Applications of Artificial Intelligence and PMU Data: A Robust Framework for Precision Fault Location in Transmission Lines

Providing continuous electric power supply to consumers is difficult for power system engineers due to various faults in transmission and distribution systems. Precise fault location (FL) in transmission lines speeds up the repair and restoration process. This paper proposes a wide neural network (WNN)Wide neural network approach for FL identification in power transmission networks. The proposed WNN uses the voltage, current magnitude, and phase angles measured by the phasor measurement unit (PMU)Phasor measurement unit. This proposed work considers the Western System Coordinating Council (WSCC)Western system coordinating council 9-bus test system with optimal PMU placement for fault analysis. The several types of faults created on different line sections of the test system are simulated using MATLAB/Simulink environment considering various fault parameters such as fault resistance, fault inception angle, and fault distance. The performance of the proposed scheme is measured by finding the absolute prediction error between actual and predicted FL. The results show that the average prediction errors for L-G, LL, LL-G, and LLL faults are 0.0121, 0.0209, 0.0139, and 0.0124, respectively. The proposed method outperforms the related machine learning-based FL estimation schemes for all the test cases considered at different fault locations. In addition, considering phase angle measurement improves the accuracy of finding the fault location compared to the voltage magnitude and current magnitude feature set.