Hybrid Fuzzy–Neural Droop Control for Fast Voltage and Frequency Stability in Renewable Agricultural Microgrids

Abstract

This paper introduces a Hybrid Fuzzy–Neural Droop Control (FNDC) approach designed for renewable agricultural microgrids that combine photovoltaic (PV), biogas, and battery energy storage systems (BESS). The control framework merges the nonlinear adaptability of fuzzy logic with the learning ability of neural networks to improve voltage–frequency stability, speed up transient responses, and reduce steady-state errors amid fluctuating renewable generation and nonlinear agricultural loads. The fuzzy component dynamically adjusts droop coefficients based on voltage deviations, power imbalances, and load variation rates. Meanwhile, the neural component continuously refines fuzzy membership parameters through an online gradient-based learning law. The FNDC is implemented and tested on a Python-based microgrid simulation platform utilizing pandapower and NumPy/SciPy. Comparative results against standard and fuzzy-only droop controllers show that the FNDC reduces settling time by up to 67%, cuts mean absolute error (MAE) by 45% and decreases RMSE by over 50% for voltage and frequency regulation. Additionally, the adaptive and decentralized design of the FNDC provides robustness against communication delays and scalability for rural deployment. The proposed strategy offers an intelligent and efficient control framework for next-generation smart agricultural microgrids powered by hybrid renewable energy sources.