مهندسی شیمی ایران

مهندسی شیمی ایران

مدل‌سازی جذب دی‌اکسید‌کربن با محلول آبی پیپرازین/ پتاسیم کربنات بااستفاده‌از شبکه‌های عصبی MLP، RBF

نوع مقاله : مقاله پژوهشی

نویسندگان
1 دانشجوی دکتری مهندسی شیمی، دانشگاه علم و صنعت ایران
2 استاد مهندسی شیمی، دانشگاه علم و صنعت ایران
چکیده
در این مقاله، تأثیر شرایط عملیاتی و محلول آبی بر نرخ جذب دی‌اکسید کربن بااستفاده‌از روش سطح پاسخ و شبکۀ عصبی مدل‌سازی‌شد. در این مطالعه، دما، میزان بارگذاری، فشار تودهای CO و فشار تعادلی CO به عنوان متغیرهای ورودی و شار انتقال جرم CO₂ به عنوان متغیر خروجی برای مدلسازی در نظر گرفته شدند. از نتایج یادگیری شبکۀ MLP با سه لایۀ پنهان و به‌ترتیب تعداد 10، 40 و 10 نورون در هر لایه و تابع آموزش لونبرگ مارکوات، MSE و R2 به‌ترتیب برابربا0/0018616 و 0/9924 به‌دست‌آمد. شبکۀ RBF نیز توانست به MSE و R2 به‌ترتیب برابربا 0/0004 و 0/99849 بعداز 200 اپوک دست‌یابد. اگرچه RBF میانگین مربعات خطای کمتری ارائه داد، MLP به دلیل معماری ساده‌تر، هزینه محاسباتی کمتر و تطابق کیفی نزدیکتر با سطح RSM به عنوان مدل ارجح انتخاب شد.
کلیدواژه‌ها
موضوعات

عنوان مقاله English

Modeling of CO2 Absorption with Aqueous Piperazine/Potassium Carbonate Solution Using MLP and RBF Neural Networks

نویسندگان English

F. Bahmanzadegan 1
A. Ghaemi 2
1 PhD. Student of Chemical Engineering, Iran University of Science and Technology
2 Professor of Chemical Engineering, Iran University of Science and Technology
چکیده English

In this article, the effect of operating conditions and aqueous solution on the CO₂ absorption rate was modeled using response surface methodology (RSM) and artificial neural network (ANN). The model inputs consisted of temperature, CO2 loading, bulk CO2 pressure, and equilibrium partial pressure of CO2, while the CO2 mass transfer flux was considered as the output. Multi-Layer Perceptron (MLP) was trained with three hidden layers containing 10, 40, and 10 neurons, respectively, using the Levenberg-Marquardt training function. The MLP with three layers and 60 neurons, trained with the Trainlm learning function, achieved an MSE of 0.0018616 and an R² of 0.9924 Radial Basis Function (RBF) also reached an MSE of 0.0004 and an R² of 0.99849 after 200 epochs. Although RBF provided a lower MSE, MLP was selected as the preferable model because of its simpler architecture, lower computational cost, and closer qualitative agreement with the RSM surface.

