A GA-PSO Hybrid Algorithm Based Neural Network Modeling Technique for Short-term Wind Power Forecasting

Abstract

This article proposes a hybrid neural network modeling techniquefor forecasting of wind power generation based on an integrated algo -rithm combining genetic algorithm (GA) and particle swarm optimi-zation (PSO). The share of wind energy in electric power generationkeeps growing supported by favorable environmental policies aimingat achieving low-emission targets. However, due to the intermittent anduncertain nature of wind flow, integration of wind power into electricpower systems brings operational challenges to address. Accurate windpower generation forecasting tools play a key role to address the chal -lenges. A multi-layered feed-forward artificial neural network modeloptimized by a combination of genetic algorithm and particle swarmoptimization algorithm is developed in this work for wind power gen -eration forecasting. The proposed technique is tested based on practicalinformation obtained from Goldwind Smart Microgrid in Beijing. Theperformance of the proposed method is superior to neural networkmodels optimized using GA and PSO separately, as well as the bench-mark persistence approach.