1、Designation: C 1067 00 (Reapproved 2007)Standard Practice forConducting A Ruggedness or Screening Program for TestMethods for Construction Materials1This standard is issued under the fixed designation C 1067; the number immediately following the designation indicates the year oforiginal adoption or,
2、 in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. Asuperscript epsilon (e) indicates an editorial change since the last revision or reapproval.1. Scope1.1 This practice covers a procedure for detecting sources ofvariation in a test me
3、thod.The procedure should be used duringthe development of a test method, before the interlaboratorystudy is executed, such as those in Practices C 670, C 802, andE 691. Interlaboratory studies can be expensive to execute.Resources will probably be more efficiently used if sources ofvariation in a t
4、est method are eliminated prior to performingthe interlaboratory study. The procedure also is useful fordetermining sources of variation in an existing test method thathas been found to have poor precision.1.2 This practice covers, in very general terms, techniquesfor planning, collecting data, and
5、analyzing results from a fewlaboratories. Annex A1 provides the details of the procedurewith an example and Annex A2 gives the theoretical back-ground.1.3 The practice does not give information pertinent toestimating within- or between-laboratory precision.1.4 This standard does not purport to addre
6、ss all of thesafety concerns, if any, associated with its use. It is theresponsibility of the user of this standard to establish appro-priate safety and health practices and determine the applica-bility of regulatory limitations prior to use.2. Referenced Documents2.1 ASTM Standards:2C 670 Practice
7、for Preparing Precision and Bias Statementsfor Test Methods for Construction MaterialsC 802 Practice for Conducting an Interlaboratory Test Pro-gram to Determine the Precision of Test Methods forConstruction MaterialsE 456 Terminology Relating to Quality and StatisticsE 691 Practice for Conducting a
8、n Interlaboratory Study toDetermine the Precision of a Test MethodE 1169 Guide for Conducting Ruggedness Tests3. Terminology3.1 Definitions:3.1.1 determination value, nnumerical quantity calculatedas directed in the test method using direct measurementsobtained in accordance with the procedures give
9、n in the testmethod.3.1.2 replication, nthe act of obtaining two or moredetermination values under specified conditions. The numberof replications must be finite and the scope of the replicationoperation may be narrow or broad, but must be specified.3.1.3 For definitions of other statistical terms u
10、sed in thisstandard, refer to Terminology E 456.3.2 Definitions of Terms Specific to This Standard:3.2.1 factor, nan element in the test procedure or labora-tory environment that is a potential source of variation in testresults.3.2.2 ruggedness, adjthe characteristic of a test methodthat produces t
11、est results that are not influenced by smalldifferences in the testing procedure or environment.3.2.3 screening, nthe detection of significant sources ofvariation as compared to chance variation.3.2.4 variable, na number or quantity that varies.4. Summary of Practice4.1 The practice requires that th
12、e user develop, from theo-retical or practical knowledge, or both, a list of factors thatplausibly would cause significant variation in test results if thefactors were not controlled. The technique is limited to theanalysis of the effects seven factors and requires considerablyless effort than would
13、 be required to collected data for seven1This practice is under the jurisdiction of ASTM Committee C09 on Concreteand Concrete Aggregates . This practice was developed jointly by ASTM Commit-tees C01, C09, D04, and D18, and is endorsed by all four committees.Current edition approved June 1, 2007. Pu
14、blished October 2007. Originallyapproved in 1987. Last previous edition approved in 2000 as C 1067 00.2For referenced ASTM standards, visit the ASTM website, www.astm.org, orcontact ASTM Customer Service at serviceastm.org. For Annual Book of ASTMStandards volume information, refer to the standards
15、Document Summary page onthe ASTM website.1Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.factors in a full factorial study. Procedures exist for analysis ofsmaller and larger numbers of factors (see Guide E 1169), butseven is a conve
16、nient number for many test methods forconstruction materials. The seven-factor analysis requires 16determinations by each laboratory. The procedure can beusefully executed by a single laboratory, but sometimes addi-tional information can be obtained if it is repeated in one or twoadditional laborato
17、ries.4.2 The procedure requires that two levels of each factor beidentified, then 16 determinations be done on a prescribedcombinations of factor levels. The levels assigned to a factormay be quantitative or qualitative (for example, brass versussteel).4.3 The disadvantage of this type of analysis i
