PERFORMANCE ENHANCEMENT OF PHOTOVOLTAIC PANELS IN SOLAR ENERGY SYSTEMS USING DEEP LEARNING ALGORITHM

Abstract

Solar photovoltaic (PV) systems are gaining importance increasingly as it directly converts solar radiation into electrical energy which is renewable and environment friendly. Where it has a numerous advantage, some disadvantages are also there like its dependency on environmental conditions. The power developed by solar panel decreases if it does not get uniform radiation. Sometimes due to nearby buildings, passing clouds etc. PV module might be partially shaded because of which power output of solar panel may get decrease this is called partial shading conditions. It causes significant reduction in the system power output. To overcome this, maximum power point-tracking under partial shading condition by continuous duty cycle variation schemes have been proposed, in which proposed LUO converters are connected to PV module to enable maximum output voltage at any given condition. In this proposed system a new method of Artificial Neural Network based Maximum Power Point Tracking (ANN-MPPT) has been implanted, which is capable of tracking the Maximum Power Point in the presence of other local maxima. The proposed scheme tracks Maximum Power Point (MPP) by continuous variation of converter’s duty cycle without the use of costly components such as signal converters and microprocessors thereby increasing the compactness of the system. The converter's goal is to operate at unity power factor and provide input currents with a tolerable harmonic content in a grid interface. Finally, to validate the proposed controls, we will conduct a series of numerical simulations using MATLAB 2021a /Simulink software and hardware is implemented in DSPIC30F4011 Controller.