Comparative Studies of Wind Power Forecasting Using Random Forest and LSTM Models

As the world puts more focus on renewable energy, wind energy is the key to reaching sustainable energy targets. Yet, its unpredictable nature causes problems for grid stability and energy management. Thus forecasting of wind power is essential to improve grid stability and security. We have discussed a methodology for producing an effective and reliable wind power forecasting model using machine learning models such as Random Forest (RF) and Long Short-Term Memory (LSTM) networks. Random Forest is a technique under supervised learning which aggregates results from several decision trees. It has good skills in finding the complex interrelations between several weather inputs and wind power outputs. LSTM is an RNN made to work with data in a sequence, ideal for capturing the patterns within time, for example, wind speeds. Therefore, we put both models in training that come from a combination of previous data on winds, weather factors (including temperature, pressure, humidity), and wind power by turbines. The result demonstrates that both RF and LSTM models perform well with the available data, but the former is more precise in determining the generation of wind power.