Citation

Parkinson DB. Reliab. Eng. 1980; 1(1): 29-42.

Copyright

(Copyright © 1980, Elsevier Publishing)

DOI

10.1016/0143-8174(80)90012-8

PMID

unavailable

Abstract

A reliability analysis is described in which the sample moments of the design variables are used to define the first four moments of a function representing the condition of failure, or malfunction, in the space of the design variables. These moments are then employed in transforming the failure function to a space of a normal random variable so permitting an estimate of failure probability to be made. A four-moment Reliability Index is defined and some examples of the technique are provided. The extension of the technique to stationary stochastic processes, for time-dependent problems, is discussed, and some alternative transformation procedures are compared.