Fault Diagnosis Method of Electrical Control System based on Improved Genetic Algorithm

As a fundamental part of intelligent apparatus, the fault diagnosis of electrical control systems is very imperative, since this traj ectory of diagnosis aims at improving reliability and safety of the system. Traditional genetic algorithms have been utilized extensively for fault diagnosis; their limitations, though, lie in accuracy and real-time performance. Therefore, this paper provides an attempt at a better method of the fault diagnosis of electrical control systems employing Genetic Algorithms. The algorithm includes innovative lock-up of genetic operations, peer-based feature extraction, and multi-obj ective optimization, hence augmenting the performance of such a diagnosis. From experimental results, the overall recognition accuracy of diagnosis by the improved genetic algorithm exceeds 97.8% average recognition rate after running five different typical fault modes, with a maximum accuracy value of 99%, thus outperforming its closest competing algorithm by a margin of nearly 8.6%. Moreover, the algorithm flies in the face of the competition in terms of time taken for diagnosis, now only requiring 2.26 <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$s$</tex> on average-again, almost precisely half of that taken by the traditional method. Thus, the proposed algorithm does not only enhance accuracy in fault recognition, but it is also more efficient and represents a useful tool for ensuring reliability and safety in electrical control systems.