基于机器学习的汽车发动机零部件故障预测模型研究

打开文本图片集

中图分类号：U463.6 文献标识码：A 文章编号：1003-8639（2025）09-0115-03

Research on Machine Learning-based Fault Prediction Models for Automotive Engine Compol

【Abstract】As the core power component a vehicle，the automotive engine is subject to sudden failures that tencreate safety risks incur high repair costs.Traditional threshold-based alarms scheduled disassembly inspections suferfromhigh false-positiveratessevere latency，making itdificult tocapturetheincipient，latent degradation individual parts.With the increasing accuracy onboard sensors the advancement edgecomputing capabilities，data-driven fault prediction has become a keyresearch direction in the industry.This paper therefore presents a systematic investigation machine-learning prediction models that fuse multi-source timeseries data.Bydeveloping adaptive feature-enginering techniques dynamic optimization algorithms，the proposed approach achieves precise detection subtle fault signatures under complex operating conditions，providing reliable technical support for preventive-maintenance decisions.

【Key words 】machine learning;automotive engine; component failure; fault prediction

0 引言

近年来，随着车载多源传感技术的演进，振动频谱、热力学参数及油液状态等实时数据的高频采集，为基于数据驱动的故障预测研究提供了基础条件[。（剩余4500字）