Downloaded from https://www.everyspec.com AMCP 706N 1 "Trhe information required fur the use of tions. The components must perform their this statistical model is listed hero: assigned functions when they are needed. S The probability of completing these fune- 1. Detailed analysis of the mission profile tions satisfactorily in the operating environ- to determine stress levels and durations for meat is called s-reliability. The total time each hardware level and to convert actual that a specific component must operate may test time to mission equivalents, be less than the total nidsson time, Reli- ability Is, therefore, a function of the environ- 2.
The subsystem (or system) reliability mental and usage stresses, the operating and equation, nonoperating conditions, the part reference- fallur• rates, and the time duration of the 3. Test data, including total test times and environments and usage stresses. The corn- associated environmental stresses and failures ponent reliability model Is based on the from all test sources. following assumptions (modified from Ref. 33): The model permits reliability assessment to begin on a piecemeal basis with informa- I, Failure rate is independent of time.
tion derived from development tests, engi- neering evaluation tests, and from qualifica- 2, Part failures are s-independent of each tion tests, Later on, information from other.,j production tests also can be used, 3. Part failure rate is independent of the The tests may differ with respect to levels, history of the component. conditions, and durations; they may be applied at component, equipment, and sub- 4. A specific assumption must be made system levels, ajid the test conditions include about the failure rate for all phases of the a wide variety of environmental conditions; mission, even benign subphases, and their durations may differ considerably. 5.
There is a great deal of uncertainty in The calculation of the best estimate of the estimation process, much of which is component failure rate in a specific environ- related to tMe suitability of the model, and ment will be discussed first. Then techniques the remainder is due to statistical uncertainty for incorporating data from higher level teats (calculated presuming the model is perfect). (equipment and subsystem) in the calculation of component failure rates will be presented. 6.
The k-factor approach is useful for Once the basic failure rate is established, converting failure rates from one set of con- methods are developed for combining these ditions to another, especially for storage-like estimates to generate the best estimate of conditions. mission reliability. A procedure for estimat- Ing the uncertainty of these estimates also is 7. Each special environment can be repre- described. sented by an additive term onto the reference failure rate, The size of the term is s-indepen- 2.12.2 COMPONENT MODEL dent of the presenrrm of other terms.
It is u if each special environment aggravated exactly Consider components in a complex system one s-independent failure mode in the item. which must operate successfully over a de- fined mission (Ref. 31). During the mission, Due to Assumption No, 5 there is little the component is exposed to eavironmental point in making minor statistical refinements stresses Jf different degrees, kinds, and dura- in the estimate. For example, s-bias in the