Accession Number:

AD0814877

Title:

RELIABILITY PREDICTION ON THE BASIS OF RESPONSE 'PEAK' STATISTICS FOR NON-GAUSSIAN PROCESSES.

Descriptive Note:

Summary technical rept. 1 Dec 65-30 Nov 66,

Corporate Author:

MINNESOTA UNIV MINNEAPOLIS DEPT OF ELECTRICAL ENGINEERING

Personal Author(s):

• Smits, T. I.
• Lambert, R. F.

Report Date:

1967-02-01

Pagination or Media Count:

39.0

Abstract:

Reliability prediction for dynamical systems threatened by two failure mechanisms, overload and wearout, is systematized in terms of the concept of a wear-dependent failure rate. Using orthodox failure theories, these failure rates are expressed in terms of the response peak statistics crests level crossings, maxima, or rises stress reversals. These failure rates are seen to be bounded above, and can be redefined for wearout having no definite critical level. Averaging the wear-dependent failure rates over the wear distribution in an operational system yields the system failure rate. Under certain hypotheses, including the modelling of the wear accumulation as a renewal process, the truncated normal wear distribution is found appropriate. Empirical response peak data for a bimodal system driven by normal-strength Poisson impulsive noise is fitted to a Weibull distribution. Certain Weibull parameter choices are used to digitally computer system wearout reliability and compare it with the normal-derived result of Freudenthal, indicating overoptimism of the latter. The drastic effect of accelerated testing on the shape of the wearout reliability curve is indicated. Relevance to current models of turbulence gust loading is discussed, and need for future studies acknowledged. Author

Descriptors:

• *STATISTICAL DISTRIBUTIONS)
• *RELIABILITY)
• (*FAILURE(MECHANICS)
• (*STOCHASTIC PROCESSES
• STATISTICAL FUNCTIONS
• VIBRATION
• NOISE
• MATHEMATICAL PREDICTION
• FAILURE(ELECTRONICS)
• DIGITAL SYSTEMS
• ACCELERATED TESTING
• PROBABILITY
• FATIGUE(MECHANICS)

Subject Categories:

• Operations Research
• Manufacturing and Industrial Engineering and Control of Production Systems
• Mechanics

Distribution Statement:

APPROVED FOR PUBLIC RELEASE