Wind power prediction using optimized MLP-NN machine learning forecasting model

In recent years, the utilization of wind turbines to harness wind power has experienced significant growth, driven by technological advancements and increasing emphasis on sustainability. Developing nations, including India, are strategically implementing wind power initiatives in regions characterized by high annual average wind speeds. However, due to the inherently unpredictable nature of meteorological variables, wind energy production lacks a predictable schedule. Consequently, the energy sector faces the challenge of accurately forecasting wind velocity to effectively manage and optimize wind electricity generation. This research aims at Andhra Pradesh, India, to see whether wind energy is feasible. This paper presents a multilayer perceptron-neural network (MLP-NN) forecasting model that utilizes machine learning. The model extracts features with the Harris Hawks Optimization-K-Nearest Neighbors (HHO-KNN) technique. By employing this strategy, open-access data gathering will be used to improve wind power predictions. As the result, the proposed model attains an accuracy of 97.0996%, a precision of 99.16731%, an F1 score of 97%, and a recall of 96%. According to the results, the Prakasam district has the most potential for producing wind energy. Even though the sites are small in comparison with larger-scale projects, they have the potential to produce considerable amounts of electricity. Strategic decisions in renewable energy planning are informed by this research, emphasizing the significance of sophisticated forecasting approaches for wind energy optimization.