R
0.0032% 0. (s′X i′)2
i
1.0 ؋ 10−6
Table 10-2.4923 0. in.0032% 0.001 62. in.0 2..73 2. in.50 0.1.57 5..39%
Symbol
˙ m
Description
Mass flow rate
Units
lbm/sec
Calculated Value
138.. H2O
Nominal Value
0. UR.88%
Table 10-2. ′ ؍bX / Xi bX
i
i
Symbol
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [24] Differential pressure head across venturi (at 68°F)
Units
.71 6. bR /R
0.57 ؋ 10−10 4. H2O
Nominal Value
0.20%
Relative Combined Standard Uncertainty.1.984 3.TEST UNCERTAINTY
ASME PTC 19. uR /R
0. (b′X i′)2
i
Symbol
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [24] Differential pressure head across venturi (at 68°F)
Units
.44%
Relative Expanded Uncertainty.4923 −0.150%
0% 0% 0% 0. i′
1.1.45 9.1-4 Relative Uncertainties of Independent Parameters (Example 10-2: Calibrated Case)
Independent Parameters
Relative Systematic Standard Uncertainty.999 6.37 100
Relative Random Standard Uncertainty. lbm/ft3 in.1-2 Relative Contributions of Uncertainties of Independent Parameters (Example 10-2: Uncalibrated Case)
Independent Parameters
Relative Systematic Standard Uncertainty Contribution.984 3.001 62. in.400%
53
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. lbm/ft3 in.62
؋ ؋ ؋ ؋ ؋
10−5 10−8 10−9 10−10 10−7
0 0 0 2.1-3 Summary: Uncertainties in Relative Terms for the Uncalibrated Case
Calculated Result
Relative Systematic Uncertainty.4
Relative Random Uncertainty..

37 100
Relative Sensitivity.1-5 Relative Contributions of Uncertainties of Independent Parameters (Example 10-2: Calibrated Case)
Independent Parameters
Relative Systematic Standard Uncertainty Contribution.4 lbm/sec 0..`.``.22%
Relative Expanded Uncertainty.0 2.`.`. lbm/ft3 in.95 /R 138. 08/17/2010 06:03:02 MDT
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
. steady-state conditions for two hours prior to the test and for one hour during the test.1-1.4923 0. ′ i′ ) 2 (b X
i
Symbol
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [24] Differential pressure head across venturi (at 68°F)
Units
. The compressor was operated at normal.097% 0. A simplified schematic depicting the test measurement locations is shown in Fig.097%
Symbol
˙ m
Description
Mass flow rate
Units
lbm/sec
Calculated Value
138. 10-4. UR..44%
Table 10-2. 10-4.39% 0.1-6 Summary: Uncertainties in Relative Terms for the Calibrated Case
Calculated Results
Relative Systematic Uncertainty. H2O
Nominal Value
0.001 62.984 3.`.1 Define the Measurement Process A test was conducted to determine the adiabatic efficiency of an air compressor at normal operating
--`.1.```..50
Relative Random Standard Uncertainty Contribution. bR /R
0..50 0.95 /R
0..20% 0.1-2005
TEST UNCERTAINTY
Table 10-2.1.ASME PTC 19.````````.72 ؋ 10−7 0 0 2. ′ i′ ) 2 (s X
i
3..0 ؋ 10−6
Table 10-2. and (c) the importance of applying all known engineering corrections and using appropriate engineering relationships as part of the test results analysis process.1-7 Summary: Comparison Between Calibrated and Uncalibrated Cases
Calibrated Case
˙ m bR /R sR /R uR /R UR.. Each array was connected to a digital pressure indicator in a manner which yielded a spatially averaged pressure measurement.44% 0. The mean value for each measured parameter was calculated using eq. in.4 lbm/sec 0. i′
1. uR /R
0. The resulting averages are presented in Table 10-4.22% 0.1.4
Relative Random Uncertainty. (4-3.1). in.88%
(b) treatment of correlated sources of uncertainty in a practical manner.-`-`.999 6..`.1. Measurements of the total pressure and total temperature at the inlet and exit of the compressor were collected at one minute intervals resulting in 60 discrete measurements of each parameter over the test period.57 ؋ 10−10 5.44%
Uncalibrated Case
138. The compressor inlet and exit total temperatures were measured using multipoint thermocouple stagnation probe arrays which were permanently installed in the compressor inlet and
54
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.62 ؋ 10−7
0 0 0 2.20% 0.`---
conditions. The compressor inlet and exit total pressures were measured using multiport impact pressure arrays which are permanently installed in the compressor inlet and exit.20%
Relative Combined Standard Uncertainty.57 ؋ 10−10 4.4923 −0. sR /R
0.

-`-`. was used to measure both the inlet and the exit temperatures.34 1.`. lbm/ft3 in.002 0. A common digital temperature indicator.```.`..4
Independent Parameters
Uncertainty Contribution of Parameters to the Result (in Result Units Squared)
Absolute Systematic Standard Uncertainty Contribution. bX
3.61 ؋ 10−2
exit.001 62.``. b X+
1.9 ؋ 10−6 7.1-2005
Table 10-3-1 Absolute Contributions of Uncertainties of Independent Parameters (Example 10-3: Uncalibrated.0E–03 0.07 ؋ 10−2
0 0 0 4. Nonsymmetrical Systematic Uncertainty Case)
Independent Parameters
--`.002 0.86 ؋ 10−3 1.75E–03 5. in.0E–04 1.14 −11.`. in. in.002 −0. H2O
Nominal Value
0.`---
Parameter Information (in Parameter Units)
Absolute Negative Systematic Standard Uncertainty. The adiabatic efficiency of the air compressor was initially calculated using the following simplified engineering relationship:
55
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
p
( P2 /P 1)[(␥ −1)/␥] − 1 T2/T 1 − 1
(10-4.. H2O
Absolute Sensitivity. lbm/ft3 in.37 100
Absolute Positive Systematic Standard Uncertainty..TEST UNCERTAINTY
ASME PTC 19. The thermocouples from each multiprobe array were connected in parallel in a manner which yielded a spatially averaged temperature measurement.````````.999 6. sX
i
0 0 0 0.29 ؋ 10−4 4..`.984 3. with built-in cold junction compensation..`.0E–03 0.75E–03 5. 08/17/2010 06:03:02 MDT
.15
Absolute Random Standard Uncertainty. in. (sX i ) 2
i
0. (bX i ) 2
i
Symbol.9 ؋ 10−6 1.35
Absolute Systematic Standard Uncertainty. − bX
5.75E–03 5...0E–04 1. Xi
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [24] Differential pressure head across venturi (at 68°F)
Units
..6898
Absolute Random Standard Uncertainty Contribution.0E–03 0..1)
where P1p measured P2p measured T1p measured ature T2p measured ature compressor inlet total pressure compressor exit total pressure compressor inlet total tempercompressor exit total temper-
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale..05
Symbol Xi
C d D h
Description
Discharge coefficient Throat diameter Inlet diameter Water density at 60°F [24] Differential pressure head across venturi (at 68°F)
Units
.108 0.8 86.002 0.279 1. Pretest and posttest calibrations were performed on all instrumentation. i
140.0E–04 1.

Uncertainty. The difference between the two test results was compared to the previously reported uncertainty to illustrate the importance of applying all known engineering corrections and using appropriate engineering relationships as part of the results analysis process.606
Expanded Absolute Uncertainty.95+
1. bR
0.70 95.ASME PTC 19.35
Expanded Absolute Positive Uncertainty. sR
0.7.```.030 0.2 List Elemental Error Sources Based upon a review of the measurement methods and instruments employed for the test.4 Using the previous simplified engineering relationship. a test result and associated uncertainty were calculated and are presented in Tables 104.-`-`.95
1. UR. Recognizing that the use of the simplified engineering relationship in the computation of
56
10-4.52 520.`. 10-4.2
Absolute Random Standard Mean Value. UR. assumed to be 1. 10-4.14
Expanded Absolute Negative Uncertainty.540
Symbol
˙ m
Description
Mass flow rate
Units
lbm/sec
Calculated Value
138. the test result was reevaluated using more rigorous engineering relationships to eliminate some of the assumptions noted.21
Symbol
˙ m
Description
Mass flow rate
Units
lbm/sec
Offset.1-2005
TEST UNCERTAINTY
Table 10-3-2 Summary: Uncertainties in Absolute Terms (Example 10-3: Uncalibrated..``.1-1 Elemental Random Standard Uncertainties Associated With Error Sources Identified in Para. qR
−0.1-2 and 10-4.07
Table 10-4.0 960.`. UR.`.2 through 104..170 0.1-3.6.`.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. Nonsymmetrical Systematic Uncertainty Case)
Absolute Systematic Standard Uncertainty.````````. The details of this comparison are discussed in para.`. 10-4.. uR
0.600
the test result introduces an error that was not accounted for in the reported uncertainty. 08/17/2010 06:03:02 MDT
.4
Absolute Random Standard Uncertainty.276
Combined Standard Uncertainty of the Result.95−
1.40 p adiabatic compressor efficiency
Simplifying assumptions for this relationship include (a) use of ideal gas properties for dry air (b) negligible potential energy change (c) negligible heat loss to surroundings (d) constant specific heats and a specific heat ratio of 1.. The details of the uncertainty analysis are discussed in paras..0
0. the following lists of elemental error sources were compiled for each of the measurements: (a) Compressor Inlet and Exit Pressure Measurements (1) random error associated with incomplete sampling of the average pressure over the duration of the test and random variability in the pressure measurement instrumentation (2) systematic error resulting from imperfect calibration and drift of the digital pressure indicator (3) systematic error resulting from environmental influences on the digital pressure indicator (4) systematic error resulting from imperfect spatial averaging (5) systematic error due to the inability of the impact pressure arrays to fully realize total pressure
--`.300 0..
␥p ratio of the specific heats. Units Xi sX
i
Symbol
P1 P2 T1 T2
Description
Inlet pressure psia Exit pressure psia Inlet temperature R Exit temperature R
14.

sR
0.00366
Absolute Expanded Uncertainty.00772
57
--`.`.021 0.48E–6
NOTE: (1) These systematic standard uncertainties have some components that are correlated.1 Typical Pressure and Temperature Locations for Compressor Efficiency Determination
Table 10-4.52 520.TEST UNCERTAINTY
ASME PTC 19.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.600
7.`..1-2005
Fig.00231
Absolute Combined Standard Uncertainty.51E–6 1.62E–6 [Note (1)] 2.14E–6 1.``.0 960..792 [Note (1)]
0.54E–6 5. (i b X )2
i
Absolute Random Standard Uncertainty Contribution.8355
Absolute Random Standard Uncertainty.`.```. bX
i
Symbol
P1 P2 T1 T2
Description
Inlet pressure Exit pressure Inlet temperature Exit temperature
Units
psia psia R R
Nominal Value.````````.1-3 Calculated Result
Absolute Systematic Standard Uncertainty.00367 −0. UR
0. bR
0. 10-4. 08/17/2010 06:03:02 MDT
.030 0. (i s )2
X
i
0.`.0
Absolute Sensitivity. The correlated terms are not shown in the table.21E–6 1.198 0.170 0.00629 +0.00309
Symbol
Description
Computed adiabatic compressor efficiency
Units
Nondimensional
Calculated Value.-`-`.70 95.00203
Absolute Systematic Standard Uncertainty Contribution.. Xi
14. R
0.`.646 [Note (1)] 0.300 0. sX
i
Parameter Information (in Parameter Units)
Absolute Systematic Standard Uncertainty.59E–6 [Note (1)]
1.53E–7 1..1-2 Independent Parameters
Uncertainty Contribution of Parameters to the Result (in Result Units2)
Absolute Random Standard Uncertainty.. i
−0. uR
0.
Table 10-4..0409 +0.

