1、Designation: C 1297 03Standard Guide forQualification of Laboratory Analysts for the Analysis ofNuclear Fuel Cycle Materials1This standard is issued under the fixed designation C 1297; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision
2、, 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 guide covers the qualification of analysts to per-form chemical analysis or physical measurements
3、of nuclearfuel cycle materials. The guidance is general in that it isapplicable to all analytical methods, but must be appliedmethod by method. Also, the guidance is general in that it maybe applied to initial qualification or requalification.1.2 The guidance is provided in the following sections:Se
4、ctionQualification Considerations 4Demonstration Process 5Statistical Tests 61.3 This standard does not apply to maintaining qualifica-tion during routine use of a method. Maintaining qualificationis included in Guide C 1210.1.4 This standard does not purport to address all of thesafety concerns, if
5、 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:C 1009 Guide for Establishing a Quality Assuranc
6、e Pro-gram for Analytical Chemistry Laboratories Within theNuclear Industry2C 1068 Guide for Qualification of Measurement Methodsby a Laboratory Within the Nuclear Industry2C 1128 Guide for Preparation of Working Reference Mate-rials for Use in the Analysis of Nuclear Fuel CycleMaterials2C 1156 Guid
7、e for Establishing Calibration for a Measure-ment Method Used to Analyze Nuclear Fuel Cycle Mate-rials2C 1210 Guide for Establishing a Measurement SystemQuality Control Program for Analytical Chemistry Labo-ratories Within the Nuclear Industry2C 1215 Guide for Preparing and Interpreting Precision an
8、dBias Statements in Test Method Standards Used in theNuclear Industry22.2 ISO Standard:ISO Guide 30 Terms and Definitions Used in Connectionwith Reference Materials33. Significance and Use3.1 This is one of a series of guides designed to provideguidance for implementing activities that meet the requ
9、ire-ments of a sound laboratory quality assurance program. Thefirst of these, Guide C 1009, is an umbrella guide that providesgeneral criteria for ensuring the quality of analytical laboratorydata. Other guides provide expanded criteria in various areasaffecting quality, producing a comprehensive se
10、t of criteria forcontrolling data quality. The approach to ensuring the qualityof analytical measurements described in these guides is de-picted in Fig. 1.3.2 The training and qualification of analysts is one of theelements of laboratory quality assurance presented in GuideC 1009, which provides som
11、e general criteria regarding quali-fication. This guide expands on those criteria to provide morecomprehensive guidance for qualifying analysts. As indicatedin Guide C 1009, the qualification process can vary in ap-proach; this guide provides one such approach.3.3 This guide describes an approach to
12、 analyst qualificationthat is designed to be used in conjunction with a rigorousprogram for the qualification and control of the analyticalmeasurement system. This requires an existing data base whichdefines the characteristics (precision and bias) of the system inroutine use. The initial developmen
13、t of this data base isdescribed in Guide C 1068. The process described here isintended only to qualify analysts when such a data base existsand the method is in control.3.4 The qualification activities described in this guide as-sume that the analyst is already proficient in general laboratoryoperat
14、ions. The training or other activities that developed thisproficiency are not covered in this guide.3.5 This guide describes a basic approach and principles forthe qualification of laboratory analysts. Users are cautioned to1This guide is under the jurisdiction of ASTM Committee C26 on Nuclear FuelC
15、ycle and is the direct responsibility of Subcommittee C26.08 on Quality AssuranceStatistical Applications, and Reference Materials.Current edition approved July 10, 2003. Published August 2003. Origianllyapproved in 1995 as C 129795. Last previous edition approved in 1995 as C 1297.2Annual Book of A
16、STM Standards, Vol 12.01.3Available form American National Standards Institute, 11 West 42nd Street,13th Floor, New York, NY 10036.1Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959, United States.ensure that the qualification program implemented meet
17、s theneeds and requirements of their laboratory.4. Qualification Considerations4.1 When a qualification program is being established,consideration should be given to analyst selection criteria, thetraining program, and practical demonstration. The criteria thatgovern when qualification is achieved s
18、hould be documentedalong with methods for determining the knowledge and skill ofthe analyst.4.1.1 Analyst selection should be based on establishedcriteria that are related to the complexity of the method thatanalysts are expected to perform. Criteria should include theminimum education required, any