18、s that themethod only estimates simple effects of each factor and doesnot detect interactive effects among factors.5. Significance and Use5.1 The purpose of a ruggedness evaluation is to determin-ing how sensitive the test method is to changes in levels ofpertinent operating factors. Normally, opera
19、ting conditions fora test method are defined along with an allowable tolerance. Aruggedness analysis determines that effect of worst-case varia-tion in operating conditions within this tolerance range. Themethod then can be revised with smaller tolerances on operat-ing conditions to improve the prec
20、ision.5.2 A major reason for poor precision in test methods is thelack of adequate control over the sources of variation in testingprocedures or testing environments. These sources of variationoften are not controlled adequately because they were notidentified during the development of the test proc
21、edures.5.3 All new test methods must be subjected to an interlabo-ratory program for purposes of developing a precision and biasstatement. These programs can be expensive and lengthy, andthe result may be that the determination is made that themethod is too variable to be published without further r
22、evision.Interlaboratory studies typically give the subcommittee anindication that the method is too variable, but they do notusually give a clear picture of what is causing the variation.Application of this ruggedness practice using one or a fewlaboratories may be a much more economical way to deter
23、minethese causes.5.4 Many existing test methods were published before therewas a requirement that precision and bias statements bedeveloped. Since this became a requirement, most of these testmethods have developed precision and bias statements, and theresult is that many have been found to suffer f
24、rom relativelylarge amount of variation. Use of this practice represents arelatively simple way to investigate the causes of variation intest methods, so that a subcommittee will have some guidanceas to which parts of the test method need to be studied furtherfor revision.5.5 The procedure can be us
25、ed for a program within a singlelaboratory, but involvement of at least three laboratories isrecommended, particularly if the single laboratory were to bethe one in which the test method was developed. This isparticularly important for new test methods. The originatinglaboratory is so much a part of
26、 the development of the testmethod that it is difficult for it to be objective in spotting anyproblems in the clarity of the test method directions. Twoadditional laboratories will probably contribute fresh criticalreview of the validity of the test method and provide assistancein clarifying the ins
27、tructions of the test method when needed.6. Materials6.1 The number and types of material shall cover the rangeof material properties to which the test method is applicable.The test method does not apply to material types or propertyvalues outside the range evaluated. Three to five materials willusu
28、ally be sufficient.6.1.1 Some preliminary testing may help the laboratoriesinvolved determine the materials that shall be used in thescreening program.7. Procedure7.1 Determine the number of laboratories that will partici-pate in the program and which materials each will use in theprogram. The maxim
29、um amount of information is obtained ifall laboratories include all materials in their part of theprogram, however cost can be reduced by each laboratoryusing a different material. Caution must be exercised ininterpreting the results since laboratory-dependent cannot beseparated from material-depend
30、ent effects.7.2 Factors that are likely to have the greatest effect on thevariability in the test results are selected for study. Levels ofthese factors are determined, selecting the minimum andmaximum levels that would plausibly occur in the execution ofthe test method if there were no particular e
31、fforts to controlthem. Only two levels are allowed. Levels often representquantitative properties, such as temperature, pressure, etc, butthey may also represent nonquantitative values, such as old vsnew, wet vs dry, etc. In this standard, factors are assigned letterdesignations, AG, and the two lev
32、els of each factor aredesignated with upper and lower cases of these letters, as inTable 1.7.3 Assign combinations of factor levels to experimentaldeterminations according to Table 1. The 8 determinations willbe done in duplicate, therefore, the full study on each materialwill require 16 determinati
33、ons.7.4 Construct a 16 row by 16 column results matrix from the16 determinations values (d1 d16) as shown in Table 2. Theabsolute values of the determinations in each row are identical,only the signs vary. Calculate Z and W statistics as shown in theequations below.Zr5(116di, where di8s are the 16 r
34、esults in each row r!. (1)TABLE 1 Pattern of Assigning Levels to Seven FactorsDetermination NumberFactor 1 2 3 4 5 6 7 8A aaaaAAAAB bbBBbbBBC CcCcCcCcDDDddddDDEeEeEEeEeF FffFFffFGGggGgGGgC 1067 00 (2007)2Wr5Zr216(2)7.5 The W statistic for row 1 represents the simple sum ofthe determinations and are
