Impact Analysis of Soiling on Performance of Photovoltaic using ARIMA and LSTM

Solar panels in outdoor settings are frequently exposed to pollutants like dirt, dust, sediment, and bird droppings, which is results in a reduction of output total power generation by as much as $50 \%$. It is imperative to consider the effect of soiling on panels in order to precisely assess the power output of these systems. This paper proposes time series forecasting to predict power generated in clean and soiled panel. This prediction is done based on the data collected real time for clean and soiled panel. The proposed algorithm used for this purpose is ARIMA and LSTM. A comparison between ARIMA and LSTM is done through various parameters, study shows that LSTM outperform and predict soiling loss accurately as compared to ARIMA, ARIMA provides better computational time.