基于案例推理的轴系故障智能诊断方法研究  

Research on Intelligent Diagnosis Method of Shafting Fault Based on Case Reasoning

作　　者：张荣国 刘金林[1] 房诗雨 古铮 ZHANG Rongguo;LIU Jinlin;FANG Shiyu;GU Zheng(College of Power Engineering,Naval University of Engineering,Wuhan 430033)

机构地区：[1]海军工程大学动力工程学院,武汉430033

出　　处：《舰船电子工程》2023年第6期140-144,199,共6页Ship Electronic Engineering

摘　　要：针对船舶远海航行时,现有轴系故障诊断技术不完善不适用问题,提出将案例推理应用于轴系故障诊断。以学习型伪度量(Learning Pseudo Metric,LPM)作为案例间的相似度度量,运用蝗虫优化(Grasshopper Optimization Algorithm,GOA)算法优化BP神经网络(GOA-BP)训练该度量函数并预测案例相似度,以此建立轴系案例推理故障诊断模型。最后通过轴系故障数据对该方法进行性能测试,并将其与基于BP神经网络的学习型伪度量模型(LPM-BP)、基于遗传算法(GA)优化BP神经网络的学习型伪度量模型(LPM-GA-BP)、基于欧式距离的最近邻算法(KNN)以及支持向量机(SVM)等算法进行对比,结果表明,论文研究方法能有效提升故障诊断效率,对轴系故障诊断具有一定的推广应用价值。Aiming at the problem that the existing shafting fault diagnosis technology is not perfect and applicable when the ship is sailing at sea,the case-based reasoning is applied to shafting fault diagnosis.Taking learning pseudo metric(LPM)as the similarity measure between cases,the grasshopper optimization algorithm(GOA)algorithm is used to optimize BP neural network(GOA-BP)to fit the measurement function and predict the case similarity,and the case-based reasoning fault diagnosis model of shafting is established.Finally,the performance of this method is tested by shafting fault data,and compared with the learning pseu⁃do metric model based on BP neural network(LPM-BP),the learning pseudo metric model based on genetic algorithm(GA)opti⁃mized BP neural network(LPM-GA-BP),the K nearest neighbor algorithm based on Euclidean distance(KNN)and support vector machine(SVM).The results show that the method studied in this paper can effectively improve the efficiency of fault diagnosis.It has certain popularization and application value for shafting fault diagnosis.

关 键 词：案例推理 故障诊断 学习型伪度量 蝗虫优化 船舶轴系 

分 类 号：U664.2[交通运输工程—船舶及航道工程]