A dynamic reconfiguration method based on neuro-fuzzy control algorithm for partially shaded PV arrays

Abstract

The partial shading on PV arrays causes power decrease, hot spots, and damage to its components. In addition, the performance of the PV array is reduced for the internal mismatch. This work proposes a novel dynamic, on line, low costs, and automatic method to mitigate these effects. The method considers the mismatch due to partial shading and cell variability. In this method, the entire PV array is fully dynamic, and auxiliary PV modules are not required. The PV modules are electrically rearranged while the same interconnection scheme is maintained. The method is based on the temperature, the global voltage, and the global current of the entire PV array. These global measurements reduce the number of sensors, signal processing, computing time, electrical connections, and implementation cost. The electrical rearrangement of the PV modules is controlled by a neuro-fuzzy algorithm, a connection control, and a switch matrix. The proposed approach was implemented in hardware and validated experimentally in a real PV array. The results show that the method has 97% effectiveness in distinguishing the cause of power decrease, a 100% effectiveness in locating the shaded PV module, and a 10% increase in the PV array output power.