Enhanced Power control in Microgrids Using Adaptive Neuro-Fuzzy Control

Droop control is the most commonly used controller in inverter-interfaced microgrids as a primary control to regulate active and reactive power exchange. However, feeder impedance variations and slow response to dynamic load changes often restrict its performance, resulting in power-sharing inaccura-cies. To address these limitations, this paper introduces an Adaptive Neuro-Fuzzy Inference System (ANFIS)-based virtual impedance controller that dynamically adjusts the inverter's reference voltage to improve power control and response time. The proposed controller continuously compensates for impedance mismatches by adjusting the virtual voltage, and it ensures more accurate tracking of active and reactive power with minimal deviations. By leveraging the combined strengths of fuzzy logic and neural networks, ANFIS eliminates the need for manual tuning and enhances control adaptability in nonlinear microgrid environments. The proposed controller is validated using a 100 k W battery energy storage in islanded and grid-connected in various operating conditions and load changes in charging and discharging mode, and comparative results with traditional controller highlights the superior tracking accuracy, reduced response time, and improved system resilience of the proposed approach, making it a viable solution for enhancing microgrid performance.