基于BIRCH的分布式光伏系统短期发电功率预测方法

Short-Term Power Generation Power Prediction Method of Distributed Photovoltaic System Based on BIRCH Clustering

当前分布式光伏系统短期发电功率预测结构多设定为目标式,预测范围在实际发电环境下受限,导致平均绝对预测误差增加。为此设计基于利用层次方法的平衡迭代规约和聚类(Balanced Iterative Reducing and Clustering Using Hierarchies,BIRCH)的分布式光伏系统短期发电功率预测方法。首先,明确预测指标,采用多层级的方式设计预测结构;其次,结合BIRCH原理,设计发电功率预测模型;最后,采用梯度回归处理的方式来实现最终预测。测试结果表明,对比传统变分模态分解-麻雀搜索算法-反向传播(Variational Mode Decomposition-Sparrow Search Algorithm-Back Propagation,VMD-SSA-BP)光伏系统短期发电功率预测小组、传统时序动态回归光伏系统短期发电功率预测小组,此次所设计的方法得出的平均绝对预测误差被较好地控制在2.1以下,预测效果更佳,针对性更强,误差可控,具有实际的应用价值。

At present,the short-term generation power prediction structure of distributed photovoltaic system is mostly set to target type,and the prediction range is limited in the actual generation environment,which leads to an increase in the average absolute prediction error.Therefore,the design and verification analysis of short-term generation power prediction method for distributed photovoltaic system based on Balanced Iterative Reducing and Clustering Using Hierarchies(BIRCH)are proposed.Firstly,the forecast index is defined and the forecast structure is designed in a multi-level way.Secondly,combined with BIRCH clustering principle,a generation power prediction model is designed.Finally,gradient regression processing is used to realize the final prediction.The test results show that:compared with the traditional Variational Mode Decomposition-Sparrow Search Algorithm-Back Propagation(VMD-SSA-BP)short-term generation power forecasting group and the traditional time series dynamic regression photovoltaic system short-term generation power forecasting group,the average absolute prediction error obtained by this method is well controlled below 2.1,which shows that the designed short-term generation power forecasting effect is better,more targeted,and the error is controllable,which has practical application value.