Microsoft Word - 211.docx CHEMICAL ENGINEERING TRANSACTIONS VOL. 61, 2017 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Guest Editors: Petar S Varbanov, Rongxin Su, Hon Loong Lam, Xia Liu, Jiří J Klemeš Copyright © 2017, AIDIC Servizi S.r.l. ISBN 978-88-95608-51-8; ISSN 2283-9216 Rate-Based Modelling and Simulation of Large-Scale CO2 Capture Using a Piperazine-Promoted Aqueous Ammonia Solution Ruize Lua,b,c, Kangkang Lib, Jian Chena,*, Hai Yub, Moses Tadec aState Key Laboratory of Chemical Engineering, Department of Chemical Engineering, Tsinghua University, Beijing 100084, China bCSIRO Energy, 10 Murray Dwyer Circuit, Mayfield West, NSW 2304, Australia cDepartment of Chemical Engineering, Curtin University of Technology Australia, GPO Box U1987, Perth, WA 6845, Australia cj-dce@tsinghua.edu.cn Due to the fast reaction rate of piperazine (PZ) with CO2, it has the potential to act as a promoter in aqueous ammonia (NH3)-based CO2 capture processes. A rigorous, rate-based model for the NH3–PZ–CO2–H2O system was developed using Aspen Plus®, and validated against experimental results. Absorption and desorption processes were simulated under real flue gas conditions to gain a practical understanding of the behaviour and characteristics of interactions between PZ-promoted NH3 solution and CO2. Adding PZ significantly increased the CO2 absorption rate in the NH3-based CO2 capture process via a fast reaction between PZ carbamate and CO2. The temperature along the column was higher than in the absence of PZ, and additional NH3 was released into the solution, which led to higher NH3 loss. Adding PZ also reduced the stripping heat, resulting in a smaller energy requirement for solvent regeneration. 1. Introduction Aqueous ammonia (NH3) has gained significant attention as a potential alternative to traditional amine solvents in post-combustion CO2 capture process, due to its low cost, low regeneration energy, high CO2 absorption capacity, and resistance to degradation (Mumford et al., 2015). Its commercial application for CO2 capture is restricted by the low absorption rate of CO2, and the high volatility of NH3 (Molina et al., 2015). Piperazine (PZ) is widely recognized as an effective rate promoter in amine solutions and the NH3-based PCC process. Fang et al. (2014) showed that the addition of PZ significantly increased the mass transfer coefficient of CO2 into aqueous NH3 by three to fourfold. Previous bench-scale research using wetted-wall columns has confirmed that the addition of small amounts of PZ to both unloaded and CO2 pre-loaded aqueous NH3 solutions can significantly increase the CO2 absorption rate (Li et al., 2013). To understand the overall performance of blended NH3/PZ solvent for CO2 capture, a rigorous model is needed to accurately predict its thermodynamic properties, including vapour–liquid equilibrium, chemical reaction equilibrium, and kinetic performance such as mass and heat transfer, materials and energy balance, and chemical kinetics. Yu et al. (2016) investigated CO2 absorption by NH3/PZ in a laboratory-scale packed column, and used the results to develop and validate a rate-based model. In this paper, a rigorous, rate-based model was developed using Aspen Plus® for CO2 absorption and desorption using PZ-promoted aqueous NH3, and validated against experimental results. Process simulations under large-scale conditions provided us with insight into the thermodynamic properties and kinetic behaviour of the capture process. Our rate-based model and improved understanding of PZ-promoted, aqueous NH3-based CO2 capture will guide further improvement, scale-up and techno-economic assessment of this process. DOI: 10.3303/CET1761310 Please cite this article as: Lu R., Li K., Chen J., Yu H., Tade M., 2017, Rate-based modelling and simulation of large-scale co2 capture using a piperazine-promoted aqueous ammonia solution, Chemical Engineering Transactions, 61, 1873-1878 DOI:10.3303/CET1761310 1873 2. Model description and validation 2.1 Model description The embedded models for the NH3–CO2–H2O and PZ–CO2–H2O systems provided by Aspen Plus v8.4 were combined to develop a thermodynamic model for the NH3–PZ–CO2–H2O system (Aspen Technology, 2010). The electrolyte non-random two-liquid (NRTL) model was chosen to calculate activity