数据挖掘技术在火力发电厂锅炉管壁超温分析中的应用 (1)论文.pdfVIP

数据挖掘技术在火力发电厂锅炉 管壁超温分析中的应用 王 锋 华能南京电厂，江苏南京六合区凤凰南路 98#，210035 Studies of Data Mining on the overheating of Boiler Tubes in the Thermo-Power Plant wangfeng HuaNeng Nanjing Power Plant,No.98,FengHuang Road(South),LiuHe District ,Nanjing Jiangsu 210035 ABSTRACT ：The overheating of boiler tubes is a universal 策树分析推测出了有关锅炉管壁超温点等目标。 problem in current thermo-power plant. It affects boiler ’s 关键词：数据挖掘；关联分析；决策树；神经网络；管 reliability and safety property and enterprise economic profit 壁超温 seriously. This paper takes the 2 boilers of HuaNeng Nanjing Power Plant as the study object. by collecting, gathering, 1 前言 pretreating the data, establish the middle analytic database of the boiler tubes, then select three kinds of the method of Data 目前随着火力发电厂锅炉参数及容量的日益增 Mining—— Back Propagation Neural Network, Association 大，锅炉的结构的日趋复杂，其运行的可靠性越来越 Rules, Dicision Tree to mine running data of the boiler ,finally, 大地影响着电厂整体的经济性和安全性，特别是锅炉 multipurpose are presented ： one is to estimate some “四管”（过热器、再热器、省煤器和水冷壁）爆漏， temperature point of the boiler tubes by the Back Propagation 因其引发的事故率高，对电厂安全、电量和经济技术 Neural Network, two is to search the related factors that 指标影响大，是长期困扰火力发电厂生产的一大难 influence the overheating of boiler tubes by Association Rules, 题，因此，成为电力生产严防的二十五项重大事故之 and the other is to analysis and calculate the point of the 一，一直备受重视。华能南京电厂自 1994 年投产至 overheating of boiler tubes by Decision Tree, and so on. 今也曾多次发生锅炉管壁泄漏，特别是从 2000 年 5 月至 9 月间，#1 炉先后发生 5 次因锅炉管壁长期超温 KEY WORD: Data Mining; Association Analysis; Decision