基于Elman神经网络的电力负荷预测模型研究
摘要随着我国电力事业的发展，电网管理越来越趋于现代化，电力负荷预测问题的研究 也越来越受到人们的注意，它是研究电力系统规划问题、电力系统经济运行自动化的重 要依据。因此，寻求合适的电力负荷预测方法从而最大限度地提高电力负荷预测精度是 非常必要的。83390
针对传统的电力负荷预测方法，并结合人工神经网络（ANN）在分析非线性问题上 所拥有的优势，本文提出了基于 Elman 神经网络的电力负荷预测模型研究。我们通过 MATLAB 软件进行编程，从输入层、隐含层、输出层出发，设计并建立了基于 Elman 神 经网络的负荷预测模型。本文在预测模型建立之后，利用随机数读取的历史电力负荷数 据，进行训练，从而实现仿真预测。同时我们将预测出的数据与真实数据进行对比，得 出两者之间的误差率，并对其进行分析。
BP 神经网络是实际工程中最常用的电力系统负荷预测方法，它是一种静态神经网 络，在用于预测动态负荷时很容易陷入局部最小点，所以预测精度就很难有实质性的提 高。而本文具有动态递归性的 Elman 神经网络模型，采用的是自适应学习速率的下梯度 下降法，能提高网络的学习速率且有效地抑制网络陷入局部最小点，从而使结果更加有 效，增加可信度，有效地提高预测精度。
毕业论文关键词：电力负荷；负荷预测；BP 神经网络；Elman 神经网络
Abstract With the development of our country electric power enterprise, power management is more and more devoted to be modern,the research of power load forecasting problem is becoming more and more popular, it is the study of power system planning problem and the important basis of automation of electric power system economic operation。Therefore, to seek appropriate power load forecasting method so as to maximize the power load forecasting accuracy is very necessary。
In view of the traditional power load forecasting method, and combined with artificial neural network (ANN) has advantages in analyzing nonlinear problems, this paper puts forward the model of power load forecasting based on Elman neural network research。We through the MATLAB software programming, starting from the input layer, hidden layer and output layer, design and set up based on Elman neural network load forecasting model。Establishment of forecasting model is presented in this paper, the use of random Numbers read the history of the power load data, training, so as to realize the simulation prediction。At the same time, we will predict data compared with real data, it is concluded that the error rate between the two, and to analyze it。
The BP neural network is commonly used in the practical engineering of the power system load forecasting method, it is a kind of static neural networks, when used to predict the dynamic load is easy to fall into local minimum point, so the prediction precision is hard to have a substantial increase。And this article with dynamic recursion Elman neural network model uses the adaptive learning rate under the gradient descent method, which can improve the network's learning rate and effectively restrain network into a local minimum point, which makes the results more effectively, increasing credibility, effectively improving the prediction accuracy。
Key words: power load;Load forecasting;The BP neural network;Elman neural network
目 录
第一章 绪 论 1
1。1 选题的目的和意义 1
1。3 电力负荷预测方法