Storage reliability assessment for electromechanical components with small sampling based on prior information prediction

The storage reliability of electromechanical products such as relays and contactors, which are widely used in the aerospace and military fields, will directly affect the performance of the systems in which they are used. For the existing problem of storage reliability assessment for small samples of aerospace relays and other electromechanical products produced on a small scale, a particle filter and Bayesian theory based storage reliability evaluation method is proposed. Firstly, with the application of a particle filter, the distribution of the degradation model parameters is estimated by combining the initial distribution of degradation parameters with actual degradation data to predict the distribution of the degradation data for each test time. Secondly, we consider the predicted distribution to be prior information, then calculate the prior estimation of degradation data distribution hyper-parameters within the constraints of reliability distribution function information entropy maximization. Then we fuse the tested degradation data from the samples with the Bayesian formula to compute the posterior estimation of the hyper-parameters. After that, we obtain the interval estimation of storage reliability by solving a non-central t distribution. Finally, a specific aerospace electromagnetic relay was taken as an example to illustrate the method in detail and verify the effectiveness of the proposed method.