含高比例分布式光伏的母线负荷预测方法

高比例分布式光伏的大规模接入对母线辖区的负荷预测产生了较大影响,导致母线辖区内负荷偏离用户用电负荷的真实状况。文章考虑了高比例分布式电源对负荷形态的影响,提出了基于互信息与混合模型的母线辖区内负荷预测模型,对分布式电源相关输入因子采用互信息系数进行相关性分析,并通过由XGBoost算法与极限学习机算法组成的混合模型对数据进行训练。最后,使用某地母线辖区内负荷数据进行实例验证,结果表明,考虑分布式电源接入后的母线辖区负荷预测精度高于常规预测方法,文中所建立的预测模型具有良好的预测精度。

Bus load forecasting considering large-scale distributed PV interconnection

Large-scale interconnection of distributed photovoltaic has a great impact on load forecasting in bus,which means the load curve in the substation-area can not reflect the real load curve of customer.In this paper,given that the distributed PV integration influence to load curve,the load forecasting of bus based on hybrid model and mutual information is proposed.Mutual information is used to analyze photovoltaic correlation input factors.Hybrid model including XGBoost and extreme learning machine is used to train by input data.At last,the case verification is carried out on actual data of bus-area,the results show that the prediction model established in this paper has good prediction accuracy and stability.