Research on Grey AMSAA-ELP model on the basis of discovery learning driven reliability growth of complex equipment

Abstract

A Grey AMSAA-ELP model based on discovery learning driven reliability growth of complex equipment is proposed in accordance with the practical cases where system's reliability can be improved by discovery learning from the development process of other related equipment. The characteristics of the model are analyzed and a maximum likelihood estimation (MLE) calculation formula is built in the cases of time terminated testing and fixed failure number testing (commonly known as Type I censoring and Type II censoring). It is pointed out that if MLE has multiple extrema, the quasi- Monte-Carlo method can be used for parameter calculation. Furthermore, test methods for goodness of fit are given. Finally, failure data from an engine with certain complex equipment are analyzed by means of the proposed method, which reveals that test data fitting the Grey AMASS-ELP model outperforms that of the Grey AMSAA, and its assessment results are more consistent with engineering practices as well.