19、 prerequisite training, and theoverall experience required. The selection criteria should bedefined and documented.4.1.2 The method-specific analyst training program shouldbe an established program with a prescribed training proce-dure. Some mechanism such as an oral or written test should beused to
20、 allow an analyst to demonstrate knowledge andunderstanding of the chemical, physical, instrumental, andmathematical concepts used to execute the method. It isadvisable to monitor progress during training to ensure that theanalyst has a reasonable chance of passing the qualificationtest.4.1.3 The pr
21、actical demonstration of the analysts ability togenerate results with the analytical method should be comparedto established criteria. The comparison criteria should bedefined and documented.NOTE 1The qualification of analysts, like many other laboratoryprocesses, has the potential for undetected er
22、rors. There are two types oferrors that occur. One is to fail an individual who should have beendetermined to be qualified. The other error is to pass an individual whoshould not have been determined to be qualified. The potential for theseerrors to occur and the potential consequences to the labora
23、tory should becarefully considered when determining the laboratorys qualificationmethodology. A statistical approach includes choosing the significancelevel at which the determination of qualification will be made. Thisproduces a quantitative value of the two possible risks. This is describedfurther
24、 in Appendix X1.5. Demonstration Process5.1 The suggested approach to practical demonstration foranalyst qualification that is described in the remainder of thisguide involves a comparison of the performance of the analystwith the performance of all qualified analysts on a particularanalytical metho
25、d. The performance is measured by the analy-sis of reference materials (see ISO Guide 30) and comparisonof the results to the data base for the analytical method. Thisapproach requires a data base that describes method perfor-mance. The comparison described in this guide is statistical innature and
26、therefore statisticians should be involved early on inthe process of defining qualification. Other types of compari-sons may serve to qualify equally well; however, such com-parisons are not addressed in this guide. If used, they should bedefined and documented.5.2 The data base for a given analytic
27、al method is generatedby all qualified analysts who run reference material samples onan established schedule or frequency. The data base is used toestablish the bias and precision of the method as routinely usedin the laboratory. The data base is established through ameasurement control program as p
28、resented in Guide C 1210.For a new method, a data base should be established accordingto Guide C 1068 and the analyst should be qualified againstthat data base.5.3 If changes in a method occur or changes in the executionof a method occur that render the existing data base represen-tation of the meth
29、od questionable, the qualification of analystsshould be suspended until the data base is verified or a newdata base is generated. When a new data base is generated, theold data base should be archived (retained for future reference)as a part of the documentation of the laboratory qualityassurance pr
30、ogram.5.4 A predetermined number of reference material samplesshould be selected for the analyst after training has beencompleted. The analyst should analyze the samples overseveral days, and not in a single session, to simulate morerealistically the conditions under which the data base wasestablish
31、ed.5.5 Since the samples may be at different concentrationlevels, the analysts demonstration results are normalized usingestablished parameters from the existing data base for eachcontrol standard. The normalized data are used to test forconformity to the data base. Statistical tests for the statist
32、icaldistribution (normality) as well as precision and bias aresuggested in Section 6. These terms are described in GuideC 1215.5.6 If the results of all three tests are satisfactory, the analystis qualified on that method. If the analyst does not qualify,retraining should be required before being al
33、lowed to retest forqualification. The analyst should be given a different set ofreference material samples each time retesting is allowed tomaintain the independence of successive tries. That will allowFIG. 1 Quality Assurance of Analytical Laboratory DataC1297032the same statistical tests to be use
34、d on each set of results. SeeFig. 2 for a schematic of the qualification process.6. Statistical Tests6.1 There are a number of statistical procedures appropriatefor performing the statistical tests on the analysts demonstra-tion data set to determine qualification. The procedures de-tailed in Append