کلیدواژه‌ها English

Absorption
Piperazine
Potassium Carbonate
Carbon Dioxide
Neural Network
MLP
RBF
[1]        Ghaemi, A. & Hemmati, A. (2021). Modeling of Two Parameters Isotherm of CO2 and H2S Adsorption Using Metal Organic Frameworks (MOFs). Iranian Chemical Engineering Journal, 20(114), 22-36, [In Persian]. doi: 10.22034/ijche.2021.239817.1026
[2]        Ghaemi, A., Shahhosseini, S., & Maragheh, M. G. (2009). Nonequilibrium dynamic modeling of carbon dioxide absorption by partially carbonated ammonia solutions. Chemical Engineering Journal, 149(1-3), 110-117.
[3]        Ramezanipour Penchah, H., Ghaemi, A., & Ghanadzadeh Gilani, H. (2022). Efficiency increase in hypercrosslinked polymer based on polystyrene in CO2 adsorption process. Polymer Bulletin, 1-22.
[4]        Brunetti, A., Scura, F., Barbieri, G., & Drioli, E. (2010). Membrane technologies for CO2 separation. Journal of Membrane Science, 359(1-2), 115-125.
[5]        Song, C., Liu, Q., Deng, S., Li, H., & Kitamura, Y. (2019). Cryogenic-based CO2 capture technologies: State-of-the-art developments and current challenges. Renewable and sustainable energy reviews, 101, 265-278.
[6]        Rastegar, Z., & Ghaemi, A. (2022). CO2 absorption into potassium hydroxide aqueous solution: experimental and modeling. Heat and Mass Transfer, 58(3), 365-381. [7]        Ahmadi, M., Bahmanzadegan, F., Qasemnazhand, M., Ghaemi, A., & Ramezanipour Penchah, H. (2024). Experimental, RSM modelling, and DFT simulation of CO2 adsorption on Modified activated carbon with LiOH. Scientific Reports, 14(1), 13595.
[8]        Mohseni, A. H. and Ghaemi, A. (2020). Experimental Modeling of CO2 Absorption into Monoethanolamine Amine Using Response Surface Methodology. Iranian Chemical Engineering Journal, 19(111), 43-54, [In Persian].
[9]        Pashaei, H., Ghaemi, A., Nasiri, M., & Karami, B. (2020). Experimental modeling and optimization of CO2 absorption into piperazine solutions using
RSM-CCD methodology. ACS omega, 5(15),8432-8448.
[10]      Behroozi, A. H., Akbarzad, N., & Ghaemi, A. (2020). CO2 reactive absorption into an aqueous blended MDEA and TMS solution: experimental and modeling. International Journal of Environmental Research, 14, 347-363.
[11]      Ghaemi, A., Dehnavi, M. K., & Khoshraftar, Z. (2023). Exploring artificial neural network approach and RSM modeling in the prediction of CO2 capture using carbon molecular sieves. Case Studies in Chemical and Environmental Engineering, 7, 100310.
[12]      Tabarzadi, P., & Ghaemi, A. (2024). Modeling and optimization of CO2 capture in spray columns via artificial neural networks and response surface methodology. Case Studies in Chemical and Environmental Engineering, 10, 100783.
[13]      Etemad, E., Ghaemi, A., & Shirvani, M. (2015). Prediction of CO2 mass transfer flux in amine solutions using neural networks. Journal of Iranian Chemical Engineering, 14(79), 98-106, [In Persian].
[14]      Naderi, K., Foroughi, A., & Ghaemi, A. (2023). Analysis of hydraulic performance in a structured packing column for air/water system: RSM and
ANN modeling. Chemical Engineering and Processing-Process Intensification, 193, 109521.
[15]      Ghaemi, A., Jafari, Z., & Etemad, E. (2020). Prediction of CO2 mass transfer flux in aqueous amine solutions using artificial neural networks. Iranian Journal of Chemistry and Chemical Engineering, 39(4), 269-280.
[16]      Pashaei, H., Mashhadimoslem, H., & Ghaemi, A. (2023). Modeling and optimization of CO2 mass transfer flux into Pz-KOH-CO2 system using RSM and ANN. Scientific reports, 13(1), 4011.
[17]      Zafari, P., & Ghaemi, A. (2023). Modeling and optimization of CO2 capture into mixed MEA-PZ amine solutions using machine learning based on ANN and RSM models. Results in Engineering, 19, 101279.
[18]      Norouzbahari, S., Shahhosseini, S., & Ghaemi, A. (2015). Modeling of CO2 loading in aqueous solutions of piperazine: Application of an enhanced artificial neural network algorithm. Journal of Natural Gas Science and Engineering, 24, 18-25.
[19]      Garg, S., Shariff, A. M., Shaikh, M. S., Lal, B., Suleman, H., & Faiqa, N. (2017). Experimental data, thermodynamic and neural network modeling of CO2 solubility in aqueous sodium salt of l-phenylalanine. Journal of CO2 Utilization, 19, 146-156.
[20]      Pakzad, P., Mofarahi, M., Izadpanah, A. A., & Afkhamipour, M. (2020). Experimental data, thermodynamic and neural network modeling of CO2 absorption capacity for 2-amino-2-methyl-1-propanol (AMP)+ Methanol (MeOH)+ H2O system. Journal of Natural Gas Science and Engineering, 73, 103060.
[21]      Hemmati, A., Ghaemi, A., & Asadollahzadeh, M. (2021). RSM and ANN modeling of hold up, slip, and characteristic velocities in standard systems using pulsed disc-and-doughnut contactor column. Separation Science and Technology, 56(16), 2734-2749.
[22]      Khoshraftar, Z., & Ghaemi, A. (2023). Modeling of CO2 solubility in piperazine (PZ) and diethanolamine (DEA) solution via machine learning approach and response surface methodology. Case Studies in Chemical and Environmental Engineering, 8, 100457.
[23]      Noroozian, M., Shahhosseini, S., & Ghaemi, A. (2023). Artificial intelligence and response surface methodology to predict CO2 capture using piperazine‐modified activated alumina. Environmental Progress & Sustainable Energy, 42(3), e14117.
[24]      Khoshraftar, Z., & Ghaemi, A. (2023). Modeling and prediction of CO2 partial pressure in methanol solution using artificial neural networks. Current Research in Green and Sustainable Chemistry, 6, 100364.
[25]      Khoshraftar, Z., Ghaemi, A., & Taheri, F. S. (2023). Polyethylenimine-functionalized halloysite nanotube as an adsorbent for CO2 capture: RSM and ANN methodology. Current Research in Green and Sustainable Chemistry, 7, 100389.
[26]      Khoshraftar, Z., & Ghaemi, A. (2023). Prediction of CO2 solubility in water at high pressure and temperature via deep learning and response surface methodology. Case Studies in Chemical and Environmental Engineering, 7, 100338.
[27]      Bahmanzadegan, F., & Ghaemi, A. (2024). Exploring the effect of zeolite's structural parameters on the CO2 capture efficiency using RSM and ANN methodologies. Case Studies in Chemical and Environmental Engineering, 9, 100595.
[28]      Fernandes, R. O., & Tavares, M. C. (2024). Mitigation of the commutation failure problem in the HVDC multi-infeed scenario in Brazil using synchronized phasor measurement. Electric Power Systems Research, 235, 110829.
[29]      Norouzbahari, S., Shahhosseini, S., & Ghaemi, A. (2016). Chemical absorption of CO2 into an aqueous piperazine (PZ) solution: development and validation of a rigorous dynamic rate-based model. RSC advances, 6(46), 40017-40032.
[30]      Cullinane, J. T. (2005). Thermodynamics and kinetics of aqueous piperazine with potassium carbonate for carbon dioxide absorption. The University of Texas at Austin.