Intelligent Type-3 ANFIS-based adaptive secondary control for parallel grid-forming inverters in islanded smart microgrids

This article presents an adaptive secondary control method for islanded AC microgrids using an Enhanced Type-3 Adaptive Neuro-Fuzzy Inference System (T3-ANFIS). The considered microgrid structure consists of two parallel grid-forming (GFM) inverters operating under different loading conditions. In conventional droop-based systems, decentralized power sharing causes steady-state frequency and voltage deviations, which can negatively affect system stability and power quality. Therefore, an effective secondary control layer is required. To overcome these issues, a hybrid adaptive control structure is introduced. The proposed architecture contains two T3-ANFIS structures working together. The first T3-ANFIS is used to learn and represent the nonlinear system dynamics, whereas the second T3-ANFIS adaptively adjusts the PI controller parameters in real time according to the learned system behavior. Moreover, a Correntropy Kalman Filter (CKF) is integrated into the system for robust state estimation under non-Gaussian disturbances, improving overall system reliability. Simulation results show that the proposed controller provides better frequency restoration, voltage regulation, and power-sharing performance compared to the conventional control approach under different operating conditions. Moreover, it achieves faster dynamic response, reduced settling time, lower steady-state error, and decreased total harmonic distortion (THD), highlighting its effectiveness for inverter-based islanded microgrids.