基于人工神经网络的中长期负荷预测算法

Mid and Long-term Load Forecast Algorithm Based on Artificial Neural Network

当前中长负荷预测的大部分方法都衍生于传统的线形统计理论,难以解决复杂的非线性问题。文中结合BP人工神经网络技术,利用人工神经网络所具有的非线性映射和函数逼近功能对中长期电力负荷进行了研究,提出了一种中长期电力负荷预测的思路。并利用北京市的实际数字对未来若干年的用电量进行了预测,实验结果表明,该算法具有较好的准确性和可行性。

Most mid and long-term load forecast methods are derived from linearstatistics, which is weak in solving complex non-linear problems. This article advanced an approach to use the algorithm based on back-propagation neural network and taking benefit from its special features, to solve the problem of mid and long-term load forecast. A forecast is also held in this article to predict the power usage of future years by using figures from Beijing in past years.The result shows the approach has good veracity and feasibility.