Life Prediction of Metro Vehicle Relay Based on Multi-kernel Gaussian Process Regression
Electromagnetic relays play a crucial role in the automatic control systems in metro vehicles,which are extensively utilized.When a relay fails,it can result in significant delays,including vehicle going offline and necessitating casual repairs.This paper proposes a life prediction model for metro vehicle relay based on multi-kernel Gaussian process regression and employs Pearson correlation to analyze the characteristic parameters of the relays and utilizes Gaussian kernels and adaptive Gaussian kernel functions to model the covariance,as well as cubic polynomials to model the basis functions.The remaining life of the relay is modeled as a Gaussian distribution,and the model provides point estimates and variance for the remaining life.It calculates the probability density function and cumulative distribution function for the remaining life,deriving interval estimates for the remaining life and the probability of operating for at least a specified duration.A complete life cycle test is conducted on the developed relay life prediction test bench,and a case analysis is performed,demonstrating the effectiveness of the model algorithm.
metro vehicleelectromagnetic relayremaining life predictionmulti-kernel Gaussian process regressionkernel function
NETL
NSTL
万方数据