ASME PTC 19.1-2005

TEST UNCERTAINTY

**Table 10-4.3.2-1 Inlet and Exit Pressure Elemental Systematic Standard Uncertainties
**

Absolute Systematic Standard Uncertainty for Inlet [Note (2)] psia, b P1

0.02

Error Source

Calibration and drift of indicator [Note (3)]

**Source of Information for Estimation of Limits of Error
**

Calibration and drift uncertainty reported by calibration laboratory based upon pretest and posttest calibration Published information provided by indicator vendor Reported uncertainty provided by compressor vendor based upon test rig data Engineering judgment

**Limits of Error for Inlet [Note (1)] psia, B P1
**

0.04

**Limits of Error for Exit [Note (1)] psia, B P2
**

0.25

**Absolute Systematic Standard Uncertainty for Exit [Note (2)] psia, b P2
**

0.125

Environmental influences on indicator [Note (3)] Spatial averaging [Note (3)]

0.01

0.005

0.06

0.03

0.01

0.005

0.3

0.15

Realization of total pressure [Note (3)] Total

Negligible

Negligible

Negligible

Negligible

Eq. 6.8

N/A

0.021

N/A

0.198

NOTES: (1) All limits of error are estimated at 95% confidence. It is assumed that these estimates are based on large degrees of freedom and that the population of possible error values associated with each elemental systematic error source is normally distributed. (2) Elemental systematic standard uncertainties are calculated using eq. (4-3.5). (3) It is assumed that the indicated elemental systematic standard uncertainties for the inlet and exit pressure measurements are not correlated because the inlet and exit measurements are substantially different in magnitude. For this particular example, assuming that these are not correlated will elevate the estimated uncertainty in the test result.

(b) Compressor Inlet and Exit Temperature Measurements (1) random error associated with incomplete sampling of the average temperature over the duration of the test and random variability in the temperature measurement instrumentation (2) systematic error resulting from imperfect calibration and drift of the thermocouple probes (3) systematic error resulting from imperfect calibration and drift of the digital temperature indicator (4) systematic error resulting from imperfect calibration and drift of the cold junction reference (5) systematic error resulting from environmental influences on the digital temperature indicator (6) systematic error resulting from imperfect spatial averaging

58

(7) systematic error due to the inability of the stagnation probes to fully realize total temperature 10-4.3 Calculate Random and Systematic Standard Uncertainties 10-4.3.1 Random Standard Uncertainty. The elemental random standard uncertainties associated with the error sources identified in para. 10-4.2 were evaluated by calculating the absolute standard deviation of the mean for each measured parameter [see eq. (4-3.2)]. The results are summarized in Table 10-4.1-1. 10-4.3.2 Systematic Standard Uncertainty. The elemental systematic standard uncertainties associated with the error sources identified in para. 104.2 were evaluated and totaled for each measured

--`,,``,````````,`,,```,`,,-`-`,,`,,`,`,,`---

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

TEST UNCERTAINTY

ASME PTC 19.1-2005

**Table 10-4.3.2-2 Inlet and Exit Temperature Elemental Systematic Standard Uncertainties
**

Absolute Systematic Standard Uncertainty for Inlet [Note (2)] °R, b T 1

0.55

Error Source

Calibration and drift of thermocouple probes [Note (3)] Calibration and drift of indicator [Note (3)]

**Source of Information for Estimation of Limits of Error
**

Calibration and drift uncertainty reported by calibration laboratory based upon pretest and posttest calibration Calibration and drift uncertainty reported by calibration laboratory based upon pretest and posttest calibration Calibration and drift uncertainty reported by calibration laboratory based upon pretest and posttest calibration Published information provided by indicator vendor Reported uncertainty provided by compressor vendor based upon test rig data Engineering judgment

**Limits of Error for Inlet [Note (1)] °R, B T1
**

1.1

**Limits of Error for Exit [Note (1)] °R, B T 2
**

1.1

**Absolute Systematic Standard Uncertainty for Exit [Note (2)] °R, b T 2
**

0.55

0.4

0.2

0.2

0.1

Calibration and drift of cold junction reference [Note (4)] Environmental influences on indicator [Note (3)] Spatial averaging [Note (3)]

0.5

0.25

0.5

0.25

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

--`,,``,````````,`,,```,`,,-`-`,,`,,`,`,,`---

0.2

0.1

0.1

0.05

0.1

0.05

1.0

0.5

Realization of total temperature [Note (3)] Total

Negligible

Negligible

Negligible

Negligible

Eq. (6-2.1)

N/A

0.0646

N/A

0.792

NOTES: (1) All limits of error are estimated at 95% confidence. It is assumed that these estimates are based on large degrees of freedom and that the population of possible error values associated with each elemental systematic error source is normally distributed. (2) Elemental systematic standard uncertainties are calculated using eq. (4-3.4). (3) It is assumed that the indicated elemental systematic standard uncertainties for the inlet and exit temperature measurements are not correlated because the inlet and exit measurements are substantially different in magnitude. For this particular example, assuming that these are not correlated will elevate the estimated uncertainty in the test result. (4) It is assumed that the indicated elemental systematic standard uncertainties for the inlet and exit temperature measurements are correlated as an error in the common cold junction reference will cause equivalent errors in the inlet and exit measurements.

parameter as shown in Tables 10-4.3.2-1 and 104.3.2-2. 10-4.4 Propagate Random and Systematic Standard Uncertainties The individual parameter standard uncertainties are propagated into terms of the test result by a Taylor series expansion as given in Nonmandatory

59

**Appendix C. The absolute random standard uncertainty for the test result is [see eq. (7-3.1)]
**

sR p [( P1s P1 ) 2 + ( P2s P2 )2 + ( T1sT1 )2 + ( T2sT2 )2]

p 0.00231

1⁄ 2

(10-4.2)

where the absolute sensitivities are [see eq. (7-2.1)]

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

ASME PTC 19.1-2005

TEST UNCERTAINTY

P1 p

−[ ␥ − 1)/␥ ][( P 2/P 1) −1/␥][ P2 / ( P1 ) 2] [( T 2 /T 1 ) − 1]

(10-4.3)

**The uncertainty interval for the test result is [see eq. (7-5.3)]
**

R ± UR p 0.8355 ± 0.00772 (10-4.11)

p −0.0409(psia −1)

P2 p

[( ␥ − 1)/␥ ][( P 2/P 1) −1/␥][1/ P 1] [( T2 /T 1 ) − 1]

10-4.6 Report

(10-4.4)

p +0.00629(psia −1)

A summary presentation of the uncertainty analysis at the test result level is reported in Tables 10-4.1-2 and 10-4.1-3. 10-4.7 Importance of Applying Known Engineering Corrections As discussed in subsection 4-1, it is important that all known engineering corrections are applied and appropriate engineering relationships are used as part of the results analysis process. Failure to do so leads to additional errors in the reported test result which are not accounted for in the test uncertainty analysis. This could, in turn, lead to expression of an uncertainty interval for a test result that does not encompass the true value. For this particular example, the assumptions of the analysis method used to compute the adiabatic efficiency of the compressor were not identified as sources of error. To illustrate the potential significance of having overlooked these sources of error, the test result will be recalculated using more exact relationships that do not require the assumption of constant specific heats and the assumed specific heat ratio of 1.4. By definition, the adiabatic efficiency of a compressor is the ratio of the work input required to raise the pressure of a gas to a specified value in an isentropic manner to the actual work input.

p

ws wa

(10-4.12)

T1 p

[ T 2]{[(P 2 /P1 ) (␥−1)/␥] − 1}

[(T 2 − T1 ) 2] p +0.00367(°R−1)

(10-4.5)

T2 p

[− T1]{[( P 2 /P 1 )

( ␥ −1)/ ␥

2

] − 1}

[( T 2 − T 1 ) ] p −0.00203(°R−1)

(10-4.6)

Note: In computing the preceeding sensitivity coefficients, it was assumed that the specific heat ratio is independent of air temperature.

**The absolute systematic standard uncertainty for the test result is [see eq. (8-1.2)]
**

bR p [( P1 b P1 ) 2 + ( P2 b P2 )2 + ( T1 bT1 )2

+ ( T2 b T2 )2 +2 T1T2bT1T2]

1⁄ 2

p 0.00309

(10-4.7)

**where bT1T2p the covariance of the error sources common to T1 and T2 and is determined as (see subsection 8-1)
**

bT1T2 p 0.25(°R) · 0.25(°R) p 0.0625(°R 2)

(10-4.8)

**10-4.5 Calculate Uncertainty The combined standard uncertainty in the test result is [see eq. (7-5.1)]
**

uR p ͓͑b R͒ 2 + ͑ S R͒ 2͔

1⁄ 2

where wsp entropic compressor work wap actual compressor work Assuming negligible change in the potential energy of the gas being compressed and negligible heat loss to the surroundings, the adiabatic efficiency can be expressed as a function of stagnation enthalpies as follows:

p

60

h2s − h 1 h 2 − h1

(10-4.13)

p 0.00386

(10-4.9)

**and the expanded uncertainty of the result is [see eq. (7-5.2)]
**

U R p 2u R p 0.00772 (10-4.10)

The assumptions required for using this equation are presented in subsection 1-3.

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

TEST UNCERTAINTY
ASME PTC 19.7 and compared with the previously reported test result. The value for h2s.1-2005
Fig.14)
1 s p const
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. This illustrates
61
P
P2
(10-4. to determine the relative pressure corresponding to isentropic compression. (10-4. which accounts for variable specific heats. 10-4. The values for h1 and h2 are evaluated at the measured inlet and exit conditions. and h2s. however.7. must be evaluated at a state corresponding to the measured exit pressure but assuming isentropic compression has occurred.) where h1p stagnation enthalpy at measured inlet conditions h2p stagnation enthalpy at measured exit conditions h2sp stagnation enthalpy assuming an isentropic compression process An h-s diagram of the actual and isentropic processes of an adiabatic compressor is illustrated in Fig. failure to use the more correct engineering relationships would lead to presentation of an uncertainty interval for the test result that does not encompass the true value. 08/17/2010 06:03:02 MDT
.
P r2 p P r1
where P1p measured compressor inlet total pressure P2p measured compressor exit total pressure Pr1p the relative pressure determined from ideal gas property tables for air at the measured compressor inlet total temperature Pr2p the relative pressure at the exit corresponding to isentropic compression Next. h2. In this case. The results of this analysis are summarized in Table 10-4.7 The h-s Diagram of the Actual and Isentropic Processes of an Adiabatic Compressor (Used by Permission From McGraw-Hill Co. the value for h2s is evaluated from air property tables at a state corresponding to the value of Pr2 determined above.13) indicates an error in the previously reported test result that is greater than the magnitude of the previously reported expanded uncertainty. Comparison of the previously reported test result with the more exact value determined by eq. 10-4. ideal gas property tables for air are used to evaluate h1. Assuming ideal gas properties for dry air. This is done by first using the following relationship.

ASME PTC 19.1-2005

TEST UNCERTAINTY

**Table 10-4.7 Evaluation of Analysis Error
**

Symbol

h1 h2 Pr 1 Pr 2

Parameter Description

Units

Average Value

124.27

**The performance of a hydraulic system can be evaluated using Bernoulli’s equation.
**

Hp P v2 + +z 2g

h2s

UR ...

Stagnation enthalpy at measured Btu/lbm inlet total temperature Stagnation enthalpy at measured Btu/lbm exit total temperature Relative pressure at measured ... inlet total temperature Relative pressure at exit ... corresponding to isentropic compression Stagnation enthalpy at exit Btu/lbm corresponding to isentropic compression Computed adiabatic compressor ... efficiency using eq. (10-4.13) Previously reported adiabatic ... compressor efficiency using eq. (10-4.1) Previously reported expanded ... uncertainty in the test result Difference between results from ... eq. (10-4.13) and eq. (10-4.1)

231.06 1.2147 7.8931

where Hp total head, m Pp system pressure, Pa p fluid density, kg/m3 vp fluid velocity, m/s gp gravitational constant, 9.81 m/s2 z1p elevation, m The effective head produced at a specified pressure is a measure of pump performance. The effective head in terms of the measured test parameters and physical property data is

⌬P p P2 + 8 Q 2

+ ( z 2 − z 1) g

212.35

0.8248 0.8355

0.00772 −0.0107

2 d4 2

where

the importance of applying all known engineering corrections and using appropriate engineering relationships as part of the test results analysis process. 10-5 PERIODIC COMPARATIVE TESTING 10-5.1 Problem Definition Periodic testing of equipment is a common situation where measurement uncertainty must be considered. For this example, a pump is considered to perform consistently if the supply pressure measured at 100% of rated flow is consistent with prior test results. The pump design data is presented in Table 10-5.1-1 and the test data is presented in Table 10-5.1-2 After each test, it is necessary to evaluate the test results. As part of this effort the effect of measurement uncertainty must be considered. For the first test in Table 10-5.1-2 the available information is limited since it was a factory test. (See Figs. 10-5.1-1 and 10-5.1-2.) The conclusions that can be drawn from Fig. 10-5.1-3 are as follows: (a) The pump is operating consistently when compared to the factory test results since the uncertainty bands overlap. (b) The pump is operating better than the minimum required design condition. The confidence for this conclusion is better than 95%.

