天燃气负荷预测的建模与应用研究  

Modeling and Application Research on Gas Load Forecasting

作　　者：代丽娴[1] 

机构地区：[1]广西梧州学院计算机科学系,广西梧州543002

出　　处：《计算机仿真》2011年第10期180-183,共4页Computer Simulation

摘　　要：研究天燃气负荷预测问题,由于天燃气负荷受人口增多用量增大及天气、季节、节假日等因素影响,具有周期性和随机性的变化规律,形成一种非线性特性,传统预测方法无法进行准确的预测,预测精度比较低。为了提高天燃气负荷的预测精度,提出一种基于RBF神经网络的天燃气负荷预测方法。首先对天燃气负荷历史数据进行预处理,剔掉一些异常的数据,然后将数据输入到RBF神经网络中学习,采用遗传算法对RBF神经网络参数进行优化,从而建立最优的天燃气负荷预测模型。采用某企业的天燃气负荷数据对模型的性能进行验证,实验结果表明,相对于传统预测方法,RBF神经网络提高了天燃气负荷预测精度,是一种较好的天燃气预测方法。Research natural gas load forecasting problems.As gas load is affected by the weather,seasons,festival and other factors,the changing rule is of randomicity and periodicity,the traditional forecasting methods cannot accurately forecast it,and prediction accuracy is quite low.In order to improve the prediction accuracy of natural gas load,this paper proposed a gas load forecasting method based on RBF neural network.Firstly,the historical data of natural gas load is pretreated to picking off some abnormal data,then the data are used as inputs of the RBF neural network for training and the RBF neural network parameters is optimized by genetic algorithm.Then the optimal gas load forecasting model is established.The model performance is verified by som gas load data from an enterprise,and experimental results show that compared with the traditional forecasting methods,the RBF neural network has improved the natural gas load forecasting accuracy and is a better forecasting method of natural gas.

关 键 词：天燃气 负荷预测 人工神经网络 

分 类 号：TU996[建筑科学—供热、供燃气、通风及空调工程]