基于混沌神经网络理论的小电网短期电力负荷预测

通过对小电网负荷数据的特点分析,将时间序列处理、混沌理论和神经网络理论相结合提出了一种基于混沌神经网络理论的电力负荷预测模型。利用Matlab对实际数据进行了仿真计算。通过实例计算,并和不用相空间重构的神经网络的负荷预测算法的各种误差指标的分析比较说明,利用相空间重构对历史数据序列进行拆分或重构可以提高负荷预测的精度。

Short-term load forecast for small power net based on chaos-artificial neural network theory

According to the summarized features of load data in small power net,a chaos-artificial neural network theory based on load forecast model is presented,which combines time series processing,chaos theory and neural network theory together.Actual data are used for forecast simulation with Matlab.The forecast error indices of neural network with and without phase space reconstruction are compared through computation.Results prove that using phase space reconstruction to split and reconstruct the historical data can improve the precision of load forecast.