coefficients, enthalpies and Gibbs energies for non-ideal, CO2-loaded NH3 solution. The Redlich–Kwong equation of state was applied to calculate fugacity coefficients for the vapour phase. Henry’s law was applied to CO2, NH3, and N2. The chemical equilibrium constants were calculated using the following equation: lnKeq,j=a+ b T +cln(T)+dT (1) where Keq,j is the equilibrium constant of reaction j; T is the temperature in Kelvin; and the constants a, b, c and d are retrieved from Aspen Plus databanks (Aspen Technology, 2010). The differential heat of absorption was calculated from the vapour–liquid equilibrium data obtained from the Flash module calculations. The commonly used equation derived from the Gibbs–Helmholtz equation was applied: ∆Habs≈R [ ∂lnf̂CO2 ∂(1/T) ] P,x (2) where ∆Habs is the differential heat of absorption; f̂CO2 is the fugacity of CO2; and R is the gas constant. Transport properties are required when describing mass and heat transfer in a rate-based model. Various transport property models and corrections embedded in Aspen Plus were chosen to calculate the density, viscosity, thermal conductivity, diffusion coefficient and surface tension for the NH3-PZ-CO2-H2O system. The detailed setting of transport properties models can be found in our previous work (Yu et al., 2016). The rate-based model using the Aspen Plus RadFrac distillation module was employed to simulate CO2 absorption by PZ-promoted aqueous NH3. The chemical reaction model was consistent with the electrolyte NRTL model. The reactions of the formation of bicarbonate and carbamate were considered to be kinetically controlled. Reduced power law expressions were used for the kinetically controlled reactions: rj=kjT n exp(- Ej RT ) ∏(Ci) ∂i N i=1 (3) where rj is the rate of reaction; kj is the pre-exponential factor; n is the temperature exponent; Ej is the activation energy; Ci is the concentration of component i; and ∂i is the stoichiometric coefficient of component i. In this study, the concentration basis was molarity, the factor n was set to zero, and kj and Ej were derived from the work of Pinsent et al. (1956) and more recently Bishnoi and Rochelle (2002). 2.2 Model validation The model was verified against experimental data in a relatively large temperature and CO2 loading range. VLE data and differential heat of absorption data were used to validate the model for subsystems, including NH3– CO2–H2O system, PZ–CO2–H2O system and NH3–PZ–H2O system. For NH3–PZ–CO2–H2O system, differential heat of absorption data and data from CO2 absorption process were used. In general, the proposed model can satisfactorily predict thermodynamic properties of the NH3-PZ-CO2-H2O system. 0.0 0.2 0.4 0.6 0.8 0 10 20 30 40 50 60 70 80 90 100 Exp. 3.1/0.86 M NH 3 /PZ Pred. 3.1/0.86 M NH 3 /PZ Pred. 3.1 M NH 3 Pred. 0.86 M PZ H e a t o f a b s o rp ti o n ( k J /m o l C O 2 ) CO2 loading (mol CO2/mol Am) T = 40 ºC 0.0 0.2 0.4 0.6 0.8 0 10 20 30 40 50 60 70 80 90 100 Exp. 3.1/0.86 M NH 3 /PZ Pred. 3.1/0.86 M NH 3 /PZ Pred. 3.1 M NH 3 Pred. 0.86 M PZ H e a t o f a b s o rp ti o n ( k J /m o l C O 2 ) CO2 loading (mol CO2/mol Am) T = 60 ºC Figure 1: Calculated differential heat of absorption for 3.1/0.86 M NH3/PZ system compared with experimental data at 40 °C and 60 °C from Liu et al. (2012), “Exp” means experimental data, “Pred” means model prediction. 1874 Figure 1 shows that the model generally overestimates the heat of absorption at low loadings for the NH3–PZ– CO2–H2O system. Comparison of the heat of CO2 absorption in aqueous NH3 and aqueous PZ suggests that both the predicted and experimental heat of CO2 absorption are higher in PZ than in NH3 within the studied loading and temperature range. The heat of CO2 absorption in NH3/PZ mixtures should be higher than in NH3 alone, which is consistent with the model’s prediction. The rate-based model was validated using the experimental data of the CO2 absorption process under various conditions at different NH3 and PZ concentrations, temperatures and lean solution flow rates. To make a reliable comparison, the simulation conditions were consistent with the experimental conditions; detailed information of each case can be found elsewhere (Yu et al., 2016). 