application of machine learning in prediction of carbon dioxide capture in an amine plant

In this paper, by studying the process of carbon dioxide capture using amine solutions and the applications of machine learning and artificial intelligence in chemical engineering, we have described the basics in this field, and in the following, the appropriate variables for process modeling along with the range of their changes in input selected. then, using the available data (1200 data) in different process conditions and their preprocessing, the process was modeled using the artificial neural network algorithm and the high correlation coefficient and minimum error was obtained on the test data for the output variables of the model. this model includes 15 input variables such as the amine concentration entering the absorption tower, amine recirculation rate, reboiler pressure, condenser temperature, the mass flow rate of flue gas flow, the mass fraction co2 in flue gas, boil-up ratio and etc., also 9 output variables such as removal percentage, cooler and condenser duty, mass fraction of co2 in the clean gas flow, rich and lean amine and etc. the noteworthy point in this modeling is the use of various variables with a wide range of changes, which will provide the scale and application of this modeling for different operating conditions.