Performance degradation analysis of reliability growth of airborne power supply based on logistic regression

Abstract

Abstract With the improvement of reliability design of nonlinear airborne power supply system, it is more difficult to analyze the performance degradation of components under lean failure data in reliability growth test. In this paper, a state space model of performance degradation was established for nonlinear airborne power system, and the method of performance degradation modeling and analysis was studied by using nonlinear logistic regression process. Taking electrical, mechanical and thermal stress as characteristic failure loads, a nonlinear logistic regression analysis model was established to describe the performance degradation under comprehensive stress. The state of the degradation space is divided by the Markov distance, and the performance degradation trajectory is transformed into the degradation state space by using the maximum likelihood model and correlation degree model of observed variables, the degradation space trajectory of the performance of the characteristic parameters under multi-input variables is obtained. The correctness and validity of the nonlinear logistic regression model and the degradation process analysis method are verified by the test on the brushless motor reliability test platform.