重庆理工大学学报2025，Vol.39Issue(7)：27-34,8.DOI:10.3969/j.issn.1674-8425(z).2025.04.004
数据驱动的氢燃料电池客车故障预测
Fault prediction of hydrogen fuel cell buses based on data-driven
摘要
Abstract
To address the safety risk management issues in the operation of hydrogen fuel cell buses,this paper proposes a data-driven fault prediction method which involves integrating,cleaning,and preprocessing the historical operational data of fuel cells.With the employment of the data analysis techniques,the accuracy and consistency of the data are ensured and the missing,abnormal,and duplicate values are remedied.The imbalanced indicators are tackled by SMOTE(Synthetic Minority Over-sampling Technique)and random under-sampling methods,thereby enhancing prediction accuracy.An ensemble learning approach is utilized to train base models,and a safety monitoring system is built by integrating the three best-performing machine learning models(Random Forest,XGBoost,and Bagging).The performance and prediction capabilities of the multi-indicator fault prediction model are analyzed.Results indicate the model accurately predicts various driving states of hydrogen fuel cell buses,providing reliable and precise monitoring for their operation.关键词
数据驱动/大数据/集成学习/氢燃料电池客车故障预测Key words
data-driven/big data/ensemble learning/fault prediction of hydrogen fuel cell buses分类
交通运输引用本文复制引用
程培高,李健平,何梓豪,孙仁云..数据驱动的氢燃料电池客车故障预测[J].重庆理工大学学报,2025,39(7):27-34,8.基金项目
成都市科技局重点研发项目(2022-YF05-01047-SN) （2022-YF05-01047-SN）
四川省科技厅重点研发项目(23ZDYF0506) （23ZDYF0506）
