节能形势预警理论与调控机制研究

Study on Energy-saving Early Warning Theories and Control Measures

我国从2010年起定期发布各地区节能目标完成情况"晴雨表",形象地展示各地区节能目标完成进度情况。节能"晴雨表"的背后,实则是依靠一套复杂的节能形势预测预警理论与方法来支撑。以国内某市为例,采用支持向量机模型,在地区(市)、区(县)、行业三个层面建立了包括基础数据(先行自变量指标)输入和预警等级输出两大部分的节能形势预警系统,系统要预测的直接目标是复合指标单位经济增加值能耗下降率和能耗总量增速。本着"前松后紧"的原则,依据每月节能完成情况与年度目标的差距,在10%上下浮动区间内划分三级三色预警等级。完整的节能形势预警调控程序包括预警监测、专家会商与预警启动、预警调控方案实施及预警解除4个环节,同时本着有关原则,在工业、三产、公共机构等重点领域建立了节能预警调控措施库。今后仍需进一步扩大模型应用数据样本量,不断筛选更合适的先行预测指标,拟合更精确的预测模型;预警等级划分的合理性和可行性尚需广泛征求意见;节能形势预警应尽快赶上"互联网+"的潮流,建立能耗即时预测预警和互联共享平台。

Since 2010,China has regularly released reports on the completion of energy-saving targets in each region,showing the completion progress of energy-saving targets.Actually,this relies on complex theories and methods for the prediction and early warning of energy-saving situation.Tak ing a city in China as an ex- ample ,this paper establishes an early warning system for energy-saving situation ,which consists of basic data （leading indicators） input and warning grade output,at region （city）,district （county） and industry levels based on the model of support vector machine （SVM）.The system aims to predict the decline rate of energy consump- tion per unit of GDP and the total energy consumption growth.Following the principle of ＇earlier-loose and later-tight＇ ,the early warning of energy-saving situation is divided into three grades and three colors with 10% up and down floating according to the gap between monthly completion and annual targets.A complete early warning control program consists of early warning monitorlng,expert consuhation and warning start, im- plementation of control measures and lifting of warning.Meanwhile,control measure libraries are established for important fields related to secondary industries,tertiary industries and public institutions.In the future,it is necessary to further expand the sample size of the applied data of the model for continuously selecting more accurate indicators and models for early warning prediction,the rationality and feasibility of warning grade di- vision still need extensive consultation,and the early warning of energy-saving situation should catch up with the trend of Internet Plus as soon as possible and establish a real-time energy consumption predication and early warning Intemet sharing platform.