基于改进混合蛙跳算法的SVM潜水轴流泵故障诊断

Fault diagnosis of submersible axial flow pump based on ISFLA-SVM

为了提高潜水轴流泵故障诊断的准确性,提出了一种采用改进混合蛙跳算法(ISFLA)优化的支持向量机(SVM)的故障诊断模型。SVM参数的选取对故障分类结果影响很大,改进混合蛙跳算法克服了SFLA中种群趋同性和易陷入局部最优解的不足,利用ISFLA对SVM参数进行优化处理能够提高故障识别精度。为了验证该模型,搭建了潜水轴流泵故障实验平台,采集了正常状态、转子不对中和动静碰摩三种状态下的数据,分别用BP神经网络、蚁群优化ACO-SVM和ISFLA-SVM算法对它们进行分类故障识别。实验结果表明,ISFLA-SVM故障分类识别精度高于其他两种分类算法。

A fault diagnosis model of support vector machine(SVM)optimized by improved shuffled frog-leaping algorithm(ISFLA)is presented to improve the accuracy of fault diagnosis of submersible axial flow pump.The SVM parameters have a great impact on the fault classification results.ISFLA overcomes the population convergence and the defect of trapping into local optimal solution in the traditional SFLA.Using ISFLA to optimize SVM parameters can improve the accuracy of fault classification.To verify the proposed model,we set up an experimental platform for submersible axial flow pump faults,collecting the data under three states:normal state,rotor misalignment,rub-impact fault,and exploit back propagation neural network(BPNN),ant colony optimization(ACO)-SVM and ISFLA-SVM algorithms for fault classification,respectively.Experimental results show that the fault classification accuracy of ISFLA-SVM is higher than other two algorithms.