Hybrid TSA‐RBFNN based approach for MPPT of the solar PV panel under the effects of tilt angles variations and environmental effects

Abstract

This paper proposes a hybrid approach to improve the performance of photovoltaic (PV) system and track the maximum system power. The proposed hybrid approach is the combination of Tunicate Swarm algorithm (TSA) and Radial Basis Function Neural Network (RBFNN), hence it is called TSA-RBFNN. The main contribution of this paper is to “achieve the best output from solar system by tilt angles variations and environmental effects, like dust accumulation, water drops, partial shading, and maximum power point tracking (MPPT) of the solar PV panel. Here, tilt angle and orientation angles are important factor for obtaining the maximal power of the photovoltaic system with consequently the power fed to load in the PV system. The voltage, current, and PV system power are used to analyze the effect of any particle size and any weight of coal for the performance of PV modules. TSA is used to measure the PV module voltage, current, and power, then it generates a possible dataset in the offline way. The dataset has electrical parameters that are used to create a model by using RBFNN in the online way. In addition, the determination of cleaning frequency is also developed for dirty PV modules depending upon the dust deposition velocity, then the correlation among deposited dust density including power performance of photovoltaic module. The proposed method is activated in MATLAB/Simulink site under four different test cases and the efficiency is compared with other existing methods. Furthermore, optimum solutions for proposed technique, the current, voltage, power are also analyzed. In 100 iterations, the first order statistic evaluation for all the cases is obtained using proposed and existing techniques such as ANN, GBDT, SSA, and SSA-GBDT technique. In addition, the dependability, sensitivity, and accuracy of the proposed technique are also obtained. Computation time using various number of trails of proposed and existing techniques is analyzed for 100, 150, 200, 250, and 500 trails. The proposed technique achieves the trail number of 100, 150, 200, 250, and 500 trails is 48.1740 seconds, 51.2133 seconds, 71.0483 seconds, 60.00126 seconds, and 57.80132 seconds.