摘要
传统灰色风速预测模型累加处理时不能预测突变风速,使风电功率预测误差过大。采用数值逼近算法对传统灰色GM(1,1)预测模型进行优化改进,以优化的灰色GM(1,1)预测模型对未来时段风速进行预测,突变风速预测误差降低了34.3%。再将优化风速预测模型和时间序列动态神经网络相结合,构建出风电功率预测模型。应用该模型对酒泉地区某风电场现场数据进行仿真测试,预测效果可信度大于93%。
The traditional gray predictor model can not predict mutation of wind speed, leading wind power forecasting to excessive errors. The approximation algorithm was used to improve gray predictor model GM（1,1） for predicting the wind speed of future periods. The simulation results show that forecast mutation of wind speed errors decrease by 34.3%. The new wind speed forecast model is combined with time series based on dynamic neural network to predictor wind power. This model was tested at Jiuquan wind farm, and forecast credibility was greater than 93%.
出处
《电源技术》
CAS
CSCD
北大核心
2013年第4期614-616,638,共4页
Chinese Journal of Power Sources
基金
甘肃省科技厅科技支甘项目(1011JKCA172)
兰州市科技计划项目(2011-1-106)
关键词
灰色优化模型
短期风速预测
时间序列动态神经网络
风电功率预测
gray optimization model
short-term wind speed forecasting
time series dynamic neural network
windpower forecasting
