Real-time photovoltaic smoothing with supercapacitors: Low-complexity supervisory selection of conventional filters

This work introduces a hybrid power smoothing strategy for photovoltaic systems that combines three conventional filtering methods ramp-rate limitation, moving average, and low-pass filtering under a fuzzy logic-based supervisory controller. The proposed algorithm dynamically selects the most suitable filtering technique at each time step based on real-time measurements of photovoltaic power variability and the state of charge of a supercapacitor. Unlike static filters with fixed parameters, the fuzzy-driven control framework enables context aware adaptation to operating conditions, improving system responsiveness and power stability without over relying on energy reserves. The strategy was implemented in a real-time control platform and validated using a 15 kWp photovoltaic emulator and a 30 kW of supercapacitor based storage unit. Experimental results show that the proposed method reduces the standard deviation of the power output by 38.1%, the ripple index by 15.4%, and the state of charge variance by 33.4%, while limiting energy deviation to less than 0.55 kWh. These outcomes indicate enhanced power quality and improved operational sustainability of the storage system. The proposed approach offers a scalable and hardware-compatible solution for real-time photovoltaic fluctuation mitigation in low-inertia or distribution-level networks.