Fault prognostic technology of complex electronic equipment for PHM

In order to realize the complex electronic equipment prognostic and health management, it needs to research its core technology in fault prognostic. Based on the existing fault prognostic methods, this paper puts forward a support vector machine nonlinear fault prognostic model based on performance degradation data as input and reliability data as output. The working principle of fault prognostic model is to train multioutput SVM to fit the nonlinear relationship between performance degradation data and reliability data. The reliability of components can be predicted by using the trained SVM. Finally, a magnetron experimental data as an example is used to verify the prognostic model in terms of prognostics for the electronic products.