摘要
电力系统负荷预测通过对历史数据分析,预测未来需求,利用经典的Kohonen网络、Elman神经网络和粒子群优化算法建立级联网络预测模型,为了对电力系统短期精确预测,提出了处理非线性问题和解决负荷预测问题。对级联网络预测模型不但能够综合各种单一预测模型的优点,而且能够随时间的推移使结构不断变化,可以减少负荷预测的工作量。用三种神经网络模型进行短期电力负荷预测的仿真结果比较,验证了级联网络预测算法的有效性和良好的应用前景。
In this paper, a cascade network forecasting model is established using the classical Kohonen network, Elman network and the particle swarm optimization algorithm, which can solve the problems of non-linear and load forecasting. Not only can the cascade network model sum up the merits of kinds of single forecasting models, but also it can change the interior configuration, so it tallies with the character of electrical load well and reduces the workload of load forecasting. At the end of the paper, the forecasting results of three network models are compar, and the result shows that the cascade network forecasting model is very effective and has a good prospect.
出处
《计算机仿真》
CSCD
北大核心
2011年第1期311-314,共4页
Computer Simulation
基金
安徽省教育厅自然科学研究项目(KJ2009B035Z)
安徽工程科技学院青年基金(2008YQ024zd)
安徽省高校省级自然科学基金项目(KJ2008B204)
关键词
级联网络
短期负荷预测
粒子群优化
仿真
Cascaded network
Short-term load forecasting
Particle swarm optimization
Simulation
