A Comparative Analysis on Wind Speed Forecast using Optimized Neural Networks

Abstract

Electrical energy is an essential input for the improvement of a country's economy. A consistent source of electrical power is vital to support and improve living standards. Wind energy is seen as a commercially appealing and rapidly growing electrical energy resource of lower environmental impact and cost-effectiveness. Thus, the percentage of electrical power produced by wind energy in the energy sector has increased dramatically during the last few years. The fluctuating characteristics of wind power present severe challenges to electrical power transmission. Therefore, a precise wind power forecast is crucial for the successful operation of the wind farm in a reliable power distribution system. In this paper, A hybrid Neural Network model is proposed for a high-performance strategy for estimating wind speed. Weights of five neural networks, each with a different structure are adjusted with PSO and Genetic Algorithm. Trial cases for additional months of 2012 are called for validation. The results and comparisons with other wind speed forecasting models show that the introduced model provides a better wind speed forecast.

Keywords

• artificial neural network
• genetic algorithm
• particle swarm optimization (pso)
• wind speed forecasting