基于改进型Elman神经网络的电力负荷预测

发布时间：2018-04-05 03:29

本文选题：电力系统　切入点：负荷预测　出处：《西安理工大学》2017年硕士论文

【摘要】：为了保证电力系统运行的安全性和可靠性,电力调度部门有必要对未来的电力负荷进行预测。人工神经网络具有自适应性、自学习性、容错能力较强的优点,因此被运用于电力负荷预测领域,其中Elman神经网络在电力负荷预测中的应用较为广泛。首先,本文介绍了负荷预测的基本概念和研究意义,阐述了负荷预测以及Elman神经网络的国内外研究现状,并详细分析了负荷预测的误差指标以及产生误差的原因,为后续的改进工作奠定了基础。其次,对Elman神经网络的基本原理、网络结构、预测流程、学习过程和学习算法进行了重点研究。针对Elman神经网络动态信息处理能力不足、Sigmoid函数影响收敛速度、传统算法在学习过程中会出现振荡的问题,本文从Elman神经网络的网络结构、激励函数和学习算法三个方面进行改进。以安徽省某市电网的历史负荷数据为例,运用MATLAB软件进行仿真预测。结果表明,改进型Elman神经网络模型进行负荷预测取得了更好的效果,训练速度有所提高,负荷的预测值和实际值更为接近,绝对误差和相对误差都有所降低。最后,为了进一步提高负荷预测的精度,本文将小波分解所具有的良好的时频局部化特性与改进的Elman神经网络共同运用到短期负荷预测中,仍然以安徽省某市电网的历史负荷数据为例,运用MATLAB软件进行仿真观测。结果发明,预测精度得到了进-步提高。
[Abstract]:In order to ensure the safety and reliability of power system operation, power dispatching departments need to forecast the power load in the future. Artificial neural network is adaptive, self-learning and fault-tolerant ability strong advantages, so it is used in power load forecasting, the Elman neural network has been widely used in power load forecasting. Firstly, this paper introduces the basic concept and research significance of load forecasting, discusses the current research situation of load forecasting and Elman neural network at home and abroad, and analyzes the causes of error index of load forecasting errors and, laid the foundation for the subsequent improvement work. Secondly, the basic principle of Elman neural network the structure, flow prediction, learning process and learning algorithm are studied. Based on Elman neural network dynamic information processing capability, Sigmoid function influence The convergence speed, the traditional algorithm will appear oscillation problems in the learning process, this article from the network structure of Elman neural network, improve the three aspects of incentive function and the learning algorithm. The historical load data of Anhui province power grid as an example, the simulation predicted by MATLAB software. The results show that the improved Elman neural network model load forecasting has achieved better effect, improve the training speed, load forecast value and the actual value is close to the absolute error and relative error are reduced. Finally, in order to further improve the accuracy of load forecasting, this paper uses Elman neural network wavelet decomposition has good time-frequency localization characteristics and improved common use to short-term load forecasting, is still in the historical load data of Anhui province power grid as an example, by using MATLAB software. The simulation observation invention, prediction accuracy Get the advance step.

【学位授予单位】：西安理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM715

【参考文献】