A Local Quadratic Wavelet Neural Network Method for Predicting Dispersed Wind Power Generation
- 310 Downloads
In this paper, a new local quadratic wavelet neural network (LQWNN) method is proposed for short-term wind power generation prediction in power system. Due to the nonlinear structure of its local quadratic model (LQM), this method can effectively simulate the nonlinear behavior of wind power generation. In the time domain and frequency domain, wavelet function sets the effective region of LQMs in LQWNN, thus improving the learning efficiency and structural transparency of the overall model. The proposed LQWNN method employs a simple and effective particle swarm optimization (PSO) algorithm and optimizes its parameters. Through two real cases studies, namely the wind power forecast of the Irish power grid and the Jiuquan wind power plant in Gansu province, the performance of the proposed wind power forecasting method is evaluated and compared with other methods. The results show that the LQWNN method proposed in this paper is very effective in predicting wind power generation and has a good application prospect.
KeywordsNeural networks Local quadratic wavelets Particle swarm optimization Power systems Wind power forecasting
Work in this paper is supported by “Science and Technology Project of China Southern Power Grid (ZBKJXM20180015)”.
- 4.Catalao J, Pousinho H, Mendes V (2011) Hybrid intelligent approach for short-term wind power forecasting in Portugal. IET Renew Power Gener 5(3):251–257Google Scholar
- 8.Chen YH et al (2000) Evolving wavelet neural networks for system identification. Proc Int Conf Electr Eng 279–282Google Scholar