35、not used in this analysis. Statisticsfor rows 28 (W2 W8) represent the effects of the sevenfactors. The statistic for row 9 (W9) represent the total variationbetween the two replicate sets and is not used in this analysis.Statistics for rows 10 through 16 (W10 W16) are used tocalculate the error var
36、iance (X), which then is used to calculatethe test criterion (F) for each factor, as shown by the equationsbelow. Calculations are summarized in Table 3.X 5 (r51016Wr2!/7 (3)Ff5Wr2X, where Ffis the F statistic for the effect of factorf (17, represented by W2 W8, respectively)7.6 A F value of $5.59 r
37、epresents a significant effect forfactor f at a probability of 5 % for drawing an erroneousconclusion.7.7 An example of an analysis of data representing resultson 4 materials from 3 laboratories is shown in Annex A1.8. Keywords8.1 precision; ruggedness; test method; variationANNEXES(Mandatory Inform
38、ation)A1. EXAMPLE OF A RUGGEDNESS PROGRAMA1.1 This annex describes the procedure for conducting aruggedness evaluation using as an example a description of theruggedness evaluation on a test method for the measurement ofthe viscosity of asphalt.A1.2 As the first step in the ruggedness evaluation, ea
39、ch ofthe laboratories critically examined the procedure in theproposed test method. The objectives of the examination wereas follows:A1.2.1 To determine if the instructions are clear, concise,and complete,A1.2.2 To decide which factors are likely to influence testresults and therefore should be incl
40、uded in the study,A1.2.3 To pick materials that cover the range of the propertyof interest for the range of physical forms of the materials to betested, andA1.2.4 To determine the proper levels to be evaluated foreach of the chosen variables.A1.3 In this example, representatives of the three labora-
41、tories, after familiarizing themselves with the test method asspecified in A1.2, met and tried to improve the instructions forthe viscosity method. They selected variables, materials, andlevels that showed the effect of the variation. One of thelaboratories measured viscosity at 24 C, 25 C, and 26 C
42、 andfound that there was about a 10 % variation with a change ofTABLE 2 Results Matrix of 16 Determinations (d1 d16)Eight Determinations for Replicate Set 1 Eight Determinations for Replicate Set 2row1234567812345678ZW1 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z1W12 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d1
43、5d16Z2W23 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z3W34 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z4W45 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z5W56 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z6W67 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z7W78 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z8W89 d1d2d3d4d5d6d7d8d9d1
44、0d11d12d13d14d15d16Z9W910 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z10W1011 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z11W1112 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z12W1213 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z13W1314 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z14W1415 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d
45、15d16Z15W1516 d1d2d3d4d5d6d7d8d9d10d11d12d13d14d15d16Z16W16TABLE 3 Summary of Statistics for Seven Factors and RandomErrorFactor WFA W2W22/XB W3W32/XC W4W42/XE W5W52/XF W6W62/XG W7W72/XH W8W82/XW10X = (W2)/7, for W10-16W11W12W13W14W15W16C 1067 00 (2007)31 C. This was considered too large so 24.6 and
46、 25.4 C wereselected as the lower and upper temperature levels for theruggedness test. In the same manner, the effect of the othervariables were evaluated and the two levels to be evaluatedwere determined.NOTE A1.1Seven variables were selected and placed in a systematicprocedure called an incomplete
47、 Latin Square or a Youden Square (1).Thevariables are listed below and shown in a Youden Square in Table A1.1.This plan can evaluate the seven variables with eight determinations.Table A1.2 shows the variables and the levels selected for this example.A1.4 Four materials were selected to cover the ra
48、nge of thetest method and the viscosities were determined by each of thethree laboratories with one replication. The results are dis-played in Table A1.3. This plan required 16 determinations byeach laboratory on each material or 64 determinations by eachlaboratory.A1.5 Table A1.4 specifies the expe
49、rimental plan for aYouden Square for seven factors. The theory of its use iscovered in Annex A2. Table A1.4 consists of 16 rows and 16columns of coefficients each equal to 61 and arranged in adefinite pattern.A1.6 To obtain Table A1.5, first copy one row from TableA1.3 16 times in the general format of Table A1.5 and thenmultiply each entry in the new table by the corresponding entryin Table A1.4. Table A1.5 is just such a table derived from
