Solar Plant Intelligent Control System Under Uniform and Non-uniform Insolation

Abstract

This paper presents the solar plant intelligent control system under uniform and non-uniform insolation. In order to develop the effective solar plant intelligent control system, we further improve a modified fuzzy neural net. In this research an improved modified fuzzy neural net includes: 5 convolutional blocks and then a two-layered recurrent network, fuzzy units and two-layered recurrent networks. Based on the solar plant images and data of basic sensors, the solar plant intelligent control system provides an effective maximum power point tracking under uniform and non-uniform insolation. We trained the convolutional blocks and two-layered recurrent networks of a modified fuzzy neural net based on stochastic gradient descent optimizer and modified ant lion optimizer respectively. We showed benefits of the proposed 1 kW solar plant intelligent control system by experiments. The experimental results demonstrate that the proposed solar plant intelligent control system under uniform and non-uniform insolation robust to solar plant uncertainties and provides better control speed and performance, in comparison with the classical solar plant control system based on particle swarm optimization, or perturbation & observation algorithm.