A New Technique to Predict the Sources of Voltage Sags using Support Vector Regression based S-Transform

M.F. Faisal and A. Mohamed (Malaysia)


Power Quality, S-transform, SVM, SVR, Voltage sags


The concern on the quality of electrical power has become an issue amongst many electricity users around the world. When the quality of the power supply is not good, it may results in malfunction of sensitive equipments in industrial plants. It is imperative that all the power quality disturbances must be accurately detected, classified and diagnosed so that proper mitigation measures can be implemented. In this paper a new technique using the S-transform and the Support Vector Regression (SVR), was developed for diagnosing the sources of the voltage sags. The new technique was tested on 40 numbers of voltage sags and yielded satisfactory results in the prediction of sources of voltage sags. The proposed SVR based S-transform technique was also more superior than that of the learning vector quantization (LVQ) based S-transform technique.

Important Links:

Go Back