Power Quality Disturbance Recognition Employing State Vector Machine Methods

Jiansheng Huang, Zhuhan Jiang, Leanne Rylands, and Michael Negnevitsky


Power Quality Disturbance Recognition, Digital Wavelet Transform, Support Vector Machines, Cross Validation, Parameter Optimization


This paper presents a power quality disturbance recognition system employing support vector machine (SVM) techniques. Based on site measurements, a waveform generator is designed to emulate different power quality disturbances existing in modern power distribution systems. Digital wavelet transform (DWT) is then applied to the sampled waveforms for feature extraction. Thereby obtained DWT coefficients are further exploited to identify the associated disturbances through constructing an SVM classifier for each type of waveforms. Simulation results demonstrate that the SVM based classifiers can achieve significantly higher recognition rates compared with conventional methods.

Important Links:

Go Back