基于混沌分析的BP神经网络模型及其在负荷预测中的应用

BP Neural Networks Model Based on Chaotic Analysis and Its Application on Power Load Forecasting

结合混沌分析理论和BP神经网络,提出在混沌相空间建立BP神经网络模型。运用混沌方法构成训练样本及确定神经网络的网络结构,用神经网络拟合相空间相点演化的非线性关系。并利用该模型对具混沌特性的电力系统日负荷时间序列进行短期预测,对比了标准BP网络模型和混沌线性回归模型的预测结果,结果表明基于混沌分析的BP神经网络模型的预测精度较高。

Combined with the chaos theory and artificial neural networks, this paper presents a BP neural networks model based on chaotic analysis in the phase space. Training data construction and networks structure are determined by chaotic phase space, and nonlinear relationship of phase points is established by BP neural networks. As an example, the new model is applied on short term forecasting of daily power load with chaotic characteristics. Compared to the standard BP neural networks and chaotic linear regressive model, the results show higher precision of BP neural networks model based on chaotic analysis in training and forecasting.