文章摘要
徐诗鸿,张宏志,林湘宁,等.基于改进评价指标的波动性负荷短期区间预测[J].电力系统自动化. DOI: 10.7500/AEPS20190123002.
XU shihong,ZHANG hongzhi,LIN xiangning, et al.Improved Evaluation Index Based Short-term Interval Prediction of Fluctuation Load[J].Automation of Electric Power Systems. DOI: 10.7500/AEPS20190123002.
基于改进评价指标的波动性负荷短期区间预测
Improved Evaluation Index Based Short-term Interval Prediction of Fluctuation Load
DOI:10.7500/AEPS20190123002
关键词: 波动性负荷  区间预测  边界估计  神经网络  粒子群优化算法
KeyWords: Fluctuating load  interval prediction  boundary estimation  neural network  particle swarm optimization algorithm
上网日期:2019-11-07
基金项目:国家自然科学基金重点项目
作者单位E-mail
徐诗鸿 华中科技大学电气与电子工程学院强电磁工程与新技术国家重点实验室 375856335@qq.com 
张宏志 广东电网有限责任公司管理科学研究院 glandchang@gmail.com 
林湘宁 华中科技大学电气与电子工程学院强电磁工程与新技术国家重点实验室 linxiangning@hotmail.com 
李正天 华中科技大学电气与电子工程学院强电磁工程与新技术国家重点实验室 453874933@qq.com 
卓毅鑫 广西电网公司电力调度控制中心 375856335@qq.com 
汪致洵 华中科技大学电气与电子工程学院强电磁工程与新技术国家重点实验室 359041265@qq.com 
随权 华中科技大学电气与电子工程学院强电磁工程与新技术国家重点实验室 1601153437@qq.com 
摘要:
      针对波动性较大、不确定性较强的负荷变化规律,文中提出一种基于边界估计理论的区间预测方法,从区间宽度和累计误差两个角度对现有区间预测评价指标做出改进,提高了预测结果的合理性;在此基础上,根据对各评价指标重要性的权衡,建立了区间预测综合评价指标,并利用神经网络构建了区间预测模型,以综合评价指标最优为目标,采用粒子群算法对网络结构参数进行训练优化,从而取得理想的波动性负荷区间预测效果。仿真中通过对某波动性较强的历史负荷数据进行预测分析,并与传统的点预测和区间预测方法进行对比,验证了所提方法的有效性和优越性。
Abstract:
      Aiming at the load with large fluctuation and uncertainty, this paper implements a prediction intervals method based on the lower upper bound theory, improving existing forecast evaluation index from two aspects of interval width and cumulative error, which enhances the reasonableness of prediction results. On the basis, weighing the importance of each evaluation index, the comprehensive evaluation index for interval prediction is established, and the interval prediction model is constructed by using neural network. Aiming at the optimization of the comprehensive evaluation index, the particle swarm optimization algorithm is used to train and optimize the structure parameters, so as to achieve ideal effect of interval prediction for fluctuating load. The historical load data with strong uncertainty is used to validate the proposed method. Comparing with the traditional point and interval forecasting methods, the results and analysis of the improved interval prediction verified the effectiveness and superiority of the method.
查看全文   查看附录   查看/发表评论  下载PDF阅读器