城市燃气短期负荷预测的神经网络等维新息模型

A Equal-dimension and New-information Model of Urban Gas Load Forecasting in Application of Neural Network

应用人工神经网络理论和灰色预测理论中的等维新息建模思想 ,建立了既反映其时间序列的周期性增长趋势 ,又包括天气、气温等非线性影响因素在内的短期负荷预测的BP神经网络等维新息模型。通过改进BP神经网络 ,对哈尔滨市燃气管网系统的小时燃气用量进行了预测 ,所建立的模型不仅有较高的收敛速度和精度 ,同时也具有较强的适应性和灵活性 ,可应用于工程实践。

According to characteristics of short-term load changes for urban gas load, this paper sets up a equal dimension and new information model for urban gas load based on the neural network and gray forecasting theory. This new model reflected both time sequence periodical trend and nonlinear affection. Through improving algorithm of the artificial neural network, we forecasts gas load of the harbin gas network. The result of forecasting system showed that it is feasible and efficient and could applied in practice.