62

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

d 2p inside pipe diameter, m ⌬Pp pressure change between taps 1 and 2, Pa P 2p pressure at tap 2, Pa Qp fluid flow rate, m3/s z 2p elevation of pressure tap 2, m z 1p elevation of tap 1 or free-surface elevation, m The density can be estimated numerically using the relationship [6],

p 766.17 + 1.80396 T K −

3.4589 T 2 K 1000

(10-5.1)

where TK p TC + 273.15 p the absolute temperature and TC is the fluid temperature in Celsius Estimates of the density, , using eq. (10-5.1) are reported to have a systematic standard uncertainty of 0.587 kg/m3. The random error in the curve fit was judged to be negligible and so the random uncertainty is set to zero. During testing it is not feasible to operate exactly at a specified operating condition. For a pump test, the applied flow might be slightly different for each test. This will result in a slight change in the resultant differential pressure. The variation

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

TEST UNCERTAINTY

ASME PTC 19.1-2005

**Table 10-5.1-1 Pump Design Data (TC ؍20°C )
**

Flow

m /s

0.000 0.126 0.189

3

Differential Pressure

gpm

0 2000 3000

**Percent Rated Flow
**

0 100 150

kPa

827 689 552

psi

119.9 99.9 80.1

**Table 10-5.1-2 Summary of Test Results
**

Factory Test [Note (1)]

Raw Data Flow, m3/s P exit, Pa d exit, m z exit, m z inlet, m T, °C Resultants , kg/m3 ⌬P, Pa ⌬P, psi 0.126 .. .. .. .. 20.15

Field Tests

A

0.125 840000 0.254 2.40 15.00 19.71

B

0.126 845000 0.254 2.40 15.00 20.01

C

0.130 820000 0.254 2.40 15.00 20.31

D

0.123 836000 0.254 2.40 15.00 20.78

E

0.129 841000 0.254 2.40 15.00 21.10

. . . .

997.7 712000 103.3

997.8 718000 104.1

997.7 724800 105.1

997.7 707300 102.6

997.6 710200 103.0

997.5 726400 105.4

NOTE: (1) The factory test data only provides resultant information.

Fig. 10-5.1-1 Installed Arrangement

63

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

ASME PTC 19.1-2005

TEST UNCERTAINTY

Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS

--`,,``,````````,`,,```,`,,-`-`,,`,,`,`,,`---

Fig. 10-5.1-2 Pump Design Curve With Factory and Field Test Data Shown

Fig. 10-5.1-3 Comparison of Test Results With Independent Control Conditions

in flow may be handled by normalizing the test results. This is accomplished by adding an additional random term. A normalization coefficient can be estimated from the factory test data in Table 10-5.1-1. A best fit correlation of the data is

⌬P p 827,000 − 376,000Q − 5,710,000 Q 2 (10-5.2)

⌬P N p b( Q N − Q ) + ⌬ P

(10-5.3)

where ⌬PNp expected pressure change based on the nominal (specified) test conditions, Pa ⌬Pp measured pressure change, Pa QNp nominal (specified) test flow rate, m3/s Qp measured test flow rate, m3/s bp slope of eq. (10-5.2), Pa · s/m3

64

Licensee=Mott MacDonald Ltd/5956936002 Not for Resale, 08/17/2010 06:03:02 MDT

The normalizing coefficient is the slope of eq. (10.5.2) at the specified test conditions.

9178 T 1.2-2 shows the nominal value. For this type of evaluation.2 Comparison With Independent Control The field test data for the pump can be compared with the factory test results or the minimum rated pressure output. The uncertainties for each test are presented in Table 10-5.4)
At the 100% rated flow condition (see Table 105. and the value for the slope is b p −1 810 000 Pa · s/m3. eq.80396 − 6. with test A considered the control test. 10-5. The partial derivatives for eq.420. The partial derivatives necessary to estimate the sensitivity coefficients for this problem are
Symbol.2) is
∂ ⌬P N ∂Q
Έ
QN
p b p −376. Xi
P2
Q
d2
where
p ratio of two resultants ⌿altp alternate (or variable) resultant ⌿controlp resultant used for the baseline or control For the pump test C. and total uncertainties for ⌬P.2. Table 10-5.126 m3/s. The uncertainty is calculated using the method in subsections 7-1 through 7-4.1-1). (10-5. QN p 0.80396 − 1000
K
b a p
TC p
a
p2 p
c
Qc p
d2 p
a
− ∂⌬Pc ⌬Pc ∂d2c
65
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.3 Comparative Uncertainties Correlation of terms is an important consideration in comparison testing [23. the uncertainties are independent and a simple comparison of the test results with the benchmark value is adequate.1-2005
The slope of eq.
P2 p
a
QN − Q 1.4) are
∂ p ∂⌿a ⌿a
− ∂ p ∂ ⌿c ⌿c
2d 4 2
−b
2
−32Q
2d 5 2
8Q
2
z1 z2 b TK
2d 4 2
+ (z2 − z1 ) g
g
− g
These may be combined with the partial derivatives presented earlier to estimate the sensitivity coefficients for the comparative analysis. and the systematic.4) becomes
A p
707 kPa p 0.2-3.2-3. random. The results are plotted in Fig.000 Q N
The uncertainty for each test can be calculated as shown in Table 10-5.000 − 11. 24] where two different operating conditions or constructions are being compared by use of a ratio
p ⌿ alt ⌿ control
(10-5. The final data reduction equation becomes:
⌬P p P2 + 8 Q 2
+ ( z 2 − z 1) g + b ( Q N − Q )
2d 4 2
10-5. (10-5. 10-5. (10-5.2-1.985 718 kPa
Formulas for Absolute Sensitivity
∂⌬P p1 ∂P 2
16Q
The test results for the comparative analysis are summarized in Table 10-5.9178TK 1000
∂⌬Pa ⌬Pa ∂P2a
∂⌬Pa ⌬Pa ∂a
∂⌬Pa ⌬Pa ∂ba
− ∂⌬Pc ⌬Pc ∂P2c
Qa p
∂⌬Pa ⌬Pa ∂Qa
∂⌬Pa ⌬Pa ∂z2a
∂⌬Pa ⌬Pa ∂Tca
− ∂⌬Pc ⌬Pc ∂Qc
d2 p
a
∂⌬Pa ⌬Pa ∂d2a
∂⌬Pa ⌬Pa ∂z1a
The temperature sensitivity coefficient is computed using the chain rule:
∂ ⌬P N ∂ ∂ ⌬P N 8Q 2 p p + (z 2 − z1) g ∂ TK ∂ ∂T K 2d 4 2
a p
z2 p
a
z1 p
a
6.TEST UNCERTAINTY
ASME PTC 19. The uncertainty for the comparative analysis can be computed using the method from subsection 8-1. 08/17/2010 06:03:02 MDT
.

9916 1. Contribution.76 −120.0069 0.0168 0.ASME PTC 19..0048 0.0؋10−3 0..0169 0. 08/17/2010 06:03:02 MDT
. uR
7.71 997.9889 1.95.2-3 Summary of Results for Each Test
Test
Factory Test A B C D E
⌬P.23؋107 2. sR
3.0؋103 0.29؋103 1.0
3.0 −47810 9786 −9786 26.46؋106 9.0091 0. kPa
.0؋106 0 0 0 7.23؋103 25.0000 1..0125 0.0097 0.0049 0.0207 0. 15 15 15 15 15
0.0070 0. bX
i
Uncertainty Contribution of Parameters to the Result (in Result Unit Squared)
Absolute Random Standard Uncertainty.95
14..```.0049 0.2-1 Uncertainty Propagation for Comparison With Independent Control
Independent Parameters
Parameter Information (in Parameter Units)
Absolute Systematic Standard Uncertainty.`.5 1.25 0.40 15...16 0 0
Table 10-5.590
Symbol
⌬P
Description
Differential pressure
Units
Pa
Calculated Value
718.8 5.5 6. 6. i
1.0؋10−3
Absolute Absolute Systematic Random Standard Standard Uncertainty Uncertainty Contribution.10 0 0
3.125 840. bR
6.2-2 Summary: Uncertainties in Absolute Terms
Absolute Systematic Standard Uncertainty.95
0.`.0؋10−3 3500 1.````````.0070
U.0069 0.254 2. sX
i
Symbol
Q P2 d2 z2 z1 Tc b
Description
Units
Nominal Value
0. 4 4 4 4 4
U⌬P.`..530
Combined Standard Uncertainty of the Result.5 6.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.8 −1.0049
s
0.11؋107 1.0168 0.951
Table 10-5.82؋106
Absolute Sensitivity.`.476
Total Absolute Uncertainty.000
Absolute Random Standard Uncertainty.. kPa
.5 6.6 5000
1.1-2005
TEST UNCERTAINTY
Table 10-5. kPa
712 718 725 707 710 726
b X.0049 0. UR.0170 0.9847 0.00 19..-`-`..50؋104 44.``.0؋10−3 3000 0 0 0 0.86؋106 1.0170
66
--`.50؋104 1.`..0125 0.0111
b
0.5
s X. kPa
.0048 0.5 6. (b X i )2 (s X i )2
i
i
Flow m3 Exit pressure Pa Exit diameter m Exit elevation m Inlet elevation m Fluid temperature °C Fluid density kg/m3 Correlation Pa·s/m3 coefficient
3.0069 0..

The absolute systematic standard uncertainty of the result (from Tables 10-5. The exception is the calibration standard uncertainty. Standard protocol is to bring a replacement gage to the test location on the day of the test and replace the gage just prior to the test.`. For test C. all of the systematic standard uncertainties are considered fully correlated except for the exit pressure that is only partially correlated. This calibration standard uncertainty is treated as a fully correlated standard uncertainty associated with the pressure gage. For eq.. when all of the systematic standard uncertainty terms are correlated and the corresponding sensitivity coefficients are equal (i.control) is zero.e. SR.-`-`.2).```.1-2005
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`. with test A considered the control test.````````.500 Pa.`.. (10-5.4) the systematic standard uncertainty summation based on eq. from Table 10-5.. (8-1.3-1 is 0..TEST UNCERTAINTY
ASME PTC 19. much of the systematic error is not correlated.500 Pa. i.`.`---
Fig.0069. The partial correlation for the pressure measurement must be derived from the elemental uncertainties. the absolute random standard uncertainty component..2-3 and 10-5. associated with each gage is 3. 08/17/2010 06:03:02 MDT
. b2 p P2 bP2
a
+ b + b + b + b + b + b + b + b + b + b + b + b + b + b
2
a
Qa
Qa
2
d2
a
d2
2
a
a
a
2
z2
a
z2
2
a
z1
a
z1
2
a
Tc
a
Tc
2
a
BN
a
BN
2
a
P2
c
P2 2
2
c
Qc
Qc
2
d2
c
d2
2
c
c
c
z2
c
z1
2
c
TC
c
TC
2
c
BN
c
BN
2
c
+ 2 P2 P2 b P2 P2 + 2 Qa Qcb QaQc + 2 d2 d2 b d2 d2
a c a c a c a
c
+ 2 a c b a c + 2 z 2 z 2 b z 2 z 2 + 2 z 1 z 1 b z 1 z 1
a c a c a c a
c
+ 2 Tc Tc bTc Tc + 2 bN bN b bN bN
a c a c c a c a
c
For this example. 10-5.2 Comparison of Test Results Using the Initial Field Test as the Control
The uncertainty of the comparison ratio. Thus.. which is known to be 2. For this program the test procedure requires that the pressure gage be calibrated just prior to conducting the test. Since the gage is randomly selected and several gage brands are used..`. . bP. each test is conducted with a different gage.3-1) would be
67
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.alt p i.`. The absolute systematic standard uncertainty.``.

