| 用神经网络法预测特大型电网短期负荷的初探 |
| |
| 引用本文： | 孟宪生,余铭正,武寒,郭佳田. 用神经网络法预测特大型电网短期负荷的初探[J]. 华东电力, 2000, 28(6): 4-6 |
| |
| 作者姓名： | 孟宪生 余铭正 武寒 郭佳田 |
| |
| 作者单位： | 1. 上海电力学院,上海,200090
2. 华东电力调度局,上海,200002 |
| |
| 摘 要： | 影响特大型电网短期负荷的因素十分复杂,负荷变化属非平稳的随机过程.用神经网络BP算法并采用变步长、加动量因子以及不等权、半随机初始解等方法,使迭代迅速收敛,减小预测误差、提高学习精度;用目标误差作为结束网络迭代条件;用测点模型矩阵预测不同季节、不同时段的负荷,在短期负荷预测中取得较满意结果.
|
| 关 键 词： | 神经网络 短期负荷预测 特大电网 |
| 文章编号： | 1001-9529(2000)06-0004-03 |
| 修稿时间： | 2000-04-28 |
A Preliminary Investigation on Short-term Load Predication of Extra-large Power Grid Using ANN Method |
| |
| Meng Xiansheng Yu Mingzheng. A Preliminary Investigation on Short-term Load Predication of Extra-large Power Grid Using ANN Method[J]. East China Electric Power, 2000, 28(6): 4-6 |
| |
| Authors: | Meng Xiansheng Yu Mingzheng |
| |
| Affiliation: | Meng Xiansheng Yu Mingzheng (Shanghai Institute of Electric Power Shanghai 200090) Wu han Guo Jiatian (East China Electric Power Dispatching Bureau Shanghai 200002) |
| |
| Abstract: | Factors influencing the extra large power network are very complicated,and variation of load is of a non steady stochastic process.The paper presents the study and application of Artifical Neural Network (ANN) in the filed of short term load forecasting of extralarge power network.The BP(Back Propagation)based on the ANN with methods of variable step,momentum parameter,non uniform weights and semi stochastic initial solutions,which cause a rapid converge of iteration,minimization of predicating error and enhancemant of studying precision.The target error is used as the ending condition of network iteration,The load of different seasons and periods are forecasted with the matrix of models of forecasting points. A satisfactory results of short term load predication has been obtained. |
| |
| Keywords: | artifical neural network forecasting short-term load |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
