A hybrid GA-ANFIS optimization framework for enhanced transient stability and adaptive control of droop-based grid-forming inverters

The increasing penetration of inverter-based resources (IBRs) is fundamentally transforming power system dynamics, leading to reduced system inertia and heightened stability challenges. Although droop-based Grid-Forming (GFM) control provides a practical and decentralized solution, its performance remains highly sensitive to the tuning of nonlinear controller parameters. This paper proposes a hybrid, data-driven optimization framework that integrates a Genetic Algorithm (GA) for offline global optimization with an Adaptive Neuro-Fuzzy Inference System (ANFIS) for real-time adaptive control of droop-based GFM inverters. The GA stage determines an optimal parameter set by minimizing voltage and frequency deviations under multiple operating conditions, while the ANFIS model learns the nonlinear relationship between system states and optimal control actions to enable online adaptation. The proposed controller is validated through a comprehensive testing protocol, including step-load disturbances, ramp load variations, and symmetrical and asymmetrical fault scenarios, with performance evaluated against IEEE Std. 1547–2018 requirements. Simulation results demonstrate that the proposed GA–ANFIS framework significantly enhances transient stability, reduces overshoot, improves tracking accuracy, and ensures rapid post-fault recovery compared to conventional droop control and benchmark strategies. Notably, the controller maintains strict frequency regulation under dynamic loading and achieves compliant operation under fault conditions. These results establish the proposed approach as an effective and scalable solution for adaptive control of grid-forming inverters in low-inertia power systems.