Soiling Losses Estimation in Solar Panels Using Different Analytical Methods

Due to fluctuating environmental conditions many losses occurs on solar panels, soiling losses are one of them. To ensure grid stability prediction of soiling loss is essential. The two unique approaches for estimating the effect of soiling on small-scale solar plants are compared in this research. Neural network strategy is used to predict losses due to soiling. Different training-to-validation proportion is carried to choose the leading proportion for the show. There are customary strategies proposed to deterministically assess the solar power loss caused by ruining. In any case, the error of deterministic estimation cannot be disposed of due to the characteristic instability of solar power. Hence, this paper proposes a probabilistic evaluation strategy, ANN, LSTM to appraise the power loss by soiling on solar PV board. Test results indicates that R-square value and predicted power loss in LSTM is more accurate