Application of Artificial Neural Network and multiple linear regression for modeling and sensitivity analysis of a stripper column

There are many ways to produce the hydrogen. A way is through steam reforming. Because in this method CO2 presents as an impurity with hydrogen, hence CO2 should be removed. Since stripper column is keyequipment in purification process, thus, in this study, stripper column is modeled and investigated by artificial neural network as a technique of nonlinear modeling. The number of variables used for modeling is 5 and 2 as input and output variables, respectively. Next, in order to validate, this model compared with multiple linear regression (MLP) method. Determining the input effective variables on performance of thecolumn is the later purpose of results of this modeling. The results reveal that the ANN method is more powerful tool than MLP one to describe and predict the column. However, the major and minor input variable for both methods are analogous.

Application of Artificial Neural Network and multiple linear regression for modeling and sensitivity analysis of a stripper column Keywords:

Multiple Linear Regression , Artificial Neural Network , Sensitivity Analysis , Stripper Column , Nonlinear modeling