Research on LightGBM-based fault prediction for electrical equipment in artillery fire control system

In this paper, we proposed an efficient algorithm to solve the fault prediction for electrical equipment in artillery fire control system. Aiming at the modern fire control system based on integrated modular architecture, we analyze the problem of fault diagnosis and prediction. Then fault prediction for electrical equipment is translated into a time series data prediction problem. LightGBM-based prediction model is proposed, in which the data set selection, features building, training and evaluation method is discussed. At last a numerical simulation is proposed to illustrate the efficiency of the method in this paper.