基于改进ELM的柴油机故障诊断方法研究  

Research on Diesel Engine Fault Diagnosis Method Based on Improved ELM

作　　者：王波 赵凤强 乔浩 史书杰 WANG Bo;ZHAO Fengqiang;QIAO Hao;SHI Shujie(School of Mechanical and Electronic Engineering,Dalian Minzu University,Dalian Liaoning 116650,China)

机构地区：[1]大连民族大学机电工程学院,辽宁大连116650

出　　处：《大连民族大学学报》2024年第5期400-405,共6页Journal of Dalian Minzu University

摘　　要：结合变分模态分解(VMD)和极限学习机(ELM)的柴油机故障诊断方法对柴油机的故障进行诊断分类。针对柴油机振动信号非线性、非平稳性的特点,给出了基于麻雀搜索算法(SSA)优化VMD分解方法,以此达到较好的分解性能。针对柴油机故障信号故障种类多样等问题,给出了基于灰狼优化算法(GWO)优化ELM的分类模型,使分类性能更加稳定。最后将所提出的方法用于五十铃6BB1型柴油机的故障检测与识别中,其故障识别准确率达98.04%。诊断结果验证了GWO-ELM具有较高的精准度,证明该方法是可行有效的。This article proposes a diesel engine fault diagnosis method that combines Variational Mode Decomposition(VMD)and Extreme Learning Machine(ELM)to diagnose and classify diesel engine faults.Aiming at the nonlinear and non-stationary characteristics of diesel engine vibration signals,an optimized VMD decomposition method based on Sparrow Search Algorithm(SSA)is proposed to achieve good decomposition performance.A classification model based on Grey Wolf Optimization(GWO)algorithm for optimizing ELM is proposed to address diverse types of fault signals in diesel engines,making the classification performance more stable.Finally,the proposed method is applied to the fault detection and recognition of Isuzu 6BB1 diesel engine,with a fault recognition accuracy of 98.04%.The diagnostic results verify that GWO-ELM has high accuracy,and this method is feasible and effective.

关 键 词：柴油机故障诊断 麻雀搜索算法 变分模态分解 灰狼优化算法 极限学习机 

分 类 号：TK428[动力工程及工程热物理—动力机械及工程]