基于支持向量机的某区域电网电力需求的预测研究

发布时间：2018-06-01 08:29

本文选题：电力负荷预测 + 支持向量机　；参考：《北京交通大学》2014年硕士论文

【摘要】：摘要：电力需求预测是电力系统规划的重要组成部分。本论文主要对电力需求中的月度和年度负荷进行预测。它们的特点是历史数据少,受经济、社会等不确定因素影响较大。准确的负荷预测有利于提高电网运行的安全稳定性,有效地降低发电成本,保证用电需求,增强供电可靠性,从而提高电力系统的经济效益和社会效益。本文介绍了电力系统负荷预测的目的和意义,对国内外负荷预测的现状进行综述。介绍了负荷预测的基本原理,分析了各种方法的优缺点。阐述了电力负荷的分类及特点,给出了电力负荷预测的模型要求和误差指标。同时对支持向量机的文献进行综述,指出支持向量机的优点和存在的问题。并且对影响电力需求的因素做了分析。本文在分析了电力负荷预测特点的基础上,采用支持向量回归机算法对山西电网月度最大电力负荷进行预测。介绍了支持向量机算法的基本原理,建立基于该方法的负荷预测模型,给出基本算法流程图,利用MATLAB进行程序设计,实现上述算法过程。通过预测结果分析并与其他方法进行比较,验证了该智能预测方法的可行性。然后在分析了支持向量回归机的各参数对其性能有很大影响的基础上,结合月度最大电力负荷的特点,提出了利用粒子群优化支持向量负荷预测模型,并通过权衡近期数据法整理月度负荷数据。同时给出了粒子群优化支持向量机模型的原理及流程图。通过实际算例分析,与标准支持向量回归机方法的预测结果进行比较,验证粒子群优化后的支持向量回归机负荷预测模型具有预测精度高、计算量小等优势。同时提出了基于粒子群优化支持向量机负荷预测模型的改进措施,增加了惯性权重因子,并提出了三点平滑法优化数据,使模型更加完善。给出改进后的算法流程图,通过预测山西月度负荷,验证了改进型的粒子群优化支持向量机负荷预测模型的预测精度更高。在前文分析了电力需求影响因素的基础上,针对年度负荷的特点,整理输入数据,利用建立好的模型预测未来年份的最大电力负荷。最后,介绍国内外常用的基于支持向量机的算法程序。分析了网格搜索法优化支持向量回归机的特点,运用基于网格搜索法优化支持向量机的CMSVM软件对山西省电力负荷进行预测研究。并对基于支持向量机的负荷预测所需要注意的关键问题做出总结,并提出建议。
[Abstract]:Abstract: power demand forecasting is an important part of power system planning. This paper mainly forecasts the monthly and annual load in power demand. They are characterized by the lack of historical data, economic, social and other uncertain factors. Accurate load forecasting can improve the safety and stability of power system, reduce the cost of generation, ensure the demand of electricity, enhance the reliability of power supply, and improve the economic and social benefits of power system. This paper introduces the purpose and significance of power system load forecasting, and summarizes the current situation of load forecasting at home and abroad. The basic principle of load forecasting is introduced, and the advantages and disadvantages of various methods are analyzed. This paper expounds the classification and characteristics of power load, and gives the model requirements and error index of power load forecasting. At the same time, the paper summarizes the literature of support vector machine, and points out the advantages and problems of support vector machine. The factors that affect the power demand are also analyzed. Based on the analysis of the characteristics of power load forecasting, the support vector regression algorithm is used to forecast the maximum monthly power load of Shanxi power network. The basic principle of support vector machine (SVM) algorithm is introduced, the load forecasting model based on this method is established, the flow chart of basic algorithm is given, and the program is designed by using MATLAB to realize the above algorithm process. The feasibility of the intelligent prediction method is verified by analyzing the prediction results and comparing with other methods. Then, based on the analysis of the influence of the parameters of support vector regression machine on its performance, combined with the characteristics of the maximum monthly power load, a support vector load forecasting model based on particle swarm optimization is proposed. And collate monthly load data by tradeoff short-term data method. At the same time, the principle and flow chart of particle swarm optimization support vector machine model are given. Compared with the prediction results of the standard support vector regression method, it is verified that the load forecasting model based on particle swarm optimization has the advantages of high forecasting accuracy and less calculation. At the same time, an improved support vector machine load forecasting model based on particle swarm optimization is proposed. The inertia weight factor is added, and the three-point smoothing method is proposed to optimize the data to make the model more perfect. The flow chart of the improved algorithm is given. By forecasting the monthly load in Shanxi, it is verified that the improved particle swarm optimization support vector machine forecasting model has higher forecasting accuracy. Based on the analysis of the influencing factors of power demand, the input data are arranged according to the characteristics of the annual load, and the established model is used to predict the maximum power load in the future year. At last, the algorithm program based on support vector machine is introduced. The characteristics of optimized support vector regression (SVM) based on grid search method are analyzed, and the power load forecasting of Shanxi province is studied by using CMSVM software based on grid search optimization support vector machine (SVM). The key problems of load forecasting based on support vector machine are summarized and some suggestions are put forward.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TM715

【参考文献】