Accelerated life testing with several type II censored samples
McCoun, Kelly Lynn
MetadataShow full item record
The asymptotic normality of the maximum likelihood estimator for a likelihood function based on several independent Type II censored samples is established. Examples from the accelerated Hfe class of models are considered with regard to this result. This extends previous work for the single sample case. Conditional confidence interval estimation is discussed for the parameters in the Arrhenius and Eyring models based on data from several Type II censored samples. Conditional confidence interval estimates for the average life of a component under user-specified stress, the acceleration factor or the factor which expresses how many times longer a component should last at one stress relative to another, and the 7th quantile of the failure time distribution of a component under user-specified stress are also provided. Approximate confidence intervals for these parameters are given for comparison with the conditional confidence intervals.