短期风电功率预测误差及出力波动的概率建模

PROBABILISTIC MODELING OF SHORT-TERM WIND POWER PREDICTION ERRORS AND OUTPUT FLUCTUATIONS

准确刻画短期风电功率预测误差以及区域风电出力波动特征是解决大规模不确定性能源并网运行难题的基础。为准确表征风电出力波动与预测误差及气象误差的关联关系,建立高斯混合分布概率模型及利用其与观测曲线的误差构造云模型,然后建立基于正态云与混合高斯分布耦合的概率分布模型,最后采用多种概率密度分布模型对冀北地区单风电场功率预测误差、集群风电功率预测误差、气象预测误差以及不同功率波动范围的预测误差和与其对应的气象预测误差的关联关系进行统计分析。算例结果表明,所提模型拟合效果最优,从而验证了基于正态云与混合高斯分布耦合的概率模型的有效性。

Accurately depict the short-term wind power prediction error and the characteristics of regional wind power output fluctuation is solved,the basis of large-scale parallel operation problem of uncertainty in energy for accurate characterization of wind power output fluctuation and the prediction error and the error of meteorological correlation,Gaussian mixture distribution probability model is set up and use the cloud model and the error of the observation curve structures,Then based on the coupling of the normal cloud and mixture Gaussian distribution probability distribution model,finally using a variety of probability density distribution model of single wind power prediction error,the cluster wind power prediction errors,the region's weather errors as well as different power range of northern Hebei Province weather prediction error of the prediction error and the corresponding statistical analysis of correlation.The simulation results show that the proposed model has the best fitting effect,which verifies the effectiveness of the probabilistic model based on the coupling of normal cloud and mixed Gaussian distribution.