摘要
风力发电具有波动性、随机性和间歇性,因此准确预测风电场的日有功功率对风电场与电力系统的稳定运行具有重要的意义。利用C-C法对风电场的日有功功率时间序列进行相空间重构,并通过计算其最大Lyapunov指数,验证了此功率时间序列具有混沌属性。在此基础上,用相空间重构建立了RBF神经网络和最小二乘支持向量机预测模型,对预测结果采用协方差优选确定权重,进行组合预测。通过对甘肃省酒泉地区某风电场的实测数据进行仿真,证明了该组合模型的有效性和可行性,并有效提高了预测精度。
Wind power generation possesses the inherent features of volatility,randomness and intermittence,thus it is significant for wind farm and stable operation of power system to accurately predict the daily active power of the wind farm.Via the C-C algorithm,this paper reconstructed the phase space for the daily active power time series of wind farm and calculated the largest Lyapunov index,so as to prove that the power time series has the chaos characteristics.On the basis of chaotic theory,the phase space reconstruction was utilized in the RBF neural network and the least squares support vector machine,and the prediction results＇s weghts are calculated by covariance optimization for combined prediction.The simulation is derived from the data collected from the wind farm which is located in Jiuquan,Gansu Province,then the effectiveness and feasibility of the combined model has been proved by simulation results; furthermore,the prediction accuracy has been improved effectively.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2014年第6期20-24,36,共6页
Proceedings of the CSU-EPSA
基金
国家自然科学基金资助项目(51267012)
甘肃省电网公司科技项目(2010406029)
甘肃省高等学校基本科研业务费专项资金项目(1103ZTC141)
关键词
风力发电
相空间重构
RBF神经网络
最小二乘支持向量机
组合预测
wind power generation
phase space reconstruction
RBF neural network
least square support vector machine
combination prediction