76 −120.0 0.74؋10−5 0.6 5.0
0.0169
68
--`.29؋10−9 2.0 0.39؋10−9 0.56؋10−9
bi
a
ic
−2.0125 0.37؋10−5 4.55 −1.3-1 Uncertainty Propagation for Comparative Uncertainty
٢ ٢⌬p
Variable Test C
Q P2 d2 z2 z1 TC b
Control Test A
Q P2 d2 z2 z1 TC b −1.37؋10−6 −1.25 0.000 5.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.500 0.`.0
6.37؋10−6 −1.`.`.6 5000
−1.68؋10−4 5.86؋106 1 −47.0
0.39؋10−6 1.9 ؋ 10−3)2 p 8.0 0.0134 0.0 0.84 ؋ 10−3
The combined standard uncertainty of the result is
uR p
Ί (4.43 ؋ 10−3
The uncertainty for each test can be calculated as shown in Tables 1-5.001 0.001 0.5 0.```.0 0..0134 0.39؋10−6 1.37؋10−6 0.67؋10−5 1.0134 −3.0 0.786 −9.003 2500 0..0 0.0136 −0. The results are plotted in Fig.78؋10−5 −1..39؋10−6 1.
The total comparative uncertainty
U R p 2 u R p 0.0
−1.0 0.0 0.76 −120.0 0..593 1.65؋10−4 −1.6 5000
0.0 0.37؋10−6 −1.1 0.001 0.37؋10−9
bi
c
2i i bi i
a c
ac
0.0 0.0
6.0 0.39؋10−6 1.50؋10−6 1.0 0.3-2 Sensitivity Coefficient Estimates for Comparative Analysis
ia
Q P2 d2 z2 z1 TC b 2.000 0.000
6.0125 0.66؋10−5 1.810 9.19؋10−4 −2.39؋10−9
0.0010 −2.0665 0.37؋10−6 −1.786 26.810 9.37؋10−6 −1.89؋10−8 2.43 ؋ 104) p 4.22.37؋10−6 1.0372 −1.1-2005
TEST UNCERTAINTY
Table 10-5.39؋10−6 1.39؋10−6 1.0 0.0125 0.0125 0.37؋10−6 −1.ASME PTC 19.0010
2.552 −1.0129)2 + (−1.0125 0.25 0.86؋106 1 −47.3-1 and 10-5.`.0136 −0.0655 −0. 08/17/2010 06:03:02 MDT
.0125 0.0 0.43؋10−4
(RSS sum)
bR p
Ί (0.25 0.``.38؋10−5 0.37؋10−6 −1.000 0.003 3.84 ؋ 10−3)2 + (6.89؋10−8 0.003 2500 0.001 0.85؋10−5 2.0136 3.0125 0.2.04؋10−5 2.5 0.0 0.80؋10−9 4.39؋10−6
1.786 26.81؋10−8 0.69؋10−11
0.0 0.39؋10−6 −0.0719 0.786 −9..0 0.500 0.0 0.81؋10−8 2. The uncertainties for each test are presented in Table 1-5.65؋10−4 −1.0 9.0069
(RSS value)
Table 10-5.25 0.39؋10−6 1.0 0.````````.71؋10−6 1.42؋10−9 2.37؋10−6 0.1 0.83؋10−11
0.0136 0.3-2..590 1.02؋10−8 4.68؋10−4 1.39؋10−6 −0.0 0.6 5.-`-`.0655 −0.0 1. The test conclusion that can be drawn from this figure is the pump is operating consistently when compared since the uncertainty bands overlap.001 3.30؋10−5 4.`.69؋10−5 0.003 3. 10-5.001 3.0125 0.0129
٢⌬P ٢ xi
i
bi
(bii )2
sX
i
(s X i )2
i
1.0 0.0134 −3.

Estimating the Uncertainty of Spatial and Time Average Measurements.`.. Journal of Thermophysics and Heat Transfer. 3rd edition. A.. K. M. ASME PTC 19. S. Measurement Uncertainty. October–December 1993.`. Transactions of the ASME: Journal of Engineering for Power: 473–476.`---
[9]
[20]
[21]
[10]
[22]
[11]
69
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.``. ISO/TR 7066-1: 1997 (E). Experimentation and Uncertainty: Experimentation Uncertainty Analysis for Engineers. Evaluation of Correlated Bias Approximations in Experimental Uncertainty Analysis. and W.````````. S. 6th edition. AIAA Journal. 08/17/2010 06:03:02 MDT
.`. 7: 754–757. 1994.” Geneva: International Organization for Standardization. October 1975. New York: The American Society of Mechanical Engineers. May 1996. New York: The American Society of Mechanical Engineers. [12] Wyler. 2nd edition. 1998.” Geneva: International Organization for Standardization. Fluid Meters — Their Theory and Application. et al. H. G. AIAA Journal. et al.” Transactions of the ASME: Journal of Fluids Engineering. D. New York: McGraw-Hill.. 34: 1458– 1463.```. 31: 1891–1896. Steele. Montgomery. 1987. Holman. “Identifying the True Value — The First Step in Uncertainty Analysis. 93-107. L. Asymmetric Systematic Uncertainties in the Determination of Experimental Uncertainty. 34: 1013–1018. AIAA Journal.`.
[2]
[13]
[3]
[14]
[4]
[15]
[5]
[6]
[16]
[7]
[17] [18] [19]
[8]
--`. Experimental Methods for Engineers. Coleman. Steele. Use of Previous Experience to Estimate Precision Uncertainty of Small Sample Experiments. Brownlee. Statistical Theory and Methodology in Science and Engineering.” ISA paper 88-0729. Test Uncertainty.. New York: John Wiley and Sons.. G. R. James. and D... K.. 2nd edition. Peck. 104: 250–260. J. Brown. W. NIST Technical Note 1297. 1967. 2000. 6th edition. New York: American Society of Mechanical Engineers. 3rd edition. and E. W.. Figliola. “Uncertainty of Derived Results on X-Y Plots. A. E. 1992. W.-`-`. Price. R. Measurement Error Models.1. October 1993. “Bias Error Reduction Using Ratios to a Baseline Experiment — Heat Transfer Case Study. R. Theory and Design for Mechanical Measurements. Chakroun.. New York: John Wiley and Sons. K. J. P. New York: John Wiley & Sons. “Assessment of Uncertainty in Calibration and Use of Flow Measurement Devices — Part 1: Linear Calibration Relationships. 1971.. 1999. M. 1985.TEST UNCERTAINTY
ASME PTC 19.1. Beasley. “Contributions to the Theory of a Single-sample Uncertainty Analysis. Steele. ISO 7066-2: 1988(E). “Assessment of Uncertainty in the Calibration and Use of Flow Measuring Devices — Part 2: Non-linear Calibration Relationships. C. et al. 2nd edition. Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurements. L. Applied Numerical Methods for Digital Computation.. J. Moffat. New York: Harper Collins. July 1996. W. J.” ISA paper No. A. June 1982. 1992. 1997. Introduction to Linear Regression Analysis. et al. New York: The American Society of Mechanical Engineers. W. New York: John Wiley and Sons. Fuller.” Geneva: International Organization for Standardization. G. 1995.`. ASME PTC 19. New York: John Wiley and Sons. Moffat. 1988. 1994.1-2005
Section 11 References
[1] ISO “Guide to the Expression of Uncertainty in Measurement.

W. 1935.” Technome-
[27]
[28]
[29]
[30]
trics. 117: 552–556. October 1995. “Engineering Application of Experimental Uncertainty Analysis. “Comparison of ANSI/ ASME and ISO Models for Calculation of Uncertainty.`. 45–48. H.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. December 1970. Denver. February 1969..” Transactions of the ASME: Journal of Fluids Engineering.. W.. P. 1994.1-2005
TEST UNCERTAINTY
[23]
[24]
[25]
[26]
Coleman. and R.” Industrial Engineering.. and W. “Understanding Statistical Intervals.```. W. 1995. R. Grubbs. 1997. H.`.ASME PTC 19. “Procedures for Detecting Outlying Observations in Samples.. Hahn. G.” Annals of Mathematical Statistics. “On a Criterion for Rejection of Observations and the Distribution of the Ratio of the Deviation to Sample Standard Deviation. W. W.`.``. W. et al.. December 1995. “Computer-Assisted Uncertainty Analysis. Strike.” ISA Transactions.`.````````.. New York: The American Society of Mechanical Engineers.” Proceedings of the 41st ISA International Instrumentation Symposium.-`-`. G. T.
70
--`. 11: 1. et al. Thompson. 33: 339–352. Taylor. 08/17/2010 06:03:02 MDT
. G. G. 33: 1888–1896..`. “Implications of Correlated Bias Uncertainties in Single and Comparative Tests.” Computer Applications in Engineering Education... Steele. Steele. Dieck. “Rocket Impulse Uncertainty. Coleman. W. E. G. F.” AIAA Journal. 6: 214–219. Steele. and R.. Steele. H.

“Experimental Statistics. Williams.” Transactions of the ASME: Journal of Engineering for Power. J. Pressure. Statistical Theory With Engineering Applications. 1984. 1981. 21. A. Benedict. p. R. 1977. 1967. and H. J. Natrella.” Mechanical Engineering Publication. New York: Hafner. O. “Engineering Analysis of Experimental Data. Mandel. “Fitting Straight Lines When Both Variables are Subject to Error. New York: John Wiley and Sons. New York: John Wiley and Sons. M. 1952. 2nd. Fundamentals of Temperature. Applied Regression Analysis. “Estimation of Linear Function for a Calibration Line. 3rd. edition. E. New York: John Wiley and Sons. Smith. Hayward. “Guide to In-flight Thrust Measurement of Turbojets and Fan Engines. January 1969. T. 15: (1). edition. 1959. New York: John Wiley and Sons. 2nd edition. J. Design and Analysis of Industrial Experiments. Measurement Uncertainty for Fluid Flow in Closed Conduits.” Journal of Quality Technology. Berkson. Hald. G. Draper. R. 1983. “Repeatability and Accuracy.” National Bureau of Standards Handbook 91. New York: The American Society of Mechanical Engineers.TEST UNCERTAINTY
ASME PTC 19.1-2005
Section 12 Bibliography
[1] Benedict. J. 1963. New York: The American Society of Mechanical Engineers. Regression Analysis. N. P. ASME MFC-2M..
[2]
[8]
[9]
[3]
[10]
[4]
[11]
[5] [6] [7]
[12]
71
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. 08/17/2010 06:03:02 MDT
. and Flow Measurements. R.” Technometrics.” Davies. 11: (4). L.. P. Ltd. A. AGARD Report AG-237. January 1984. November 1969.

.--`. 08/17/2010 06:03:02 MDT
.-`-`...`..```.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.`.`.`..``..`.````````.

