A Deep Neural Network-Based Highly Simplified Intelligent Approach for Maximum Power Point Tracking of Dye-Sensitized Solar Panel System

Abstract

The article presents a highly simplified novel intelligent approach to track the maximum power point (MPP) of a commercially available dye-sensitized solar panel with a deep neural network under uniform irradiance. The panel comprising dye-sensitized solar cell is a third-generation solar cell technology. It sustains its performance at low illumination levels unlike the conventional solar cell technologies. Therefore, it can eliminate the shading issue caused by the flying birds or cloud. The MPP tracking (MPPT) techniques extract maximum power from photovoltaic systems. The biologically inspired neural networks (NNs) are widely applied techniques, solve the numerous engineering and nonengineering problems. It includes the solution to the problems, associated with MPPT. NN-based MPPT methods have made tracking systems more straightforward and robust than conventional ones. These are the distinctive and better methods in terms of tracking speed, performance accuracy, and fewer oscillation near the MPP. This is the first research work, applied novel artificial intelligence based MPPT technique with a proposed scheme on a commercial dye-sensitized solar panel. The experimentation has found the scheme effective with lesser input feature than in the literature to the NN. The results, from the proposed method and the perturb and observe method, are compared for validation. The study has found less oscillation near MPP and faster tracking response.