Reliability Handbook (Army Armament Research and Development Center)

摘要

This handbook was prepared for those engineers and scientists conducting reliability experiments who would like to use statistical techniques to improve the efficiency of their experiments. The following are set forth: a) The concept of reliability of functioning characteristics of weapon components in terms of stress and strength; b) The operating engineering definition of reliability; c)The complex nature of reliability; d) Ultimate reliability in terms of safety margins; e) The relationship between test and use conditions; and f) The limitations of reliability determinations imposed by testing facilities, information and cost. A two-phase testing procedure which meets the need for demonstrating high reliabilities with small sample sizes is described in a rational, objective manner. The first phase involves use of fractional-factorial experimental designs to survey effects of important environments. The second phase is a test-to-failure procedure (using the environment found most severe in the first phase) so conducted that reliability-in-use can be calculated from test results. The need to plan experiments in advance of data collection and to test to failure is emphasized. The requirements of a good experiment are treated. Several useful fractional-factorial test plans are completely laid out in the form of treatment procedures. Tests of increased severity most useful in testing to failure are described. Examples are given for applying these methods. Statistical tables, a glossary of terms, and a list of references are included.