Downloaded from https://www.everyspec.com AMCP 704.0N 10t hr will not be the actual P,0 life, but is Accelerated tests can be implemented by presumed to be close to it. speeding up the duty cycle or the environ- mental level or both. The environments must be cycled from extreme to extreme in order to reproduce in a short time the de- gradation expected over the period of 7,7 APPLICATIONS actual service life. The environmental factor selected for acceleration is determined by Accelerated life tests can be used in pro- the item tested and its failure modes. For grams operating under tight and critical example, for many mechanical components, schedules.
11f production begi~ns before re- failure is caused by mechanical wear; hence, search and development is completed, some assurance must be obtained quickly that the the acceleration us obtained by increasing the frequency arid severity of stress.V equipment has an nl-,quate lifetime and that no gross design weaknesses exist. Life notgosts design weak nesto exit. Lifsed uThe failure data at usual environmental thests oftnditoon llevels and those at accelerated environments thew~ Qonditions, must correlate in some wiy with the stresses actually applied.
A precise statistical cor- Accelerated testing also is used when re- pair parts must be manutactured simultene- relation frequently cannot be obtained be- ouspa wir p ars msht beanpoufcturedionprogr, cause much of the theory of accelerated ously with a short run production program. testing is still very crude. In such cases, In this case, failure rate data cannot be pro- accelerated environmental tests may permit vided quickly enough by life testing to in- a great deal of intuitive information to be. fluenco the unalysis of the repair comple- developed. ment.
Statletical correlation often can be ob- tained with accelerated duty cycle testing, In somhe cases (explosives, for example), The expected number of cycles of actual the earliest times to failure in tht; storage service in a given time period often can be environment and variations in times to estimated. Accelerated testing is performed earliest failure must be known with high by increasing the number of cycles in a accuracy. A large sample would have to be given time period and measuring the mean- kept in storage in a usual life test. This cycles-to-failure. The mean cycles between would be very expensive and time consum- failures (MCBF) at th#a accelerated duty ing.
Accelerated testing of critical failure cycle frequently can be related to the MCBF modes can be used to determine the range at normal duty cycles as a function of the of variability of the time to failure, with ratio of cycles per time period, useful accuracy, with a smaller sample than that required for usual life testing. For 7-8 PARAMETRIC MATHEMATICAL example, solid rocket propellants are sub- MODELS ject to catastrophic failure moues that may result in explosions. The remaining life in The ideal accelerated test should include a slored lot must be estimat-,d periodically, (Ref. 1):
generally every six months or a year, The 1 aged samples must be subjected to acceler- I, An algorithm for converting the re- ated aging in order to determine whether liability data observed at accelerated con- some critical failure mode is about to be ditions to reliability data at normal con- triggered, ditions 79 w,