Solar Photovoltaic Generation Forecasting using LSTM and SARIMAX model

In several areas of load forecasting and throughout the years, energy management has concentrated on artificial intelligence (AI). In several circumstances, time series forecasting models have been developed. In this study, two separate models, Seasonal Autoregressive Integrated Moving Average with Exogenous factors (SARIMAX) and Long Short-Term Memory Loss (LSTM), are used to forecast solar photovoltaic (PV) generation, which involves several variables. As a result, the data utilized in this work is real-time data from NREL. To solve these issues, the data has been used to train a range of deep learning and machine learning models, and the outcomes provide novel insights on the benefits and need of integrating exogenous components in a time series model. Additionally, forecasting mistakes are evaluated using short-term prediction updates. Their analysis demonstrates how accurate their model is.