基于信息熵和变精度粗糙集优化的支持向量机降温负荷预测方法

发布时间：2018-01-07 20:40

本文关键词：基于信息熵和变精度粗糙集优化的支持向量机降温负荷预测方法　出处：《电网技术》2017年01期 　论文类型：期刊论文

【摘要】：降温负荷持续增长已成为中国南方夏季最大负荷屡创新高的重要原因。提出了一种基于信息熵和变精度粗糙集优化的不确定支持向量机方法,用于中长期降温负荷预测。方法通过挖掘数据中的相互关系去除冗余信息,从输入属性变量集中寻找核心变量。该方法利用基于信息熵改进的变精度粗糙集对支持向量机的条件属性进行约简,得到最小决策表,并将该最小决策表中对应的变量作为支持向量机预测模型的输入属性变量,进行年最大降温负荷预测。且随着预测年份的推移,该支持向量机预测模型的输入属性变量亦将随之滚动更新,能够为电网规划与运行人员提供不同预测时期降温负荷预测需重点关注的影响因子。最后,利用广东省实际数据对广东电网"十二五"和"十三五"年最大降温负荷进行预测,结果表明,所提的预测方法预测效果良好,预测精度稳定,对于中长期预测过程中的各种不确定因素的影响具有较好的鲁棒性,真正实现了中长期降温负荷的动态预测。
[Abstract]:The continuous increase of cooling load has become an important reason for the maximum load in southern China to reach new highs in summer. An uncertain support vector machine (SVM) method based on information entropy and variable precision rough set optimization is proposed. It is used for medium and long term cooling load forecasting. Methods the redundant information is removed by mining the interrelation in the data. This method uses variable precision rough set based on information entropy to reduce the conditional attributes of support vector machine and obtains the minimum decision table. The corresponding variables in the minimum decision table are used as the input attribute variables of the SVM forecasting model to forecast the annual maximum cooling load, and with the passage of the predicted year. The input attribute variables of the SVM forecasting model will be updated with it, which can provide the power network planning and operation personnel with the influence factors which need to be paid more attention to in different forecasting periods of cooling load forecasting. Finally. Based on the actual data of Guangdong Province, the maximum cooling load of Guangdong Power Grid in the 12th Five-Year Plan and the 13th Five-Year Plan is forecasted. The results show that the proposed forecasting method has good prediction effect and stable prediction precision. It has good robustness to the influence of various uncertain factors in the process of medium and long term forecasting, and realizes the dynamic forecasting of medium and long term cooling load.
【作者单位】：华南理工大学电力学院;
【基金】：国家自然科学基金青年基金资助项目(50907023) 中国南方电网有限责任公司科技项目(K-GD2012-006)~~
【分类号】：TM715
【正文快照】： 并将该最小决策表中对应的变量作为支持向量机预测模型的输入属性变量,进行年最大降温负荷预测。且随着预测年份的推移,该支持向量机预测模型的输入属性变量亦将随之滚动更新,能够为电网规划与运行人员提供不同预测时期降温负荷预测需重点关注的影响因子。最后,利用广东省实际

【相似文献】