摘要
分析了北方城市的日用气负荷特点,分别介绍了小波理论和神经网络理论等技术,并对日用气负荷中的关键技术如网络结构及输入神经元的确定、小波基函数的选择及分解尺度的确定、激励函数的选择、数据的归一化等进行了详细分析,提出了适合我国北方城市的日用气负荷预测模型.
This thesis mainly analyses the characteristics of daily gas load in northern cities and introduces the technologies of wavelet theory and neural network respectively. The key technologies of the daily gas load, such as the determination of network struc- ture and the input neural neurons, the selection of wavelet basis function, the confirmation of decomposition scale, the choice of trans- fer function and the normalization of data are analyzed in detail. Then the forecasting model is proposed of daily gas load which is ap- propriate for the northern cities in China
出处
《河南工程学院学报（自然科学版）》
2013年第2期28-33,共6页
Journal of Henan University of Engineering：Natural Science Edition
基金
河南省教育厅科学技术研究重点项目(13A520016)
关键词
日用气负荷
多分辨小波
神经网络
预测模型
daily gas load
multi-resolution wavelet
neural network
forecasting model
