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

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.