电力系统短期负荷预测.docx

实用文档 电力系统短期负荷预测 POWER SYSTEM SHORT-TERM LOAD FORECASTING 专 业：电气工程及其自动化 姓 名： 指导教师姓名： 申请学位级别：学 士 论文提交日期：二零一六年十二月 学位授予单位： 天津科技大学  摘 要 电力系统负荷预测是电力生产部门的重要工作之一。准确的负荷预测，可以合理安排机组启停，减少备用容量，合理安排检修计划及降低发电成本等。准确的预测，特别是短期负荷预测对提高电力经营主体的运行效益有直接的作用，对电力系统控制、运行和计划都有重要意义。因此，针对不同场合需要寻求有效的负荷预测方法来提高预测精度。本文采用神经网络方法对电力系统短期负荷进行预测。本文主要介绍了电力负荷预测的主要方法和神经网络的原理、结构，分析了反向传播算法，建立三层人工神经网络模型进行负荷预测，并编写相关程序。与此同时采用最小二乘法进行对比，通过对最小二乘法多项式拟合原理的学习，建立模型编写相关程序。通过算例对两种模型绝对误差、相对误差、拟合精度进行分析，同时比较它们训练时间，得出标准BP神经网络具有更好的精度优势但训练速度较慢。最后针对标准BP神经网络训练速度慢、容易陷入局部最小值等缺点，对标准BP神经网络程序运用附加动量法进行修改，分析改进后网络的优点。 关键词：短期负荷预测 标准BP神经网络 最小二乘法 附加动量法  ABSTRACT Power system load forecasting is one of the most important work of the electricity production sector. The accurate load forecasting can arrange unit start-stop, reduce the spare capacity, reasonable arrangement of the maintenance plan and reduce power cost, etc. It has a direct effect on the running efficiency of the power management entities and also has the important meaning in the power system control, operation and planning. So it is important to find effective method to enhance forecast precision for different occasions. In this paper the neural network is used for the short-term load forecasting of the power system. This article introduces the method of the power load forecasting and the principles, structure, back-propagation algorithm of the neural network. Then the three-layer artificial neural network model is created for load forecasting and the program is written. At the same time, the least square method is used for comparing. By learning the polynomial fitting principle of the square method, the model is created and the program is written. Through comparing the absolute error, the relative error, the fitting precision and their training time of the two models, the BP neural network is proved to have better accuracy but slower training speed. Due to the standard BP neural network has slower train