基于小波多尺度混沌特征参数的离心泵汽蚀故障诊断

Cavitation fault diagnosis of centrifugal pump based on wavelet multi-scale chaotic characteristic parameters

提出了基于小波多尺度混沌特征参数——最小二乘支持向量机(LS-SVM)的故障诊断模型。依据试验现象和入口压力信号的分频带混沌特征参数,将不同的有效汽蚀余量(NPSH_(a))划分为4种状态;对入口压力信号进行3个尺度的小波分解,构建出各尺度低频信号的混沌参数特征向量;应用经遗传算法优化的LS-SVM实现汽蚀故障诊断。结果表明:利用分频带混沌特征能够准确描述汽蚀状态及其演变规律,更符合离心泵的实际运行情况。模型的故障诊断精度在87.5%以上,可以高效识别出离心泵的不同汽蚀状态。

Here,a fault diagnosis model for least square support vector machine(LS-SVM)based on wavelet multi-scale chaotic characteristic parameters was proposed.Firstly,according to test phenomena and chaotic characteristic parameters of inlet pressure signal in frequency division band,different effective NPSHa were divided into 4 states.Then,the inlet pressure signal was decomposed using three scales of wavelet to construct chaotic parameter eigenvectors of each scale low frequency signal.Finally,the LS-SVM optimized using the genetic algorithm was used to realize cavitation fault diagnosis.The results showed that cavitation states and evolution law can be accurately described by using chaotic characteristics in frequency division band,they are more consistent to the actual operation situation of centrifugal pump;the fault diagnosis accuracy of the proposed model is more than 87.5%,so the proposed model can effectively identify different cavitation states of centrifugal pump.