3. Process simulation Using the developed model, the effect of PZ on the performance of both absorption and desorption in a large- scale, aqueous NH3-based CO2 capture process was investigated. The simplified configuration used in this study is consisted of an absorber, a stripper and a heat exchanger. The concentrations of CO2 and NH3 in the flue gas were 11.60 % and 1.09 % and the temperature was 44.2 °C. The NH3 present in the flue gas was recovered from the cooling water in the pre-treatment column. As reported by Zhang et al. (2013), using aqueous NH3 to absorb CO2 in one absorber under typical 500 MW coal-fired power plant conditions will result in a large absorber (diameter = 40 m, packed height = 72 m) to achieve a high CO2 removal efficiency, which is not practical. As discussed in our previous publications, it is more practical to use three identical trains of CO2 capture in parallel. Each process train has a 12 m diameter absorber and deals with one-third of the total flue gas, which is approximately 1 Mt CO2 per year (Li et al., 2015). The mass flow rate of flue gas for each process train was 794.6 t/h. Table 1: Summary of CO2 capture system operating conditions and simulation results Case no. 1 2 3 4 5 Lean solvent conditions NH3 conc. (mol/kg H2O) 4 4 4 4 3.2 PZ conc. (mol/kg H2O) 0 0.4 0.2 0 0.4 CO2 loading (mol CO2/mol alk) 0.25 0.25 0.25 0.25 0.25 Temp of Leanin (°C) 35 35 35 35 35 Absorber conditions and results Packing height (m) 55 20 20 20 23 CO2 removal efficiency (%) 82.4 82.7 75.6 68.5 82.5 NH3 conc. in outlet gas (%) 5.3 9.9 7.0 5.3 7.0 Stripper conditions and results Stripper/reboiler temp (°C) 139.4 139.9 140.8 139.4 142.3 CO2 desorption rate (kg/h) 114,917 115,288 105,453 95,539 114,987 Heat duty (MJ/kg CO2) 4.09 3.97 4.41 4.82 3.75 Five cases using different NH3 and PZ concentrations were investigated. Packing heights were varied with the purpose of achieving an 80 % CO2 capture efficiency. Detailed conditions are shown in Table 1. A low NH3 concentration of 4 mol/kg H2O was chosen to reduce energy penalty from NH3 slip, as optimised in our previous work (Li et al., 2015). PZ was introduced into the NH3 solution in cases 2, 3 and 5 under a relatively low ratio to avoid precipitation and reduce degradation. To determine the effect of the total concentration of amino group, the concentrations of NH3 and PZ in case 5 were set to 3.2 and 0.4 mol/kg H2O, to ensure that the total concentration of amino group in blended PZ/NH3 solvent was equal to that of NH3 alone. 4. Results and discussion As shown in Table 1, comparing the packing height of the absorber in case 1 and case 2 indicates that PZ as a promoter increases absorption rate. The required packing height for achieving an 80 % CO2 removal efficiency was reduced from 55 to 20 m. Comparing case 1 and case 5 also shows that PZ accelerates absorption, indicating the fast absorption is not solely due to the increase in the concentration of amino group. Introducing PZ also causes greater NH3 loss. NH3 concentration in the vent gas was almost doubled by introducing PZ at a mole ratio of 10:1 (NH3 to PZ) in case 2. The trade-off between the reduction in capital costs and the increase in energy requirement for the NH3 abatement process needs to be considered when determining the optimal amount of PZ introducing into the system. The regeneration energy requirement is decreased by introducing PZ, despite the larger heat of absorption of the PZ solution. 1875 4.1 Absorber performance The rate-based model allows for a detailed analysis of absorber and stripper performance, improving our understanding of how PZ performs in the system. The temperature profiles, speciation profiles and reaction rates along the column in case 2 and case 4 were examined. The operating conditions for these two cases were the same, except for case 2, in which 0.4 M PZ was introduced into the solution. As shown by Figure 2, the reactions for PZ-related species primarily occurred near the top of the column, because of their relatively high concentrations and fast reaction rates with CO2. Figure 2(a) shows that PZ(COO-)2 increased while PZ and PZH+ decreased rapidly at the top of the column. As lean solvent entered the top of the column, the main reactants were PZCOO- and HPZCOO, and the main product was PZ(COO-)2. This indicates that the dominant reactions for PZ-related species were as follows: PZCOO - +CO2+H2O↔PZ(COO - )2+H3O + (4) HPZCOO+H2O↔PZCOO - +H3O + (5) As the concentration of PZ carbamate decreased, PZ became the main reactant for PZ-related species. In Figure 2(b), PZ and PZH+ became reactants near the top of the column, which is consistent with the speciation profile. 