Adaptive Neuro-Fuzzy Voltage Control for LCL-Filter Grid-Connected Converter

• Published in: Engineering and Technology Journal Vol. 41; no. 2; pp. 316 - 332
• Main Authors: Safa Olwie, Abdulrahim Humod, Fadhil Hasan
• Format: Journal Article
• Language: English
• Published: Unviversity of Technology- Iraq 01-02-2023
• Subjects: anfis; distributed power generation; lcl filter; mathematical models of lcl filter; pi control
• ISSN: 1681-6900; 2412-0758
• DOI: 10.30684/etj.2022.132342.1115

Inductance – Capacitance – Inductance (LCL) filter is a very attractive candidate for renewable energy system applications due to its high efficiency. High attenuation of the switching frequency harmonics, small size, low fee, and improving the overall harmonic distortion (THD). This paper presents how voltage is affected by increased loads or voltage sag. Therefore it is necessary to control it with certain controllers. The Adaptive Neuro-Fuzzy Inference System (ANFIS) is used as an intelligent controller, the voltage constraint as training data for ANFIS obtained from PI. The filter works in a good connection between the inverter and the grid and rewords unwanted harmonics from using the inverter. The mathematical models for the LCL filter are investigated. The proposed approach shows more effective results than previous performance for voltage controlling and harmonic reduction. It gives overshoot (0.5%), steady state error (0.005), settling time (0.03 sec), rise time  (0.005 sec), and improving THD 8.67% to 2.33%  by comparing these results of ANFIS respectively with the results of PI which gave(3%),(0.01),(0.2sec)and( 0.02sec).