面向阀门内漏声发射检测的支持向量机分类建模

针对阀门泄漏声发射检测的研究中存在回归建模不准确的问题,综合考虑实际应用需要,开展了阀门液体泄漏的分类建模研究。分析了阀门泄漏声发射的机理和基本特征,建立了阀门泄漏声发射信号特征量与泄漏等级的支持向量机分类模型。在工业生产现场进行阀门泄漏声发射信号采集实验,对采集到的信号进行预处理和特征提取。采用网格搜索法寻找最优训练参数,建立最优的支持向量机分类模型,模型预测结果表明:阀门泄漏预测模型识别准确率在93%以上。

Support Vector Machine Classification Modeling for Acoustic Emission Detection of Valve Internal Leakage

Aiming at the problem of inaccurate regression modeling in the current research on acoustic emission detection of valve leakage, considering the practical application needs, a classification modeling study of valve liquid leakage is carried out. The mechanism and basic characteristics of valve leakage acoustic emission are analyzed, and a support vector machine(SVM) classification model for valve leakage acoustic emission signal feature quantity and leakage level is established. The valve leakage acoustic emission signal collection experiment is performed at the industrial production site, and the collected signals are pre-processed and feature extracted. The grid search method is used to find the optimal training parameters, and the optimal support vector machine classification model is established. The model prediction results show that the accuracy of valve leakage model prediction ang identification is more than 93%.