Performance Evaluation of a Hybrid Passive Filter and AI-Based Optimization Techniques for Harmonic Distortion Mitigation in the Nigerian Power Network

Non-linear loads are the primary causes of harmonic distortion in power system network and these loads are essential house hold equipment that are essential for daily living and for comfort. However, the occurrences of harmonic distortions result to the power system failure leading to blackouts. This paper determined the impact of distortion on the current and voltage signals of Nigerian 330 kV system in the North Central region and utilized the machine learning controlled passive filter for the mitigation of the impact of the harmonic distortion. The case study region had 6 buses and the total harmonic distortion (THD) for these buses were obtained. The THD of the current signal for without, with standalone passive filter and RNN-controlled passive filter were (16.21 %, 4.73 % and 1.89 % for bus 1), (17.55 %, 4.41 % and 2.01 % for bus 2), (18.72 %, 3.88 % and 1.91 % for bus 3), (15.93 %, 3.12 % and 1.89 % for bus 4), (17.51 %, 4.83 % and 1.19 % for bus 5) and (17.04 %, 4.81 % and 1.03 % for bus 6). The outcome voltage signals obtained were (13.21 %, 4.33 % and 2.72 % for bus 1), (14.81 %, 3.73 % and 2.13 % for bus 2), (16.44 %, 4.34 % and 2.92 % for bus 3), (12.92 %, 3.93 % and 2.72 % for bus 4), (15.24 %, 4.53 %, and 2.12 % for bus 5) and (14.55 %, 4.41 % and 2.02 % for bus 6) respectively. It was observed that the RNN-controlled filter had a better performance than the standalone passive filter.