长相关随机模型FBM对光伏发电短期预测

Short-Term Forecasting of Photovoltaic Power Generation by the Long Correlation Stochastic Model FBM

针对光伏发电功率具有较强的波动性、间歇性输出,光伏功率预测精度较低,且难于给出具体预测时间长度等问题,提出了一种长相关随机模型分数阶布朗运动(fractional Brownian motion,FBM),用于光伏功率预测。首先,采用重标极差法计算长相关(long-range dependence,LRD)参数-Hurst指数,Hurst指数用于判断光伏功率数据是否满足长相关性,并通过最大李雅普诺夫指数(Lyapunov)计算出模型最大可预测时间尺度;其次,采用随机微分法建立FBM光伏功率预测模型,同时估计FBM预测模型参数值;最后,选取澳大利亚沙漠知识太阳能中心(Desert Knowledge Australia Solar Center,DKASC)、美国国家可再生能源实验室(National Renewable Energy Laboratory,NREL)以及北京国能日新科技有限公司的光伏功率数据集,从不同的地理环境、不同的气候特征、不同的规模大小电站进行验证。仿真结果表明,该模型较传统的Kalman、LSTM模型具有更高的预测精度,可为光伏并网的稳定和安全运行提供更好的理论支持,对电网调度部门具有较高的参考价值。

To address the strong fluctuation and intermittency of PV power output which affects the accuracy of PV power prediction,and makes it difficult to give the prediction time length,this paper proposes a long correlated stochastic model-fractional Brownian motion(FBM)for PV power prediction.Firstly,the long range dependence(LRD)parameter-Hurst index is calculated using the rescaled polarity method,which is used to determine whether the PV power data satisfies the long range dependence,and the maximum predictable time scale of the model is calculated by the maximum Lyapunov exponent.Secondly,the FBM iterative prediction model is established with the stochastic differentiation method and the estimated values of the FBM prediction model parameters are obtained by theoretical derivation.Finally,the photovoltaic power datasets of Desert Knowledge Australia Solar Center(DKASC),the National Renewable Energy Laboratory(NREL),and Beijing Guoneng Rixin Technology Co.,Ltd.are selected to be verified from different geographic environments,different climatic characteristics,and different sizes and sizes of power stations.The simulation results show that the model has higher prediction accuracy than the traditional Kalman and LSTM models,which can provide better theoretical support for the stable and safe operation of the grid-connected PV,and has high reference value for the grid scheduling department.