Hybrid optimized machine learning neural network models for multi step wind speed forecasting

Abstract

Over the decades, the requirement of power is a major constraint across the globe and the utilization of renewable energy has gained its significant mportance all over. Considering the economic growth, the energy sources play a vital role and developing energy from nature will facilitate proper allocation of the identified resources. Currently, the financial growth of a nation is decided based on the power resources available in the country. So, the focus lies in utilizing the resources that are available in nature and to bring out the most from it in respect of power generation. One such energy source available in nature is wind resources and is a form of renewable energy that exists in plenty in India. The well-known factor is wind energy is clean and free from pollutants. The development of wind energy from the natural wind flow is with respect to the force with which it moves or based on the speed of the wind. Apart from its basic characteristics, wind owns the capacity for generating the power needed for the regular demands of a nation. The forecasting of wind speed is essential so as to enhance the energy requirement and wind speed forecasting lay a compromise between the energy generated and the required demand. Neural network models are applied in forecasting applications due to their stability, adaptability, handling large data, non-linearity and its capability for generalization. In respect of these features of neural networks, the neural algorithms are employed for forecasting wind speed for given input parameters in renewable energy applications. Neural computing models do not require mathematical equations or mathematical models of a system but satisfies the set criterion error based on the available input and output data. Considering all these factors, this research thesis focuses on developing hybrid optimized machine learning neural network models for predicting multi-step ahead wind speed in renewable energy applications. newline