Temperature Prediction Using LSTM Machine Learning for Renewable Energy

This work is a part of a big project. We predicted temperature for predicting Solar Panel Power. While we know the expected power, we can control the power flow and energy storage. This project can lead to multiple scenarios about renewable energy. Forecasting energy will help to minimize the demand for sources, promote a low-carbon economy and promote business productivity. Researchers, renewable energy companies (solar, wind, hydro, etc...), students, and economists.

This poster presents a case study of land-level temperature prediction using LSTMs in the context of renewable energy generation in microgrids. We explore the temperature forecast through LSTMs trained with pre-processed datasets consisting of historic temperatures along with relevant historic features, such as humidity, dew point, pressure, and wind speed. The results show that the LSTM models can fit our available data and generate reasonable future temperature values according to the obtained loss values. Moreover, we explore the forecast of temperature considering its historic values in an unsupervised manner using rolling LSTMs. The findings suggest that the approach presented in this paper can be used as a foundation of a machine learning-based microgrid control.