摘要
传统风速预测模型未考虑尾流效应对风电场输出功率的影响,使风电功率预测误差过大。在尾流效应风速模型的基础上提出尾流效应影响因子矩阵,对整个风电场机组输入风速进行修正。然后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,预测效果较好。
The traditional wind speed forecasting model without considering wake effect on the influence of wind power output makes the wind power prediction error too large. Wake effect influence factor matrix was proposed based on considering wake effect of wind speed model. The traditional gray GM（1,1） forecasting model was optimized and improved by using numerical approximation algorithm, and this model was used to forecast wind speed of next time, and then wind power prediction model was built by combining the optimization of the wind speed forecasting model and time series dynamic neural network. Simulation testing was made using the model to the Jiuquan area of a wind farm field data. The results show that mutation of wind speed prediction error is reduced by 34.3% after input wind speed of the whole wind farm unit is corrected, at the same time power prediction values and actual values of output error has been reduced, and the prediction effect is better.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2014年第7期1467-1471,共5页
Journal of System Simulation
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
尾流效应
影响因子矩阵
灰色优化模型
风速预测
时间序列动态神经网络
风电功率预测
wake effect
influence factor matrix
gray optimization model
wind speed forecasting
time series dynamic neural network
wind power forecasting
