Penetration of DWT & ANFIS to Power Transmission Disturbances

Abstract

This study proposes a hybrid method to classify and estimate the location of short circuit disturbance on power transmission lines. The hybrid method uses Discrete Wavelet Transform (DWT) and Adaptive Neuro-Fuzzy Inference System (ANFIS). The transmission system is implemented in a real system, in which the electric power transmission system on the KP bus to the GS bus is with a length of 64 Km. The DWT is used to process information from each phase voltage and current transient signal as well as the zero-sequence current for one cycle after the disturbance has started. The ANFIS classification is designed to detect disturbance on each phase and ground in determining the type of short circuit disturbance. ANFIS estimation is used to measure the location of disturbance that occur on the transmission line. The training and testing data are generated by simulating the types of short circuit disturbance using software with variations in disturbance location and fault resistance. The result is that the disturbance classification is with 100% accuracy and the estimated disturbance location is with the lowest error of 0.0006% and the highest error is 0.03%.