基于随机森林算法与贝叶斯网络的逆变器温度预测方法

打开文本图片集

中图分类号：TP391.92；TQ150.5 文献标志码：A

文章编号：1001-5922（2025）09-0137-04

Inverter temperature prediction method based on random forest algorithm and Bayesian network

ZHAI Yinan， SUN Geng （College of Information Engineering，Dalian Ocean University，Dalian 116023，Liaoning China）

Abstract：As akeycomponent of the trainauxiliarysystem，theabnormal temperatureof the inverter willaffect the normaloperation of the train.Inorder to avoid the normal operationof thetrain due to the over-temperature failure of the inverter，the temperature prediction is of great practical significance.In this paper，an inverter temperature prediction model based on random forest algorithm combined with Bayesian network was proposed.The Random Forest was used to calculateand filter the characteristicsof the train assistance systemdata，and the important features related to the temperature of the inverter weredetermined through its importance index，soas to reduce thedata dimension，and then weighted fusion with the Bayes Net regresion prediction model.By continuouslyupdating the posterior distribution basedon prior knowledge，it predictedthe future temperatureof the inverter.Through experiments on the inverter dataset，the prediction accuracy of the method reached more than 95% ，which indicates the effectiveness of the proposed method.

Key words：random forest;bayes net;inverter temperature;data dimension reduction

随着我国轨道交通行业规模的持续扩大，轨道交通的安全问题成为民生安全部署的核心任务之一，轨道交通车辆辅助逆变器的温度预测对轨道交通安全至关重要。（剩余6382字）