摘要
针对水工闸门故障检测困难、检测效率低的问题,提出了一种基于小波包信息熵和改进支持向量机(SVM)算法相结合的水工闸门故障诊断方法。首先,利用小波包分解算法对弧形闸门振动信号进行分解;然后提取分解后各节点子频带信号的信息熵特征;最后将信息熵作为特征向量输入网格搜索算法优化后的支持向量机,进行故障分类器模型的训练。试验结果表明,在小样本参与训练测试的情况下,经网格搜索算法优化后的支持向量机模型仍能实现较快的训练速度和较高的分类准确率,模型的训练预测平均用时仅为1.97 s,对闸门故障的分类准确率也达到了95%,可快速有效识别闸门的故障状态。验证了该方法在水工闸门故障诊断中的可行性。
Aiming at the problems of difficulty and low detection efficiency of hydraulic gate faults, a fault diagnosis method based on wavelet packet entropy and improved support vector machine(SVM) algorithm was proposed. Firstly, the vibration signal of radial gate was decomposed by wavelet packet decomposition algorithm. Then, information entropy features of the segmented band signals were extracted. Finally, the information entropy was input as feature vector into the support vector machine optimized by grid search algorithm, and the fault classifier model was trained. Experimental results show that under the condition of the small sample to participate in the training test, the support vector machine(SVM) model optimized by grid search algorithm can achieve faster training speed and higher classification accuracy, and the average time-consuming of the training and prediction of the model is only 1.97 s, the gate of fault classification accuracy rate reached 95%, which can quickly and effectively identify the fault state of the gate. Experimental results show the feasibility of this method in hydraulic gate fault diagnosis.
作者
李凯旋
张钰奇
杨涛
李和林
段玥晨
LI Kai-xuan;ZHANG Yu-qi;YANG Tao;LI He-lin;DUAN Yue-chen(School of Mechanical and Power Engineering,Zhengzhou University,Zhengzhou 450001,China;Henan Intelligent Manufacturing Research Institute,Zhengzhou 450001,China)
出处
《水电能源科学》
北大核心
2022年第11期203-207,共5页
Water Resources and Power
基金
工信部智能制造综合标准化与新模式应用项目(2018037)
河南省水利厅水利科技攻关项目(GG202068)。
关键词
小波包分解
信息熵
支持向量机
水工闸门
故障诊断
wavelet packet decomposition
information entropy
support vector machine
hydraulic gate
fault diagnosis
作者简介
李凯旋(1995-),男,硕士研究生,研究方向为水工结构安全运维,E-mail:kaixuan216@163.com;通讯作者:段玥晨(1984-),男,博士、副教授,研究方向为复杂机械系统动力学,E-mail:duanyc1984@zzu.edu.cn。