35、ix X2 are suggested since they have proven tobe useful. Further information about these procedures isprovided by Snedecor and Cochran4and by NUREG/CR-4604.56.2 The analystss data set is first tested for statisticalnormality. If normality is rejected, the data set is rejected andthe analyst is determ
36、ined to have failed the qualification test. Ifthe data set is accepted as normally distributed, bias andprecision tests may be performed.6.3 If these statistical tests indicate that the analysts data setexhibits bias and precision estimates that are within those ofthe established data base, the anal
37、yst is determined to bequalified. If the precision and/or bias estimates are not accept-able, the data set is rejected and the analysts is determined tohave failed the qualification test.6.4 Examples of statistical tests are presented in AppendixX2.7. Keywords7.1 analyst qualification; measurement(s
38、); quality assur-ance; reference materials4Snedecor, G.W., and Cochran, W.G., Statistical Methods, 8th Ed., Iowa StateUniversity Press, Ames, Iowa, 1989.5NUREG/CR-4604, Statistical Methods for Nuclear Material Management,U.S. Nuclear Regulatory Commission, Washington, DC, 1988.FIG. 2 Steps in the An
39、alyst Qualification ProcessC1297033APPENDIXES(Nonmandatory Information)X1. STATISTICAL CONSIDERATIONSX1.1 The significance level, a, for a statistical test is setdepending on the desired risk of rejecting a qualified analyst.The smaller the significance level, the smaller the chance thata qualified
40、analyst will be rejected (Type I error). For example,if the significance level is 0.10, then there is a one in ten chancethat a qualified analyst will fail the test. However, by using asmall a, the chance of accepting an unqualified analyst is large(Type II error). Thus there is a trade-off between
41、accepting anunqualified analyst and rejecting a qualified one. Both types oferrors can be controlled at desirable low levels by requiring asufficiently large number of demonstration tests.4,5Practicallimitations usually restrict the available number of demonstra-tion tests so that only the risk of r
42、ejecting a qualified analystmay be adequately controlled by an appropriately small levelof significance.X1.2 For multiple statistical tests, another factor thatshould be considered when selecting the significance level ofeach test is the overall significance level. For example, theoverall significan
43、ce level for three independent tests would bea8 =1(1a)4. Therefore, if the significance level of eachtest was 0.05, the overall significance level would be 0.143. Inother words, the chance of a qualified analyst failing any one ormore of three independent statistical tests when each test has asignif
44、icance level of 0.05 would be 14.3 %.X2. SUGGESTED STATISTICAL TESTSX2.1 TEST 1Test for Normality:X2.1.1 Problem StatementTest whether the demonstra-tion data set is normally distributed.NOTE X2.1This test assumes that the data base itself is normallydistributed.Let,Yi5xi2 isi(X2.1)Y5(i 5 1nYin(X2.2
45、)s25(i 5 1nYi2 Y!2n 2 1(X2.3)where:xi= theith demonstration result,i= the known mean associated with theith referencematerial sample in the data base, andsi= the known standard deviation associated with theithreference material sample in the data base, and n is thenumber of demonstration results.X2.
46、1.2 Test statistic:W 5b2n 2 1!s2(X2.4)where:b 5(i 5 1kaiYn21112 Yi! (X2.5)Yiare sorted in ascending order,k = n/2, rounded down, andaiare the Shapiro-Wilks coefficients.4 ,5X2.1.3 Acceptance RegionUse Shapiro-Wilks tables todetermine the acceptance region for a desired level of signifi-cance.4,5X2.2
47、 TEST 2Testing the Variance (Precision):X2.2.1 Problem StatementTest whether the standardizeddemonstration results have a variance different from thevariance of a standard normal distribution.Ho: s25 1 (X2.6)Ha: s21 (X2.6)X2.2.2 Test Statistic:X25n21!s2s2(X2.7)where:s2=1.X2.2.3 Acceptance RegionUse
48、chi-square tables to deter-mine the acceptance region for a desired level of significanceand n1 degrees of freedom.4,5X2.3 TEST 3Testing the Mean (Bias):X2.3.1 Problem StatementTest whether the standardizeddemonstration results have a mean different from the mean ofthe standard normal distribution.H
49、o: 5 0 Ha: 0 (X2.8)X2.3.2 Test Statistic:C1297034Z 5Y2s/=n(X2.9)where: = 0 and s =1.X2.3.3 Acceptance RegionUse standard normal tables todetermine the acceptance region for a desired level of signifi-cance.4,5X2.3.4 The following examples provide data and test resultsfor actual qualification at a particular laboratory.X2.4 Example 1:Analyst Testing FormMethod: 67015 Log Number: 050416 Analyst: RRRDemonstrationResultKnownMeanAKnown StandardDeviationAStandardizedResult0.62616 0.62620 0.01689 0.0026.04147 6.14100 0.08341 1.1931.74910 1.80680 0.02023 2
