Maximum Power Point Tracking for Photovoltaic System Using Adaptive Extremum Seeking Control

In order for photovoltaic (PV) systems to maximize their efficiency of power generation, it is crucial to locate the maximum power point (MPP) in real time under realistic illumination conditions. The current-voltage (I-V) characteristics of PV devices are nonlinear, and the MPP may vary with intrinsic and environmental conditions. Maximum power point tracking (MPPT) control is expected to seek the MPP regardless of the device and ambient changes. This brief presents the application of the adaptive extremum seeking control (AESC) scheme to the PV MPPT problem. A state-space model is derived via averaging method, with the control input being the duty ratio of the pulse-width modulator of the dc-dc buck converter. To address the nonlinear PV characteristics, the radial basis function neural network is used to approximate the unknown nonlinear (I-V) curve. The convergence of the system to an adjustable neighborhood of the optimum is guaranteed by utilizing a Lyapunov-based adaptive control method. The performance of the AESC is verified with simulation.