Title: Machine Learning Approaches for Power System Fault Parameters Prediction: A Systematic Review
Authors: Biobele Alexander Wokoma, Lesuanu Dumkhana, Ibitroko Biobele Alexander Wokoma
Volume: 8
Issue: 11
Pages: 59-63
Publication Date: 2024/11/28
Abstract:
Power system faults are a perennial issue that affects many industries due to the imperfection in the design and construction of power system components and the design and implementation of the interconnected system. This recurrent issue may be effectively mitigated by identifying the causal fault factors or parameters and more specifically the interaction of these factors in an interconnected power system. These parameters or feature factors are the subject of a large number of research studies in various aspects of power system faults including the diverse fields of power system network generation, transmission, and distribution in addition to other allied fields. The use of machine learning in the power field has shown to be very promising and effective and hence it is gaining more attention in the power industry. This review paper presents a systematic study of machine learning approaches that have been used for predictions of power system fault parameters. It captures firstly, the broad reviews from a chronological perspective and then presents the specific research studies that have adopted machine learning for the prediction of power system fault parameters, including that related to energy generation, transmission, and/or distribution of power networks. Reviews presented in this paper, observed that fault stability indices are frequently adopted as feature predictors in most of the research studies investigated. This is particularly true for most distribution and transmission networks. In addition, the use of artificial neural networks is somewhat more popular amongst power system researchers.