Power System Faults Detection Based on Deep Learning and Support Vector Machine Supervised Machine Learning Techniques

In this paper, fault detection and localization of power systems are investigated using two AI techniques. It is mainly about deep learning and support vector machine supervised machine-learning techniques. IEEE 09-bus power system is considered, and its results are compared to an IEEE 14-bus one using the accuracy score. The quantitative acquisition of fault data is performed using SimPowerSystems toolbox of Matlab. The simulation results show that the two approaches are accurate for the fault diagnosis of the power system. Both approaches proved to have fast and reliable operations. However, deep learning algorithm performances are considered more effective since it permits the classification of all types of faults. Simulation results demonstrate that the deep-learning technique achieves the accuracy of 100% compared to the support vector machine which had the accuracy of 86% and 88% for the 09-bus and 14-bus power systems respectively.