Application of Support Vector Machine in Water Distribution Systems' Fault Diagnosis

Aimming at the fault diagnosis of urban water supply pipeline network,an indoor experiment model of water supply pipeline network was designed.After the feature vector and nuclear function parameter were confirmed,the method for locating the bursting point of water supply pipeline network was discussed with small swatch data through testing the change of hydraulic pressure and utilizing the pattern recognition function of support vector machine(SVM) based on the theory of structural risk minimization.On the basis of the same experiment testing data,the fault diagnosis method of water supply pipeline network based on SVM was tested and compared with the method based on artificial neural network(ANN).The numerical example indicates that the fault diagnosis method based on SVM is more precise than the method based on ANN.