Analysis of Unified Testing Theory.
Abstract
Components of a munition are supplied to the arsenal by a producer. The arsenal requires that these components be reliable. During the production phase of a lot of components, testing is required to be performed by the producer to assure this reliability. The producer is required to use statistical procedures which are thoroughly detailed in appropriate military standards. Military standards 105D and 1235A together with the military specification guide for each component define the sampling procedures and criterion to be used in judging whether a lot is acceptable. The objective of this report is to analyze various testing methodologies to determine if there exists a basis for modifying existing test procedures in order to minimize the costs involved in sampling plus the costs in making wrong decisions in accepting or rejecting lots. To accomplish this objective, P.A. personnel selected the 105 mm artillery for the study. Because of the study was a feasibility study, and because the funding and term of the contract was limited, the numerical results which are tabulated in this report are only based on the 70 of the 840 lots which were available from the computer. Aside from the actual numbers in the tables, the methodology which is used is statistically sound and give a savings in testing costs for this particular example. In general, an application of the design which is recommended will approximately minimize total costs when the costs of decision errors are included with shipping costs.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 26, 1976
- Accession Number
- ADA023872