基于PCA和PNN柴油机故障诊断方法  

Based on PCA and PNN Diesel Engine Fault Diagnosis Methods

作　　者：陈峰[1] 范兴奎[1,2] 厉志达 Chen Feng;Fan Xing-kui;Li Zhi-da(Basic Department of Linyi Campus of Qingdao University of Technology,Qingdao 266520,China;School of Science,Qingdao University of Technology,Qingdao 266520,China;Department of Mechanical and Electronic Engineering,Linyi Campus,Qingdao University of Technology,Qingdao 266520,China)

机构地区：[1]青岛理工大学基础部,山东青岛266520 [2]青岛理工大学理学院,山东青岛266520 [3]青岛理工大学机械与电子工程系,山东青岛266520

出　　处：《内燃机与配件》2023年第24期107-110,共4页Internal Combustion Engine & Parts

基　　金：山东省以探索和实践数学建模创新实验班培养模式为研究内容,于2021年开展了本科教学改革研究课题Z2021114。2022年山东省大学生创新创业训练计划项目-基于PCA和PNN的柴油机故障诊断-S202210429168。

摘　　要：针对舰船柴油机智能故障诊断中因故障样本不足而导致的诊断模型准确度不高的问题,提出一种基于PCA和PNN的柴油机故障诊断方法。对改进柴油机故障分级车型的效果提出了两个步骤。使用主成分方法对样本量进行分类;随后,在提高故障诊断准确性的同时,利用概率神经网络(PNN)建立模型,强化其泛化性能。最后经过项目测试、比对测试等多方验证,得出结论:所研究的方法可以对柴油机故障进行精确诊断,其优点是精度高,运行时间短,适用性广。Aiming at the problem that the diagnostic model is not accurate due to insufficient fault samples in the intelligent fault diagnosis of ship diesel engines,a fault diagnosis method of diesel engine based on PCA and PNN is proposed.Two suggestions are put forward to enhance the efficacy of diesel engine fault classification models.First,the principal component analysis method(PCA)was used to expand the capacity of invalid samples;Subsequently,we used probabilistic neural networks(PNNs)to build models that enhanced their generalization performance and improved the accuracy of fault diagnosis.Finally,after project testing,comparison test and other verifications,it is concluded that the studied method can accurately diagnose diesel engine faults,and its advantages are high precision,short running time and wide applicability.

关 键 词：柴油机 故障诊断 主成分分析法(PCA) 概率神经网络(PNN) 

分 类 号：TK428[动力工程及工程热物理—动力机械及工程]