(The use of tolerance in this instance is different from that of PTC 1.1-2005
Nonmandatory Appendix A Statistical Considerations
A-1 UNDERSTANDING STATISTICAL INTERVALS It is often desirable to use the terms confidence. For a normal distribution.`.48 to 51. but differs from it because of sampling fluctuations. the procedure for constructing them
73
--`.10 ± (1. It is calculated as:
y ± c M ( n) s. and prediction interval. Some are also aware of tolerance intervals.. the sample size. if and are known exactly.95. Thus a 95 percent confidence interval on the population mean is an interval which contains with a probability of 0.`---
is no more difficult.
TOLERANCE INTERVAL TO CONTAIN A SPECIFIC PROPORTION OF THE POPULATION.) An eloquent treatment of these terms is given by Gerald Hahn [24] in his 1970 paper.
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. make limited reference to tolerance intervals. a confidence interval to contain . and given the required tabulations. contains the quantity of interest. Many nonstatistical users of statistics are well acquainted with confidence intervals.24)(1. This confusion is understandable since most texts on statistics devote extensive space to confidence intervals on population parameters. Statistical intervals are frequently misunderstood and misused. tolerance. it is possible to construct a statistical interval known as a confidence interval for the population mean .
where cM (n) is obtained from the first column of Table A-1. 08/17/2010 06:03:02 MDT
.`.```. tolerance. the interval calculated in this manner will contain the unknown mean 95 percent of the time. This probability is known as the associated confidence level.-`-`. This is unfortunate because tolerance intervals or prediction intervals are needed as frequently in industrial applications as confidence intervals.. a tolerance interval is used when a prediction interval is needed.. but most nonstatisticians know very little about prediction intervals despite their practical importance.ASME PTC 19.1 lists the information needed to construct all three intervals. which is reproduced here by permission. Instead of. However. (2) a tolerance interval to contain a specified proportion of the population. one can be 95 percent confident that the interval 48. Here is an explanation of when to use confidence.````````.. This article describes three different types of statistical intervals and shows where each should be used.
Consequently. CONFIDENCE INTERVAL FOR THE POPULATION MEAN The sample mean y is an estimate of the unknown mean . and almost never talk about prediction intervals. The most they can do is obtain an estimate and construct an interval which.`. At other times. and (3) a prediction interval to contain all of a specified number of future observations. The three intervals are: (1) a confidence interval to contain a population mean.24 and the 95 percent confidence interval for is:
50.. with a high probability.`. Table A.``.1 as a function of n. Engineers have come to appreciate that few things in life are known exactly.`. or in addition to.. More precisely.31). over a large number of samples. This interval contains with a specific probability. and prediction intervals.72 contains the unknown value of . For the example shown in the box cM (5) p 1. many applications require an interval to enclose a specific proportion of the population. A frequent mistake is to calculate a confidence interval on the population mean when the actual problem calls for a tolerance interval or a prediction interval. it can be stated that 90 percent of the population is located in the interval.

64 0. and cT. 20 (n)
7.21 3.37 2.75 2. Such an interval is called a tolerance interval and can be calculated for a normal population with the help of the factors cT.
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.1.90(n).69 3. For example. 5 (n)
5.02 1.63 4.29 2.95) the interval contains at least 90.77 3. 10.24
cP. cT. it can be stated with 95 percent confidence that the interval:
y ± c T. Another type of interval is one that will contain all the values of one or more future observations.83 2.24 3.39 2.57 3.45 2.51 2.99 2.64
However.49 to 55.31 2. The first of these numbers refers to the proportion (or percentage) of the population that the interval is to contain.11 2.31 2.90 (n ) s
The fact that both a population proportion (or percentage) and a statistical probability (also a percentage) are associated with a tolerance interval is sometimes confusing to the engineer.06 2.24 1.96 1.79 2.k(n) such that all of k future observations from the same normal population will be located in the interval:
74
contains at least 90 percent of a normal population The tolerance interval for the example in the box may be calculated as:
50.89 4.1 provide values of the factor cP.41 4.95(n).37 3.10 2.73 3.31)
or 44.77 0. 95 (n)
6.04 2.1 Factors for Calculating the Two-Sided 95% Probability Intervals for A Normal Distribution
Number of Given Observations
n
4 5 6 7 8 9 10 11 12 15 20 25 30 40 60 ϱ
Factors for Confidence Interval to Contain the Population Mean
cM (n)
1.37 5.25 3. 3.28. the best that can be stated is that with a chosen probability (say 0.48 2. The last five columns of Table A-1.25 4.26 0
Factors for Tolerance Interval to Contain at Least 90%.92 0.74 4.
± 1. 5.47 0.58
cP.96
cT. Thus.37 0.84 2.17 3.08 2.70 3. one may be 95 percent confident that the preceding interval contains at least 90 percent of the sampled population.68 2.48 2.72 3.08 4.86 2.78 2.28 3.99(n).67 2.56 3.50 3.63 2.36 3.63 5.22 2.10 ± (4.16 4.57 2. 95% and 99% of the Population
cT.53 3.41 0.66 2.30 6.05 2. PREDICTION INTERVAL TO CONTAIN ALL OF A SPECIFIED NUMBER OF FUTURE OBSERVATIONS.01 3.35 2.28)(1.32 0.97 2.43 4. This is known as a prediction interval. 99 (n)
8.53 3. or 99 percent of the population.14 2. in this case. 95.26 3.63 4.23 4. 10 (n)
6.90(n)s p 4.98 3.67 0. but.05 0.84 0.72 0.46 3.55 0.37 4.15 3.90 2.62 3.17 3.71 3.21 3.28 4. The second number specifies the probability that the calculated interval really contains at least the specified proportion of the population.16 2.26 4.40 3. 90 (n)
5.73 2.02 3.41 3.46 4.08 3.32 3.33 3.80
cP.78 2.07 2.10 3.89 3. 08/17/2010 06:03:02 MDT
. and 4 of Table A-1.55 2. and 20 Future Observations
cP. 2 (n)
4.1-2005
TEST UNCERTAINTY
Table A-1.43 2.33 3.78 5.21 5.58 4.85 3.33 1. shown in columns 2. an interval to contain a specified proportion of the population may still be of interest.43 2.72 2.ASME PTC 19.62 3.33 2.97 2.96
cT.26 4.64
Factors for Prediction Interval to Contain the Values of All of 1.03 2.74 2. When and are known exactly. if only sample estimates y and s of the population values and are given. there is no longer any uncertainty associated with the proportion of the population contained in the interval.62 2.85 5.41 5.93 2.71 where cT.05 1.57
cP.02
A two-sided 95 percent interval is y ± c(n)s.12 3.95 2. 1 (n)
3.14 2.43 3.59 1. 2.14 2. where c(n) is the appropriate tabulated value and y and s are the mean and the standard deviation of the given sample of size n.56 4.21 2.38 3.88 3.35 3.

Thus two future units from the sampled population will be located in the interval:
50. Inspection of the tabulations indicates that a confidence interval on the mean is always smaller than the other two intervals. the mean is one of the two parameters which uniquely defines the distribution.10 ± (3. It is seen that for the given sample of 5. Also. the length of a confidence
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.31)
HOW TO SELECT THE RIGHT INTERVAL.95 is somewhat larger than a prediction interval to contain both of two future observations. therefore. The mean is the most commonly used single value to describe a population. the confidence interval to contain the population mean is appreciably smaller than both the tolerance interval and the prediction interval. The engineer must decide upon the relevant questions. k p 2 and n p 5. The following comments are offered to serve as a guide to the engineer in this process. unlike the other two intervals.
interval approaches zero as the sample size increases (the interval converging to the point ). It is identical to the median (50 percent point) and mode (most common value) of the distribution.95. A-1 How the Lengths of the Statistical Intervals for the Example Compare
y ± c P. For example. with a probability of 0.70)(1. 08/17/2010 06:03:02 MDT
. and is often used as a standard by which competing processes are compared.k (n ) s.2(5) p 3.95.95. From Table A-1.70. Its use for such comparisons is especially appropriate when it is reasonable to assume that each of the competing processes has the same statistical variability (as measured by the process standard deviation) and. A-1.TEST UNCERTAINTY
ASME PTC 19.
with a probability of 0.1 the factor cP. it should be easy to decide upon the correct intervals. The statistician’s job is to develop correct procedures for answering relevant questions. but that the relative sizes of the tolerance and prediction intervals depend upon the proportion of the population to be contained in the prediction interval.1-2005
Fig. if two additional readings are taken from the example in the box. For the normal distribution. The relative lengths of the three intervals obtained in the preceding examples are compared in Fig. Also a tolerance interval to include at least 90 percent of the population with a probability of 0. Once the questions to be answered have been clearly stated.25 to 54. The population mean is therefore of great interest in characterizing product performance. the differences between processes can be described
75
or 45.

Standard books on elementary engineering statistics. 1969.
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.. A typical astronaut. Finally. This article also considers additional types of statistical intervals such as: A prediction interval to contain a future sample mean. also
76
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
discuss tolerance intervals. are discussed in References 1. + (51. Thus. Such intervals. while a prediction interval would be used to obtain limits to contain all of a small number of future units from the population. 3. . 49.4 + . . Further. give prime space to the concept of confidence intervals and. The assumption of equal standard deviations is frequently made. Assume that readings obtained on a normally distributed performance parameter based on a random sample of five units are: 51. you are interested in estimating the range of variation of the underlying population or of the observations in a future sample.7. September. 2. A prediction interval to contain a future sample standard deviation.” Journal of the American Statistical Association. His main concern is the worst that will happen in the one.. including confidence. Specifically.ASME PTC 19. in many cases.1-2005
TEST UNCERTAINTY
completely by differences in their means.4 − 50. Reference 4. 1970.6. 49. Similarly.
THE EXAMPLE PROBLEM The calculation of the three intervals are illustrated here by the following numerical example.” Journal of the American Statistical Association. a tolerance interval is applicable if limits are needed that contain most of the sampled population.. Reference 6. Because of random fluctuations. and 51. If. a confidence interval is established which contains the unknown value of with a specified degree of confidence. . From this informa¯ and the sample standard deviation. the sample mean y tion s are calculated by well-known expressions:
yp
͚ yi / n p (51. an engineer who is concerned with the performance of a mass-produced item. . WHERE TO GET MORE INFORMATION. would use a prediction interval to obtain specification limits to contain the performance parameter for all three units with a high probability. then a tolerance interval or a prediction interval is needed. but make no mention of prediction intervals except in a regression context. a new time-sharing computer program calculates a wide variety of statistical intervals. “Factors for Calculating Two-Sided Prediction Intervals for Samples from a Normal Distribution. In contrast. Reference 3 provides a comprehensive comparison of statistical intervals for a normal population (including more detailed tabulations than are given here) and a discussion of methods for constructing the various intervals. Hahn. or five flights in which he will be personally involved. Thus. 48. 65.31. G. 64. or even what will happen in at least 99 percent of such flights (tolerance interval).10 ip1
n
sp
΄͚
n
ip1
¯ )2 ( yi − y
n−1
΅
1⁄
2
p
Ά
͓(51. Hahn. J. a prediction interval to contain all of k future observations may be thought of as the astronaut’s interval. 2. who has been assigned to a specific number of flights. however.5. G. “Additional Factors for Calculating Prediction Intervals for Samples from a Normal Distribution.
REFERENCES 1. yn are the values of n given observations. . + 51.3. instead of characterizing typical process performance.6)/5 p 50. A confidence interval for the population standard deviation. and 5. is generally not very interested in what will happen on the average in the population of all space flights. of which his happen to be a random sample (confidence interval on the mean). December. a turbine engineer who is bidding on an order of three units based upon his past experience on five units of the same type. a sample does not provide perfect information about the population mean . and prediction intervals. such as a transistor or a lamp.6 − 50.10) 2 + . three. 08/17/2010 06:03:02 MDT
. . .4. tolerance.
where y1 . A confidence interval for a population percentile.10)2͔
(5 − 1)
·
1⁄
2
p 1. J. Prediction intervals are also required by the typical customer who purchases one or a small number of units of a given product and is concerned with predicting the performance of the particular units he has purchased (in contrast to the long-run performance of the process from which the sample has been selected). would generally be interested in a tolerance interval to enclose a high proportion of the sampled population.

