Draft Tube Pressure Pulsation Analysis Method Based On Support Vector Machine

摘要

Based on wavelet analysis of detected draft tube pressure pulsation signals, a clustering analyzing method using the support vector machine is given. The analysis results from an example are compared with those of the self-organizing neural network. The comparison shows that satisfactory results can be obtained using the clustering analyzing method even with a small number of signal samples and the method can be used in the control of hydraulic turbine operation.