Reliability Evaluation Method of Aerospace Relays Based on Fusion of Simulation Data and Small Sample Test Data

In this paper, we present a reliability evaluation approach that combines simulation data with a small amount of test data. This approach solved to lack of sufficient test data of storage degradation, which leads to failure of reliability evaluation in the initial phase of development of aerospace relays. Firstly, the Inverse Gaussian process is used to model degradation data which is strictly monotone. Secondly, based on the Bayesian theory, a posteriori estimation and a priori estimation of the model parameters are listed. Then, the process of using EM algorithm to solve hyperparameters is introduced. Finally, in combination with the actual case, the method of integrating simulation data and experimental data are used to conduct the Inverse Gaussian process degradation modeling and the reliability evaluation in combination with the Bayesian theory. The results show that the performance degradation modeling based on simulation data can effectively evaluate the reliability. And the introduction of a small amount of test data can will further improve the accuracy of reliability assessment. The proposed method could provide a new idea of reliability evaluation for products with high reliability and small samples.