Number 3.” Journal of Quality Technology. The weighting techniques described in this section should not be employed if the uncertainty intervals do not overlap to a significant degree as this is a possible indication of unaccounted for uncertainties. which is the variance of the average measurement. (A-1) yields
W1 p
͑ U X 2͒ 2
͑ U X 1͒ 2 + ͑ U X 2͒ 2
and
W2 p
͑ U X 1͒ 2
͑ U X 1͒ 2 + ͑ U X 2͒ 2
p 1 − W1
Using these same weighting factors. then the test result and its associated uncertainty may be determined by weighting the means. “Statistical Intervals for a Normal Population. Volume 2.TEST UNCERTAINTY
ASME PTC 19.” General Electric Company TIS Report 68-C-404. 4. the means and their associated uncertainty intervals from each of the independent methods should be compared as discussed in subsection 8-5. This is combined with (SXi)2. W. A-2 WEIGHTING METHOD Whenever the value of a test result is determined by several independent methods. Therefore. 6. This is true since the systematic uncertainty component of a parameter BXi is assumed to equal two times the standard deviation (square root of the variance) of the possible distribution of systematic uncertainty. Volume 2. October 1970.. the systematic and random uncertainties of the weighted mean are given by the root-sum-square relations
BX p
͑ W iBX ͒ ΅ ΄ i͚ p1
2
i
N
1⁄
2
SX p
͑ W iSX ͒2 ΅ ΄ i͚ p1
i
N
1⁄
2
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. US Government Printing Office. Nelson. Building 5.1-2005
3. eq. UXi is a combination of variances:
U Xi p 2
΄ 2
BXi
2
+ ( S X i) 2
΅
1⁄
2
Therefore. A weighting principle which is statistically valid and is based on weighting by variances is applicable in this case [5]. J. The advantages of utilizing the weighting techniques described in this section are that the uncertainty associated with a weighted test result will usually be less than the uncertainty of the results determined from each of the independent methods by themselves. November 1968. Experimental Statistics. The weighting methods presented in this section are based on the following assumptions: (a) The various methods used to determine the test result are independent to the extent that there is no appreciable correlation between the sources of uncertainty of the various methods. For two measurement methods with the means X1 and X2. Bethesda. “Two-Sample Prediction. Mary Gibbons. General Electric Information Service Department. “Summary Statistics Package — ONE-SAMS***” Document 003401. Prior to using the weighting techniques described in this section. pages 115–125. and systematic uncertainties of the various methods. Schenectady. Natrella. 5. the weighted mean of the measurements. The variance of this distribution is (BXi/ 2)2. July 1970. National Bureau of Standards Handbook 91. PO Box 43. Maryland.
Wi p
ip1
1 ͚ U
1 U Xi
N
Xi
2
2
(A-1)
where UXi represents the uncertainties of X i. Let Xi represent best estimates of a parameter by N measurement methods. (b) The assumptions presented in subsection 1-3 are valid for each of the independent methods such that the uncertainty model presented in subsection 1-3 may be used. 08/17/2010 06:03:02 MDT
. B.. Number 4. Then X . can be given by
77
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
X p
͚ W iX i ip1
N
where Wi are the required weighting factors. Hahn. 7735 Old Georgetown Road. Room 237. random uncertainties. pages 195–206. G. (Available from Distribution Unit. New York 12305).

`.ASME PTC 19. bad data points are hard to identify. These points may be caused by temporary or intermittent malfunctions of the measurement system. and should be discarded. The effect of outliers is to increase the standard deviation of the system.5) as follows:
U X p 2 ͓͑B X /2͒ 2 + ͑S X ͒ 2͔
1⁄ 2
As usual. For small samples. The odds could be increased by setting the probability of rejecting a good data point lower.`.1 General All measurement systems may produce spurious data points.``. but also may reject some
Fig.`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. This means that the odds against rejecting a good point are 20 to 1 (or less).
A-3 OUTLIER TREATMENT A-3.1-2005
TEST UNCERTAINTY
These values are combined as UX to obtain the weighted uncertainty of X according to eq.. Tests are available to determine if a particular point from a sample is an outlier.. To ease the burden of scanning large masses of data... These are the Thompson Technique [25] and the Grubbs Method [26]. computerized routines are available to scan steady state data and flag suspected outliers..`. Such points are considered to be meaningless as steady-state test data. (4-3. A-3.-`-`.`.. and UX should be reported. 08/17/2010 06:03:02 MDT
.````````.
All data should be inspected for spurious data points as a continuing check on the measurement process. this practice decreases the probability of rejecting bad data points. BX . Figure A-3.1 shows a spurious data point called an outlier. However. SX . The Thompson Technique is excellent for rejecting outliers. Errors of this type should not be included as part of the uncertainty of the measurement.```.`. In most of these tests. the probability for rejecting a good point is set at 5%. X. The suspected outliers should then be subjected to an engineering analysis.1 Outlier Outside the Range of Acceptable Data
78
--`. Two tests are in common usage for determining whether or not spurious data are outliers.

The suspected outliers should then be subjected to an engineering evaluation to determine the cause of the outliers. we conclude that −555 is a possible outlier.1. The Grubbs Method does not reject as many outliers but the number of good points rejected is smaller.125 Sample standard deviation (SX) p 140.
To illustrate the calculations for determining whether −555 is an outlier. This procedure should be repeated for all remaining data points. The Modified Thompson Technique uses the equation for S as defined in this Supplement. the diagram can be defined as a bar chart with the bars arranged in descending order of size. If ␦ is smaller than SX.
␦ p | −55 − 1.1-2005
good values.) The test for the outlier is to compare the difference (␦) with the product SX. etc. Then. If ␦ is larger than or equal to SX.8 Sample size (N) p 40 By the above equation for ␦. a value of is obtained for the sample size (N) at the 5% significance level. are attributed to Joseph Juran who defined the general principle known as the “Pareto Principle” — the “Vital Few. the absolute difference of Xj from the mean (X) is calculated as
␦ p | Xj − X |
probe readings were average ␦ and ␦ was calculated from the average for each probe. If there is not valid engineering justification to remove the suspected outliers and if removal of the outliers will significantly change the test result and its associated uncertainty.875
Using Table A-3. we say Xj is not an outlier.1 General It is often useful to display the relative sizes of the components of a whole with a bar chart. A-3. Trivial Many. the following steps are taken. the Modified Thompson Technique1 described below is recommended for identifying suspected outliers. (The probability of not rejecting a bad data point is not fixed.1. Most of the uses today. Suppose Xj. the jth observation. This limits the probability of rejecting a good point to 5%. The sample standard deviation (SX) and the mean (X) of the sample are calculated. an Italian economist who used this type of diagram in his studies of the unequal distribution of wealth. but for the purposes illustrated here.TEST UNCERTAINTY
ASME PTC 19. If there is valid engineering justification. data from similar tests. 08/17/2010 06:03:02 MDT
. the expected temporal and spatial variation/profiles of the parameters being measured.
26 −11 148 −107 126 79 −137 −52 20 −72 58 120 −216 9 179 24 124 12 −40 41 1 129 −56 40 127 −103 −38 89 2 −35 −121 25 8 10 334 −220 −60 −29 166 −555
334 and −555 are suspected outliers. The 40
Thompson used a different equation for S.2 Thompson Technique (Modified)1 Consider a sample (Xi) of N measurements.8 p 270. is the suspected outlier. It will vary as a function of sample size.3 Example There were 40 temperature probes installed in one stage of the turbine of a jet engine.125| p 556.125
From Table A-3.9
Since ␦ > SX.
79
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. we say Xj is an outlier. the suspected outliers may be removed from the analysis of the test result and its associated uncertainty. The removal of these outliers should be documented within the test report.
1
Since ␦ exceeds SX for the suspected point. The engineering evaluation should include an analysis of the instrumentation. A-4 PARETO DIAGRAMS A-4. the validity of the test should be questioned.” Mathematics were developed that described the distribution. In this Supplement. the physics of the measurement methods employed. we conclude that 334 also is a possible outlier. A-3.125 Έ p 332.
SX p 1.
␦ p Έ 334 − 1. One particular type of bar chart is called a Pareto diagram after Vilfredo Pareto. Mean (X) p 1. with extensive activity in the area of quality control. Repeating the above procedure for 334.924 ؋ 140. 334.

915 1.944 1.932 1.940 1.919 1.906 1.876 1.925 1.571 1.946 1.881 1.945 1.929 1.930 1.931 1.944 1.798 1.928 1.937 1.871 1.939 1.941 1.425 1.749 1.913 1.939 1.943 1.ASME PTC 19.656 1.944 1.942 1.943 1. 08/17/2010 06:03:02 MDT
.1 Modified Thompson (At the 5% Significance Level)
N
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
1.933 1.937 1.943 1..943 1.920 1.904 1.942 1.777 1.936 1.865 1.922 1.946 1.945 1.711 1.942 1.1-2005
TEST UNCERTAINTY
Table A-3.941 1.`.829 1.885 1..938 1.-`-`.938 1.899 1.858 1.````````.937 1.935 1.`.889 1.917 1.940 1.850 1.934 1.945 1.945 1.901 1.942 1.929 1.944 1.939 1.931 1.946 1.945 1.96
80
--`.840 1.815 1.935 1.`.923 1.910 1.916 1.893 1.941 1..911 1.934 1..`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.924 1.151 1.927 1..945 1.``.938 1.944 1.921 1.927 1.896 1.942 1.```.935 1..940 1.941 1.`.932
N
52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 ϱ
1.`.926 1.936 1.944 1.908 1.

derivation of the percentage contribution of an elemental source of standard uncertainty to the combined uncertainty is computed as the ratio of the square of the combined standard uncertainty that would be computed if the elemental source were the only source of standard uncertainty to the square of the combined standard uncertainty computed when accounting for all sources of uncertainty.``. (Note that these values were obtained from subsection 10-4).00367 −0.00203 0. the first step is to define the individual systematic and random standard uncertainties of the mean in terms of their relative individual percentage contributions to the combined standard uncertainty of a test result.00386) 2 p −6.00629 0.0409 0.`---
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. 08/17/2010 06:03:02 MDT
.0409 * 0.3%
The percentage contribution of each elemental source of systematic standard uncertainty (bXi ) to the combined standard uncertainty is
For bT1T2:
b T1 T2 % contribution ~ to uR p
2 T1 T2 bT1 T2
u2 R ؋ 100
͑ i b X i͒
u2 R
2
؋ 100
p
The percentage contribution of each correlated source of systematic standard uncertainty (bX1X2).0625 0.00386
For b P1:
b P1 % contribution ~ to uR p
p
͑P1bP1͒2
u2 R
؋ 100
(−0.792 0.600 −0.1-2005
To apply this to a test uncertainty example. and combined standard uncertainty are summarized in the following table.`.-`-`.3%
΅ ؋ 100
for sP1:
؋ 100
s P1 % contribution ~ to u R p
(−0. The percentage contribution of each elemental source of random standard uncertainty (sXi ) to the combined standard uncertainty (uR) is
of Pareto diagrams.. b T1 p 37. to the combined standard uncertainty is
2 1 2 b X1 X 2
u2 R
΄
2(0.0%
͑ i s X i͒ 2
u2 R
Similarly:
؋ 100
b P2 p 10.```.`.030 0.00386)2 p 10.TEST UNCERTAINTY
ASME PTC 19..````````.021)2 ؋ 100 (0.. As elemental sources of standard uncertainty are not combined as arithmetic sums but are instead combined as described in subsection 4-8..0409 * 0. The necessary values for parameter sensitivity.`.`. systematic standard uncertainties.4%.. Using the uncertainty model and associated assumptions presented in subsection 4-8.021 0.646 0. expressions for the percentage contributions from elemental sources of standard uncertainty were derived and are presented below.170 0. b T2 p 17..`. The second step is to create a bar chart which depicts the percentage contributions of individual systematic and random standard uncertainties to the combined standard uncertainty in descending order of size.2 Example The compressor performance example in subsection 10-4 will be used to illustrate the application
81
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`.1%
͑ P 1 s P 1͒ 2
u2 R
؋ 100
؋ 100
A-4.00367)(−0.030)2 p
(0.300 0.7%.198 0.00386)2
p 5.
Symbol
bP 1 bP 2 bT 1 bT 2 bT T 1 2 sP 1 sP 2 sT 1 sT 2 P1 P2 T1 T2 uR
Value
0.0625)
(0. random standard uncertainties.00203)(0. uR.

