Multiple Performance Parameters Fault Prediction Method Based on Gamma Process and Resemblance Coefficient

Abstract

A 20kHz signal board of radar is the research object and the circuit board reliability analysis is made by using performance degradation reliability theory. The rule of radar circuit board performance degradation in the electric impact environment was concluded with the testing data. Aiming at the insufficient of the multiple performance degradation parameters fault prediction method existing, a new method based on resemblance coefficient is present. The new method combine the signal processing technology with the performance degradation theory. The resemblance coefficient fused the multiple performance parameters degradation characteristics, so the multi-feature extraction difficulties of multiple performance parameters has been solved. Then, the Gamma process is used to describe resemblance coefficient degradation for modeling, extrapolated life information of the product, and finally, validity of this method was proved by the testing data.