Weibull component reliability evaluation with masked data

Abstract

Under the condition that the exact component causing system failure is masked in life testing, a new method is proposed to evaluate the reliability of Weibull component in a series system. Considering relaxing the symmetry assumption of masking probability, we obtained the likelihood function of the Weibull components on the basis of the masking probability does not depend on the distribution parameters. A latent variable is introduced to rewrite the form of the likelihood function, and we obtain the estimators of the unknown parameters. The confidence intervals of Weibull component life-distribution parameter are also inferred with likelihood ratio method. The approach is illustrated with a simple numerical example.