Although this application of Pareto diagrams has been used to determine the relative contributions to combined standard uncertainty of systematic and random standard uncertainties.2. 08/17/2010 06:03:02 MDT
. The ideal end result of an analysis such as this would be to take corrective action.1%. and bT1T2 should be combined as shown in Fig. s T1 p 8.2-2.2. uR. if possible.0%
Figure A-4. instrumentation. However. and bT1T2 is the largest contributor to the combined standard uncertainty. the combination of bT1. and bP1 the smallest. A-4.
As can be seen from this diagram.1-2005
TEST UNCERTAINTY
Similarly:
s P2 p 7. the method can be applied just as easily to the individual estimates of the elemental errors that contribute to systematic and random standard uncertainties.1 illustrates the relative contributions to combined standard uncertainty of individual-parameter systematic and random standard uncertainties in terms of a Pareto chart. A-4. bT2.ASME PTC 19.
Fig. the relative percentages for bT1. sT2 p 10.7%. or both.1 Pareto Chart for Random and Systematic Standard Uncertainties
82
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. since the systematic standard uncertainties for T1 and T2 are correlated. bT2. to reduce the contribution of major factors in the combined standard uncertainty through changes in methods.

``.2 Pareto Chart for Random and Some Combined Systematic Standard Uncertainties (Example 10-4)
83
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.TEST UNCERTAINTY
ASME PTC 19.`---
Fig.`..2.```.`.````````. A-4.. 08/17/2010 06:03:02 MDT
.`.`.-`-`.`....1-2005
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`..

. is determined from the estimates for the Ki elemental error sources for that parameter as
84
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. bik. For most engineering predictions of systematic uncertainty. The random standard uncertainty of the mean for each parameter is determined from Ni measurements as
si p 1
Ί Ni
΄
1 Ni − 1
͚ (Xi jp1
Ni
j
− Xi)
2
΅
1⁄
2
(B-1. There will usually be a set.3)
For the case of a single measurement (Ni p 1).2)
where
Xi p
͚ Xi jp1
Ni
Ni
j
(B-1. As discussed in para.0 for Gaussian. These estimates are combined through use of the ISO model to calculate the band about the experimental result where the true result is thought to lie with C% confidence. 08/17/2010 06:03:02 MDT
bi2 p
kp1
͚
Ki
b2 ik
(B-1. bik for the systematic errors. (B-1. 2. or the estimate of the error associated with repeated measurements of a particular parameter. and random standard uncertainties of the mean. the 95% limits of the possible distribution are estimated rather than the standard deviation of the distribution.1)
The estimate of the random error for a parameter is the random standard uncertainty of the mean.4)
The ISO Guide defines the combined standard uncertainty of the result as
2 uR p
͚ (i bi)2 + 2 i͚ ͚ i k bik ip1 p 1 k p i +1
+
ip1
I
I −1
I
(B-1. each elemental estimate. (7-4.73)(0. and. Consider an experimental result that is determined from I measured variables as
R p R ͑ X1.e. in most cases. 1. The uncertainty model requires estimates of the uncertainties for each of the elemental error sources for each parameter in the data reduction equation.ASME PTC 19. . The uncertainty model given in the ISO Guide to the Expression of Uncertainty in Measurement (1995)[1] is presented in this paragraph.1-2005
Nonmandatory Appendix B Uncertainty Analysis Models
B-1 ISO UNCERTAINTY ANALYSIS MODEL Nonmandatory Appendix B is adapted from “Computer-Assisted Uncertainty Analysis” [27]. The systematic standard uncertainty for parameter i. Systematic uncertainties can be estimated through manufacturer information. X 2. calibrations. these error distributions are assumed Gaussian (normally distributed) or rectangular (uniformly distributed).1) including the correlated
. Obtaining the standard uncertainty from the 95% estimate is simply a matter of dividing the estimate by the appropriate distribution factor. .5)
͚
I
( i si ) 2
where
i p
∂R
∂X i
(B-1. i. bi. X I͒
(B-1.. that remain after all calibration corrections are made.5) represent the systematic standard uncertainty of the result bR [eq. of elemental systematic standard uncertainties for each parameter.95) for rectangular]. The systematic standard uncertainties are related to those systematic errors. 4-3..6)
The first two terms on the right side of eq. The estimates of the elemental errors fall into two categories: systematic standard uncertainties. is taken to be the prediction of the standard deviation for a particular distribution of possible errors for that particular error source. si for the random errors where i represents each parameter.65 p (1. through sound engineering judgment. Typically. D [i. Ki.2. previous information must be used to calculate si [4].

bik.10) 85
+2
( i s i ) 2΅ ͚ ͚ i k bik + i͚ i p 1 k p i +1 p1
--`.321 14.995
127.104 3.250 3.041 4.776 2.`.`.029 3.740 1.756 2. In order to obtain the overall uncertainty in the result..`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.499 3.999
636.768 3.132 2.373 3.064 2.690 3.228 2.725 1.604 4. For instance.7)
where si is either
si p N i − 1
(B-1.030 2. at a specified confidence level. The degrees of freedom of the elemental systematic standard uncertainties bi may be known from previous information or k estimated.201 2. UR.711 1. The covariance of the random errors is assumed to be zero.922 3.551 3.925 5.729 1. 08/17/2010 06:03:02 MDT
.303 3. at a C% confidence level
UR.372 3.598 4.971 2.883 3. if one thought that the estimate of bik was reliable to within ±25%.714 1.571 2.TEST UNCERTAINTY
ASME PTC 19.453 5.617 2.135 3..703 1.045 2. R.069 2.845 2.924 8. The values for t are given in Table B-1.032 3.
bi p 1⁄ 2
k
b
ik
⌬ bik
−2
(B-1.11)
1⁄ 2
(B-1.576
0.080 2.038 3.353 2.610 6.792 3.408 5.089 7.174 3.120 2.833 1.717 1.684 1.706 1.306 2.671 1. or the correlated systematic standard uncertainty.746 1.598 12. the ISO Guide recommends that the combined standard uncertainty of the result be multiplied by a coverage factor.750 2. is determined by summing the products of the elemental systematic standard uncertainties for parameters i and k that arise from the same source and are therefore perfectly correlated [10] (see subsection 8-1).773 4.660 2.657 9.048 2.965 3.460 3.000 1.110 2.119 3.101 2.869 5.645
0.797 2.090 3.900
6.````````.708 1.197 3.106 3.074 2.179 2.182 2.980 1. (B-1.437 4.318 4.915 2.734 1. then
bi p 1⁄ 2 (0.943 1.819 3.960
0.753 1.706 4.659 3.861 2.-`-`.807
0.771 2.160 2.898 2. the Welch-Satterthwaite formula is adapted as:
Table B-1 Values for Two-Sided Confidence Interval Student’s t Distribution [9]
C
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 40 60 120 ϱ
0.947 2.`...5) to obtain the overall uncertainty in the result.021 2.015 1.878 2.056 2.707 3.057 3.`.646 3.``.819 2.701 1.314 2.619 31.365 2.1).977 2.223 3.252 3.807 2.581 3.060 2.990
63.262 2.841 4.782 1.131 2.140 4.831 2. (b) Given are the values of t for a confidence level C and number of degrees of freedom .721 1. The coverage factor is the value from the t distribution for the required confidence level corresponding to the effective degrees of freedom in the result.291
R p
ip1
Ά͚ ͚΄
I
I
ip1
͓͑i bi ͒ + ͑isi͒ ͔
2 2
4
K
·
2
͑ i s i͒
si
+
kp1
͚
i
͑ i bik͒ 4
bi
k
΅
(B-1.860 1.447 2.763 2. UR.093 2..771 1..674 3.326 3.497 3.697 1.221 4.895 1.012 2.796 1.042 2.812 1.9)
where the quantity in parentheses is an estimate of the relative variability of the estimate of bik.8)
or the degrees of freedom of the previous information if si is estimated [4].833 3.860 2.286 3.658 1.428 3.921 2.25) −2 p 8
k
With R known.```.745 3. the proper t value is obtained from Table B-1 for C% confidence and multiplied by uR from eq.067 3.725 3.920 2.1-2005
terms].317 4.587 4.`.761 1. (7-3.704 2.781 4.047 3.707 3.C p tC
I −1 I
΄ ͚ ( b )
ip1 i i I
I
2
(B-1.690 3.959 5. The covariance of the systematic errors.787 2.850 3.073 4. The ISO Guide recommends the approximation
GENERAL NOTES: (a) See [9].078 3. To find R.169 3.055 3. and the third term is the random standard uncertainty of the result sR eq.153 3.145 2.699 1.086 2.052 2.355 3.950
12.779 2.015 3.

The large sample uncertainty expression given in eq. t95 can be taken as 2 to a good approximation (93% to 95% coverage). for large degrees of freedom in the result
U R.1) including the correlated systematic standard uncertainty terms].ASME PTC 19. (7-4.1). (B-2.2)
86
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.95 p 2 ͑b 2 R + s R͒
1⁄
2
(B-2.1) is then obtained as (including the correlated systematic uncertainty terms)
΄ ͚ ( b )
ip1 i i
I
2
1⁄
(B-2.95 p 2 +2
The first two terms in the brackets in eq. Therefore. (7-5. (B-1.1-2005
TEST UNCERTAINTY
B-2 LARGE SAMPLE UNCERTAINTY ANALYSIS APPROXIMATION The method described in subsection B-1 is the strict ISO method. and the third term in the brackets is the random standard uncertainty of the result sR eq.1)
͚ ͚ i p 1 k p i +1
I −1
I
i k b ik +
͚ ip1
I
(i si )2
΅
2
2 U R. 08/17/2010 06:03:02 MDT
. (7-3. It has been shown [28-30] that for most engineering applications.1) are the systematic standard uncertainty of the result bR [eq.7) is 9 or greater. when the degrees of freedom for the result from eq.

e. . . xn ) is the joint probability density function for X1 . r (X1 .. . Xn ) of the random variables X1 .`. Xn ). . . . . . Xn ). Xn ) by means of a first order approximation to the Taylor series expansion for r (x1 . X2 presents the special case where f (X1 . . .. . variances. X1 .
E [( X − X ) 2] p
ϱ
-ϱ
( x − X ) 2 p( x ) dx p X2
The expected value of a function f (X1 . The covariance X1X2 of the random variables X1 . .. . . . . Y. Xn . The mean (or expected value) X of the random variable X presents the special case where f (X) p X. i. . x 2 . . Suppose
Y p a 0 + a1 X 1 + a2 X 2 + . X2 . . then the mean and variance of Y can be expressed in terms of the means. X2 . that are directly measured..1-2005
Nonmandatory Appendix C Propagation of Uncertainty Through Taylor Series
C-1 INTRODUCTION Experimental results are not always directly measured. . X1 .. . Xn )... Xn ) to deviations in (X1 . X2 ..`. .````````. .... xn ) dx1 dx 2 . . can be expressed as a linear combination of random variables. . X2 .. . x2 . . .. . can be expressed in terms of the variances and covariances of its arguments. Xn ). .-`-`. Xn is given by
E [f ( X 1 . . x n )
p (x 1 .
͵
ϱ
-ϱ
f ( x1 .. . The expected value of a function f (X) of a random variable X is given by
where p(x) is the probability density function for X. . X2 . . + a n X n
E [f ( X )] p
͵
ϱ
-ϱ
f ( x ) p( x ) dx
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. . X n )] p
͵͵
ϱ
-ϱ
ϱ
-ϱ
.e. . . Xn . . X2 . . . ... r (X1 . X2 .” and “covariance” will be used throughout this Appendix. . .```.. and covariances of the variables Xi . X2 ) p (X1 − X1 )(X2 − X2 ). . which are directly measured. xn ) will be assumed to be continuous with continuous partial derivatives in the neighborhood of the point (X1 ..`. 08/17/2010 06:03:02 MDT
. . X2 . Xn . i. . . the function r (x1 . . The definitions below for “mean. . . x 2 ) dx 1 dx 2 p X1X2
C-3 PRELIMINARY CONSIDERATIONS If any random variable... The aim of this Appendix is to provide a method by which the variance of an experimental result that is not directly measured. .. X2 . to be defined as a function of certain variables.
E [f ( X 1 . X2 ..``. Xi .
E [X ] p
C-2 DEFINITIONS The primary goal of Nonmandatory Appendix C is to present an expression for the variance in r in terms of the variances and covariances of X1 . X2 .`... x2 . The approach will be to relate the deviations in r (X1 . X 2 )] p
͵͵
ϱ
-ϱ
ϱ
-ϱ
( x1 − X1 )
( x2 − X2 ) p ( x1 . where Xi is the true value of the measured variable Xi . x2 . It is quite common for an experimental result.. X 2 . . . . . xn ) in the neighborhood of the point (X1 .`---
͵ ͵
ϱ
-ϱ
x p ( x) dx p X
The variance 2 X of the random variable X presents the special case where f (X) p (X − X )2. i.. .” “variance. dx n
where p(x1 . In order to facilitate this project..ASME PTC 19. Xn . ..`.. X2 . x2 .e.
87
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`. .

