Failure data analysis by models involving 3 Weibull distributions

Abstract

Two kinds of models concerning different combinations of Weibull distributions are studied, where each of the models involves three Weibull-distributions. First, the characteristics of the fitting plots determined by these models on the Weibull plotting paper (WPP) are analyzed. Then, the appropriateness for the models being applied to fitting given failure data is described. For each of the two kinds of models' formulas, parameter estimates are discussed based on the plots on WPP. Finally, an example with real-data illustrates their application. The models presented in this paper can provide flexibility in fitting and explaining failure data. They are more appropriate to be applied to analyzing complex data than models with two Weibull distributions or a mixture of them. For getting numerical estimates of model parameters, the maximum likelihood estimation method can be applied.