基于RBF神经网络的换热管污垢热阻预测  

Forecast of Fouling Resistance of Heat Exchange Pipe Based on RBF Neural Network

作　　者：彭珍[1] 张充[2] 吴立锋[3] 杨善让[4] 

机构地区：[1]华北科技学院计算机系,北京101601 [2]河北大学数学与计算机学院,河北保定071002 [3]北京邮电大学电子工程学院,北京101601 [4]东北电力大学节能与测控技术研究中心,吉林吉林132012

出　　处：《煤气与热力》2008年第6期26-28,共3页Gas & Heat

摘　　要：建立了换热管污垢热阻监测实验系统,采用偏相关系数方法,筛选出与换热管污垢热阻形成关系密切的因素。采用 RBF 神经网络法,对换热管污垢热阻的变化趋势进行了预测,预测值与实验得到的实测值基本吻合,误差小。The experimental system for monitoring fouling resistance of heat exchange pipe is established. The factors forming a close relation with fouling resistance of heat exchange pipe are screened by using partial correlation coefficient method. The variation trend of fouling resistance of heat exchange pipe is forecasted by RBF neural network. The forecasted value basically accords with the measured value obtained in experiments, with small error.

关 键 词：污垢热阻 换热管 偏相关系数 RBF神经网络 

分 类 号：TU995[建筑科学—供热、供燃气、通风及空调工程]