In the multistage reliability growth tests with instant and delayed fix modes, the failure data can be assumed to follow Weibull processes with different parameters at different stages. For the Weibull process within ...In the multistage reliability growth tests with instant and delayed fix modes, the failure data can be assumed to follow Weibull processes with different parameters at different stages. For the Weibull process within a stage, by the proper selection of prior distribution form and the parameters, a concise posterior distribution form is obtained, thus simplifying the Bayesian analysis. In the multistage tests, the improvement factor is used to convert the posterior of one stage to the prior of the subsequent stage. The conversion criterion is carefully analyzed to determine the distribution parameters of the subsequent stage＇s variable reasonably. Based on the mentioned results, a new synthetic Bayesian evaluation program and algorithm framework is put forward to evaluate the multistage reliability growth tests with instant and delayed fix modes. The example shows the effectiveness and flexibility of this method.展开更多
In this paper the AMSAA model is generalized to the reliability growth testing of mul-ti -system simultaneous development, therefore it is named the AMSAA -BISE model . The MLE and unbiased estimates of the parameters...In this paper the AMSAA model is generalized to the reliability growth testing of mul-ti -system simultaneous development, therefore it is named the AMSAA -BISE model . The MLE and unbiased estimates of the parameters for the system and a goodness of fit test from failure-truncated data and time truncated data are given.展开更多
基金supported by Pre-research Foundation of General Armament Department of China(xxxxxxxxxxxx06KG0164)and the National Doctoral Foundation of China (2005999807).
文摘In the multistage reliability growth tests with instant and delayed fix modes, the failure data can be assumed to follow Weibull processes with different parameters at different stages. For the Weibull process within a stage, by the proper selection of prior distribution form and the parameters, a concise posterior distribution form is obtained, thus simplifying the Bayesian analysis. In the multistage tests, the improvement factor is used to convert the posterior of one stage to the prior of the subsequent stage. The conversion criterion is carefully analyzed to determine the distribution parameters of the subsequent stage＇s variable reasonably. Based on the mentioned results, a new synthetic Bayesian evaluation program and algorithm framework is put forward to evaluate the multistage reliability growth tests with instant and delayed fix modes. The example shows the effectiveness and flexibility of this method.
文摘In this paper the AMSAA model is generalized to the reliability growth testing of mul-ti -system simultaneous development, therefore it is named the AMSAA -BISE model . The MLE and unbiased estimates of the parameters for the system and a goodness of fit test from failure-truncated data and time truncated data are given.
