Efficient multidimensional Maximum Power Point Tracking using Bayesian fusion

We address the topic of a unified controller for Maximum Power Point Tracking (MPPT) controller in distributed photovoltaic (PV) modules. The power produced by a PV module depends on the solar irradiance and temperature. MPPT controllers adaptively search and maintain operation at the maximum power point for changing irradiance and temperature condition, thus maximizing the panel power output and consequently minimizing the overall system cost. Various conventional MPPT algorithms have been proposed for ideal conditions, few algorithms were derived to extract true maximum power under partial shading conditions, and very few have addressed the problem of continuously changing shading conditions caused by changing weather conditions, e.g. rain, clouds. Under dynamically changing conditions, the conventional MPPT controllers can't find the true MPP (global MPP) and are often track to a local one. In this work, results are obtained for a tracking algorithm based on Bayesian fusion combined with swarm intelligence. Compared to state-of-the-art trackers, the system achieves global maximum power tracking and higher efficiency for modules with different optimal current, caused by continuously changing uneven insolation.