基于LSTM-CNN特征提取和PSO-KNN分类的自动抓梁液压系统故障诊断  

Fault Diagnosis for Hydraulic System of Automatic Grabbing Beam Based on LSTM-CNN Feature Extraction and PSO-KNN Classification

作　　者：刘文忠[1] 张世杰 金兰 王瑞辰 LIU Wenzhong;ZHANG Shijie;JIN Lan;WANG Ruichen(Xiluodu Hydropower Plant,China Yangtze Power Co.,Ltd.,Yongshan Yunnan 657300,China;School of Aeronauticsand Astronautics,University of Electronic Science and Technology of China,Chengdu Sichuan 610037,China;Zhengzhou Kede Automatization System Engineering Co.,Ltd.,Zhengzhou Henan 450000,China)

机构地区：[1]中国长江电力股份有限公司溪洛渡水力发电厂,云南永善657300 [2]电子科技大学航空航天学院,四川成都610037 [3]郑州市科德自动化系统工程有限公司,河南郑州450000

出　　处：《机床与液压》2024年第18期203-207,共5页Machine Tool & Hydraulics

基　　金：中国长江电力股份有限公司科研项目(Z412202018);国家自然科学基金面上项目(72171037);国家自然科学基金青年科学基金项目(71801168);四川省自然科学基金项目(2023NSFSC0476)。

摘　　要：针对自动抓梁液压系统故障诊断正确率低、深层特征提取困难的问题,提出一种基于长短期记忆卷积(LSTM-CNN)特征提取网络和粒子群优化K最近邻(PSO-KNN)结合的自动抓梁液压系统故障诊断模型。以自动抓梁液压系统关键节点压力信息为输入,采用LSTM提取一维特征与CNN提取的二维特征融合,采用优化后的KNN模型对提取的特征进行故障分类。基于真实数据搭建AMESim自动抓梁模型进行仿真,验证所提方法的有效性与先进性。结果表明:所提模型的诊断正确率达到97.92%,能够有效识别自动抓梁液压系统中的常见故障。The fault diagnosis for hydraulic system of automatic grabbing beam faces the problems of low accuracy and difficulty in deep feature extraction.A diagnosis model integrating long short-term memory convolutional neural network(LSTM-CNN)feature extraction network and particle swarm optimization K-nearest neighbors(PSO-KNN)was proposed.Taking the pressure information of key nodes in hydraulic system of automatic grabbing beam as inputs,LSTM was adopted for 1D feature extraction and combined with 2D feature extracted by CNN.The optimized KNN model was then utilized for fault classification based on the extracted features.The AMESim automatic grabbing beam model based on real data was built for simulation,to verify the effectiveness and progressiveness of the proposed model.The results demonstrate that the proposed model achieves a diagnostic accuracy of 97.92%and it can effectively identify common faults in hydraulic system of automatic grabbing beam.

关 键 词：液压自动抓梁 LSTM-CNN PSO-KNN 故障诊断 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]