摘要
科学、准确的化纤能耗预测对化纤行业的健康发展,乃至对整个国民经济的发展均有十分重要的意义.本文根据我国2001年～2011年对苯二甲酸(PTA)、己内酰胺(CPL)、己二酸(AA)、丙烯腈(AN)、浆粕(CP)的消费量和化纤行业能源消耗量的历史数据,利用PASW/SPSS Statistics软件建立多元线性回归和多层感知器神经网络模型.2种模型预测结果均显示,至2015年,化纤行业吨产品能耗将比2010年下降25%以上,化纤行业能源消耗在1 750万吨标煤以下.
Scientific and accurate prediction of energy consumption has a significant meaning to the healthy development of chemical fiber industry and even to the entire national economy. In this article, a prediction research is made with multiple regression and multi-layer perceptron networks using PASW/SPSS Statistics based on the energy consumption data of chemical fiber industry and the consumption data of PTA, CPL, AA, AN and AN from 2001 to 2011. Both the models show that the energy consumption per ton chemical fiber in 2015 will drop over 25% comparing to the data of 2010. Furthermore, the total energy consumption in chemical fiber industry in 2015 is predicted to be below 17.5 million tons of standard coal.
出处
《北京服装学院学报（自然科学版）》
CAS
2012年第4期59-67,共9页
Journal of Beijing Institute of Fashion Technology：Natural Science Edition
关键词
化纤
能耗
多元回归
神经网络
chemical fiber
energy consumption
multiple regression
neural network