X2 . . . . Xn we get
r (x 1 . .1 The First Order Approximation Suppose r p r (x1 . . . + an ( Xn − Xn ) ]}
2
For example. + Xn ( x n − Xn ) + higher order terms
The variance of Y will be given as follows:
2 2 Y p E [( Y − Y ) ]
p E{[ a 1 ( X 1 − X1) + a 2 ( X2 − X2 ) + . .`. the variance of Y will be given by
2 Y p
a i aj XiXj ͚ ai2 Xi2 + 2 j p ͚ ip1 i +1
n n
Consequently
r ≈ r ( X1 . . . 08/17/2010 06:03:02 MDT
. and covariances of the variables X1 .. Xi . X2 . if n p 3.`. Xn ) + X1 (x 1 − X1 ) + X2 ( x2 − X2 ) + . . . Xn .```. . .`.1-2005
TEST UNCERTAINTY
Then
E [Y ] p E[ a 0 + a 1 X1 + a 2 X2 + . Xn . Xn we have
r (X 1 . . .-`-`. xn ). If the function r (X1 . . X2 . . . C-3 we will be able to express the mean and variance of r (X1 . then
88
For the special case where the random variables X1 . . xn ) p r ( X1 .. . X2 .. Xn .
C-4.. x2 . .. Xn ) can be expressed as a linear combination of the measured variables X1 . X2 .. . X2 .. . . .. Xn . . is defined as a function of certain measured variables..
--`. suppose that the arguments of the function r (X1 . .. + an X n ]
or. .. . . and covariances of the variables Xi . X2 . Xn ) are the random variables X1 . assume that the higher order terms in the Taylor series expansion for r are negligible compared to the first order terms. . . X2 . then
2 Y p E {[ a 1 ( X 1 − X1 )
+ a2 ( X 2 − X2 ) + a 3 (X 3 − X3 ) 2]}
where xi are the sensitivity coefficients given by
Xi p
∂r ∂x i
2 Y
p
E [ a2 1 + a2 3
( X1 − X1 ) + a 2 ( X 2 − X2 )
2 2
2
( X 3 − X3 )2
+ 2a 1 a 2 (X 1 − X1 )(X 2 − X2 ) + 2a 1 a 3 (X 1 − X1 )(X 3 − X3 ) + 2a 2 a 3 (X 2 − X2 )(X 3 − X3 )]
or
2 2 2 2 2 2 2 Y p a 1 X1 + a2 X2 + a 3 X3
Now. using the definition of the mean value of a random variable. X2 . .. Xn )
and
2 r ≈ Xi Xj XiXj ͚ Xi2 Xi2 + 2 j p ͚ ip1 i +1
n n
C-4 PROPAGATION OF UNCERTAINTY/ERROR THROUGH TAYLOR SERIES In subsection C-3 we found that if a random variable. . X2 . . variances. . .
Y p a 0 + a1 X1 + a 2 X2 + . . X2 . ..`.. Then in the neighborhood of X1 . . Furthermore.`. . Xn ) in terms of the means. If we expand r through a Taylor series in the neighborhood of X1 . Xn ) in the neighborhood of X1 . X2 .`---
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. Xn ) + X1 (X 1 − X1) + X2 ( X2 − X2 ) + ..````````. X2 . Now suppose an experimental result. Xn are all independent we get
2 r ≈
͚ Xi2 Xi2 ip1
n
Table C-4 presents some useful formulas for propagating variance through the first order approximation to the Taylor series for an experimental result r. . . Xn ) ≈ r ( X1 . . . .. Xn . r. + Xn ( X n − Xn )
+ 2a 1 a 2 X1X2 + 2a 1 a 3 X1X3 + 2 a 2 a3 X2X3
or more generally. . . . . .. X2 . . Y.. . . X2 . by means of a Taylor series approximation to r (X1 ... . X1 .. variances. . x2 . X2 . .. the mean and variance of Y can be expressed in terms of the means. .``.ASME PTC 19. . + a n Xn
using the results in para. . . . . ... could be expressed as a linear combination of the random variables.

y p B
(Ax + By )2 Ax By . 08/17/2010 06:03:02 MDT
. y′ p 1
2 V2 r p 4Vx
x p 2x
rpx
1⁄ 2
x′ p 2
V2 r p V2 x 4
S2 r p
S2 x 4x
1
2x
1⁄ 2
x p
r p ln x
x′ p 1⁄ 2
S2 r p
S2 x
V2 r p
x2 1 x p x
ln x
Vx
1 ln x
2
x′ p
r p kxayb
b a− 1 S2 Sx )2 + (bkxayb− 1Sy )2 r p (aky x
2 2 V2 r p (aVx ) + (bVy )
x p akybxa− 1.′p ∂x/x y ∂y/y
2 2 2 A2x2V2 x + B y Vy
2 2 2 2 S2 r p A Sx + B Sy
x p A.′p Ax + By y Ax + By
2 V2 r p Vy
x′ p
rp 1 y
Sr2 p S2 y
y4 1
y2
xSy
2
y p −
x x+y
ySx
2
y′ p −1
rp
S2 r p
΄(x + y) ΅ + ΄(x + y) ΅
2
2
V2 r p
2 y2(V2 x + Vy )
(x + y )2
y y .-`-`.TEST UNCERTAINTY
ASME PTC 19. p ∂x y ∂y
∂x/x V + ∂y/y V
∂r/r
x
2
∂r/r
2
y
x′ p
V2 r p
∂r/r ∂r/r . y′ p b
GENERAL NOTE: Vx p
Vy p
Vr p
--`.. y p bkxayb− 1
Sx x
Sy y
Sr r
x′ p a.```..`..``.. y p −
S2 x
(1 + x )4 1
(1 + x )
2
x
(x + y )2
x′ p
rp
x p
r p xy
(1 + x )2 1 x′ p 1+x
2 2 V2 r p Vx + Vy
V2 r p
2 2 S2 r p (ySx ) + (xSy )
x p y. y p x
rpx
2 2 2 S2 r p 4x Sx
x′ p 1.`.`.`.`.′p− x+y y x+y
V2 x
x p
x 1+x
y
(x + y )2
S2 r p
.y )
Variance (in Absolute Units) and Absolute Sensitivities
S2 r p
Variance (Dimensionless) and Relative Sensitivities
∂x S + ∂y S
∂r
x
2
∂r
2
y
V2 r p
x p
r p Ax + By
∂r ∂r ..`---
89
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.````````.1-2005
Table C-4 Taylor Series Variance Propagation Formulas
Function
r p f(x..

. x2 . Xj . .````````. In this paragraph we will briefly assess the conditions under which this approximation is meaningful.
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale. . .`. xn )
+
i p1
͚
n
xi (x i − xi )
+ (1⁄ 2!)
͚ ͚ ␥xjxk jp1 kp1
n
n
Assuming the systematic error to be independent of the random component of error this becomes
2 2 2 ␦ p  + ⑀
( xj − xj )( x k − xk ) + higher order terms
Now. Xn for an arbitrary joint probability density function p(x1 . x2 .. X2 .. Xi . Xn ) directly in terms of the variances and covariances for X1 . X2 . . C-4. . x2 . Xn ). C-5 PROPAGATION OF SYSTEMATIC AND RANDOM COMPONENTS OF UNCERTAINTY The total variance associated with a measured variable.ASME PTC 19. .`. the Taylor series expansion for r in the neighborhood of x1 .`---
so that
E [␦ ] p E[  + ⑀ ]
␦ p 
also
E [( ␦ − ␦) 2] p E {[(  −  ) + ⑀ 2]}
2 2 2 ␦ p  + ⑀ + 2 ⑀
r (x 1 .. xn becomes
--`. the total error in a measurement was given by
␦ p  + ⑀
where
␥ xx p
∂ 2r
∂x2
In this case the ratio of the second order term in the series to the first order term in the series is given by
Rp
␥ xx ( x − x ) x
So that for this case the assumption that R Ӷ1 reduces to the condition that ␥xx (x − x ) Ӷx . if the second order terms in the Taylor series expansion are retained. xn ) p r ( x1 . can be expressed as a combination of the variance associated with a fixed component and the variance associated with a random component of the total error in the measurement. xn ). . In subsection 4-2. .. X2 .. In this paragraph we will relate these two sources of variance in the measured variables. Then
r (x ) p r ( x ) + x ( x − x ) + (1/2!) ␥ xx ( x − x ) 2 + higher order terms
C-4.. x2 .2 Assessing the Validity of the First Order Approximation In para. More generally. .``. .-`-`. Let’s first consider the simplest case where r (x)... . then there is no way to express the variance in r (X1 . 08/17/2010 06:03:02 MDT
.`.1 we assumed that the Taylor series expansion for r could be reasonably approximated through the first order terms.`. . to the variance in an experimental result r (X1 . ..3 The Limitation of the Present Approach If the higher order terms in the Taylor series expansion are not small compared to the first order terms..`.1-2005
TEST UNCERTAINTY
C-4.```. . . if we define the random variable Xi as follows:
X i p ( ␦ i − i ) + Xi
where
␥ xjxk
∂ 2r p ∂ x j ∂x k
then
E [X i] p Xi E [( X i − Xi ) 2] p ␦i2 p i2 + ⑀i2 p Xi2 90
The second order terms in this expansion may be compared directly to the first order terms in order to assess their significance... . .

These results can be combined with those in para. (B-1.`.. C-3.3).1-2005
Assuming fixed errors to be independent of random errors.. the experimental result r will also be normally distributed with mean r and variance 2 r. 08/17/2010 06:03:02 MDT
.`.````````.`.``. with the use of sample statistics. this equation leads to eq.`. it can also be shown that
E [( X i − Xi )(X j − Xj )] p XiXj p ij + ⑀i⑀j
For no correlation among the random errors. and for a 95% confidence level.-`-`..`---
91
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.TEST UNCERTAINTY
ASME PTC 19..```..1 to yield
Xi Xj ⑀i⑀j ͚ Xi2 ⑀i2 + 2 j p ͚ ip1 i +1
+
n n
r2 ≈
Xi Xj ij ͚ Xi2 i2 + 2 j p ͚ ip1 i +1
n
n
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
--`. C-6 THE PROBABILITY DENSITY FUNCTION OF A RESULT The preceding paragraphs make no assumptions about the joint probability density function of the measured variables Xi .. Assuming that the first order approximation to the Taylor series expansion for r is adequate and that the measured variables Xi are jointly normally distributed.`.

. Then the distribution of Y will be approximately normal with expectation
N ip1
E(Y) p ∑ ci E(Xi)
ip1
N
and
variance
2(X) p ∑ ci2 2 (Xi).
92
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.ASME PTC 19.” Although not mathematically precise. . 08/17/2010 06:03:02 MDT
.2. this is a reasonable statement of the central limit theorem for application purposes.1-2005
Nonmandatory Appendix D The Central Limit Theorem
Paragraph G. It is a more general version of the central limit theorem than what one would find in elementary statistics textbooks. provided that Xi are independent and 2 (Y) is much larger than any single component ci2 2 (Xi) from a nonnormally distributed Xi. where E(Xi) is the expectation of Xi and 2 (Xi) is the variance of Xi.1 in [1] states “The Central Limit Theorem:
N ip1
Suppose Y p
c1 X1 + c2 X2 + . + cN XN p ∑ ci Xi.

ASME PTC 19..````````.`. 08/17/2010 06:03:02 MDT
--`.```.``.`...1-2005
D04505
Copyright ASME International Provided by IHS under license with ASME No reproduction or networking permitted without license from IHS
Licensee=Mott MacDonald Ltd/5956936002 Not for Resale.`---
.`.-`-`.`...`..