基于主成分分析和PLS算法的弧线管污垢特性研究

Study on Fouling Characteristics of Arc Line Tube Based on PCA and PLS

搭建了弧线管污垢特性实验平台,用于测量相应的污垢参数。基于主成分分析理论和偏最小二乘算法(PLS)建立了弧线管的污垢特性方程。该方程以弧线管的出入口温度、壁温及流速等参数作为模型的自变量,以污垢热阻值作为因变量。利用交叉验证原则以提取最佳主成分个数。通过对预测方程的检验,结果表明:所建5自变量预测方程能较好地实现对弧线管污垢特性的预测,而且明显优于6自变量模型。

The fouling experimental platform for arc line tube was built to measure fouling parameters. Based on both principal component analysis( PCA) and partial least squares( PLS) algorithm,the fouling equation was built which has arc line tube's inlet and outlet temperature,wall temperature and flow rate taken as the model's independent variables,and the fouling resistance taken as the dependent variable,as well as the cross validation principle used to extract the best number of principal components. Testing the prediction equation shows that 5-variable equation outperforms 6-independent variable model in predicting arc line tube 's fouling characteristics.