Modified perturb and observe MPPT with zero oscillation in steady-state for PV systems under partial shaded conditions

Abstract

This paper proposes a modified perturb and observe (MP&O) to make the conventional P&O can track the maximum power point (MPP) even under the partial shaded conditions(PSC), and doesn't need to oscillation around the MPP. It can substantially lower the power loss. With the increasing growth of renewable energy, photovoltaic (PV) systems are increasingly used to generate electrical energy from solar irradiation. In order to make the PV system can output the maximum power at all times, we can use the maximum power point tracking (MPPT) algorithm more often to reach the achievement that the system can work at the highest efficiency under the different atmosphere. Among all the MPPT methods, the Perturb and Observe (P&O) is the most commonly used because of its simplicity, lower cost and easy to accomplish. It mainly uses P-V characteristic curve to make a judgment. When the slope is zero, it is said that the system works at the MPP. However, the conventional P&O has two primary issues: (1) When the PV system occurs the partial shaded conditions (PSC), it will make the P-V characteristic curve appear multiple peak point. It will lead the conventional P&O misjudge, and make the system work at local maximum power point (LMPP), not at global maximum power point (GMPP). (2) After tracking the MPP, the work point will oscillate around it. It will cause large and unnecessary power loss. This paper proposes two methods to solve the above problems: (1) Using a function to judge whether the PSC occurs. If it occurs, set a track direction flag to search the P-V characteristic curve. If there is any LMPP can output higher power, update the information of GMPP. (2) Stop the oscillation as soon as finding the maximum point. It will minimize the loss. Through detecting the solar panel output power, we can judge that if the irradiance changes or not. This modified Perturb and Observe (MP&O) can achieve superior overall performance while maintaining simplicity of implementation. At the last, this paper also proposes the simulation to prove the feasibility of this algorithm.