基于 ANFIS 模型的光伏电站辐照度超短期预测

摘要

Solar irradiation is the most important meteorological factor for power generation of a photovoltaic (PV)station.A new ultra-short term irradiation prediction method based on adaptive neural-fuzzy inference system (ANFIS)is proposed.Firstly,extraterrestrial irradiance theoretic value is used to normalize the observed values of solar irradiation.Then input-output data pairs are formed and are utilized for ANFIS modeling.Therein,subtract clustering is used to identify the initial rule number and parameters of fuzzy model,back-propagation algorithm and least square method are used to optimized model parameters,then the future four hours irradiation prediction can be achieved by amulti-step iterative process.Undoubtedly,the experimental results verify that the method has good prediction accuracy.This method provides a new applicable solution for ultra-short term power prediction of PV stations.%辐照度是影响光伏（PV）电站发电量最重要的气象因素。提出了一种新的基于自适应模糊推理系统（ANFIS）的辐照度超短期预测方法。首先，利用地外辐照度理论值来对历史观测辐照度时间序列值进行归一化处理，形成输入输出样本对后利用 ANFIS 模型建模；然后，采用减法聚类确定ANFIS 模型的规则数和初始参数，并采用反向传播算法和最小二乘法优化模糊模型参数；最后，通过循环预测法实现对未来4 h 的辐照度预测。基于 MATLAB 的实验结果验证了该方法具备良好的预测精度，从而为发电功率的超短期预测提供了可行的解决途径。