Statistical Optimization of Carbon Dioxide Capture Performance by Tri-Solvent System of MEA-DEA-PZ from the Stored Gas Reservoir

Abstract

Aims/Objective: The current study evaluates the effectiveness of a blended tri-solvent system composed of monoethanolamine (MEA), diethanolamine (DEA), and piperazine (PZ) for removing the carbon dioxide (CO2) from the stored gas reservoir. The developed system is intended to be both environmentally benign and productive. Methods: The CO2 absorption was carried out for the total solvent (5, 10, and 15% v/v), during the course of three temperature ranges (20, 25, and 30°C) while maintaining the reservoir pressure of 1.5, 2, and 2.5 bar. The fraction of DEA: MEA has been restricted to 0.2, 0.5, and 0.8, with simultaneous loading of anhydrous PZ ranging from 0 to 2 gm. and agitation speed for step intervals of 300, 600, and 900 rpm. For the purpose of investigating the synergetic effects of the process parameters on the rapid absorption of CO2 (ξ) and the initial CO2 absorption rate (ε0), Box-Behnken Design (BBD) of response surface methodology (RSM) has been exploited. The design of experiments (DoE) assisted artificial neural network (ANN) and metaheuristic approach of hybridising ANN-whale optimization algorithm (WOA) was also developed and utilised to train and test the developed model. Three optimization models based on RSM, DoE-WOA and ANN-WOA were compared. Under the optimized operating conditions pertaining to DoE, DoE-WOA and DoE-ANN-WOA, (ε0 = 3.340, 3.460 3.513 gm./min-250 mL) and (ξ = 17.114, 18.069, 18.145 gm./250 mL) were obtained. Results: The hybridised DoE-ANN-WOA shows promising results in correlation with the experimental outcomes having error % for ε0 & ξ of 0.790 & 1.31. Conclusion: The DoE-ANN-WOA tends to be selected to predict the optimal absorption conditions as compared to other optimization techniques used in current article.