18.0 18.5 19.0 19.5 20.0 0.00% 0.02% 0.04% 0.06% 0.08% M o le p e rc e n ta g e o f P Z -r e la te d s p e c ie s Packing height (m) PZH + PZ PZCOO - /HPZCOO (a) 0.3% 0.4% 0.5% 0.6% 0.7% PZ(COO - ) 2 M o le p e rc e n ta g e o f P Z (C O O - ) 2 18.0 18.5 19.0 19.5 20.0 -1500 -1000 -500 0 500 1000 1500 PZ(COO - ) 2 PZH + /PZ HPZCOO PZCOO - P ro d u c ti o n r a te ( k m o l/ h r) Packing height (m) (b) Figure 2: Predicted profiles along the packing height in case 2: (a) PZ-related speciation in the liquid phase at the upper stages of the column; (b) production rates of PZ-related species at the upper stages of the column. Positive production rate value means net production in the liquid phase As shown in Figure 3(a), NH3 concentration generally decreased from top to bottom in both cases, as expected, while it increased slightly near the top of the column in case 2. Figure 3(b) shows that the reaction rates of NH3- related species were almost the same in both cases, except for the reactions occurred near the top of the column. Instead of being a reactant, NH3 became a product, while NH2COO- became a reactant in case 2. This indicates that the reactions occurring for NH3-related species near the top of the column were as follows: NH2COO - +H3O + ↔NH3+CO2+H2O (6) NH3+H3O + ↔NH4 + +H2O (7) Combined with the reactions for PZ-related species, the total reactions occurring near the top of the column were: NH2COO - +PZCOO - ↔PZ(COO - )2+NH3 (8) NH2COO - +HPZCOO↔PZ(COO - )2+NH4 + (9) As indicated in Eq(4) – (9), no apparent HCO3- is involved, which is consistent with the speciation predication in Figure 3(a). The rate of consumption of NH2COO- in Figure 3(b) and the rate of formation of PZ(COO-)2 in Figure 2(b) were almost the same at the top of the column, which again confirms the reaction between NH3 carbamate and PZ carbamate. In general, PZ accelerates CO2 absorption by reacting with CO2 directly in the form of PZ carbamate, which releases additional NH3 into the solution through Eq(6). 1876 0 5 10 15 20 0% 1% 2% 3% 4% 5% NH3 NH4 + NH2COO - HCO3 - CO3 2- M o le p e rc e n ta g e Packing height (m) (a) 15 16 17 18 19 20 -1500 -1000 -500 0 500 1000 NH 3 NH 4 + NH 2 COO - P ro d u c ti o n r a te ( k m o l/ h r) Packing height (m) (b) Figure 3: Predicted profiles along the packing height in case 2 (solid line) and case 4 (dash line): (a) HCO3-, CO32- and NH3 related species profiles in the liquid phase; (b) production rates of NH3-related species. Figure 4 shows the liquid and vapour temperatures along the column. The addition of PZ increased the temperature, because of the larger heat of absorption of PZ and greater amount of CO2 absorbed. The increase of temperature can both accelerate the reaction and increase NH3 loss. In summary, introducing PZ into the NH3 solution in the absorption process has two effects. One is that additional NH3 is released via the reaction between PZCOO- and NH2COO- in the solution. The other is that more heat of reaction is released, which increases the temperature along column. Both effects contribute to a faster reaction rate between NH3 and CO2, and greater NH3 loss in the absorber. 0 5 10 15 20 30 35 40 45 50 55 T e m p e ra tu re ( ºC ) Packing height (m) Liquid Temp, case2 Vapour Temp, case 2 Liquid Temp, case 4 Vapour Temp, case 4 0.12 0.16 0.20 0.24 0 2 4 6 8 P a c k in g h e ig h t (m ) Vapour mole fraction H 2 O, case 1 H 2 O, case 2 H 2 O, case 3 H 2 O, case 4 NH 3 , case 1 NH 3 , case 2 NH 3 , case 3 NH 3 , case 4 Figure 4: Predicted temperature profile along the packing height in case 2 and case 4. Figure 5: Vapour fraction of H2O and NH3 along the column in cases 1 – 4. 4.2 Stripper performance Assuming that no heat loss occurred in the CO2 stripping process, the energy requirement for CO2 regeneration can be calculated using the following equation: QR=HS+HD+QC (10) where the sensible heat, HS , is the energy required to heat the solvent from inlet temperature to outlet temperature; desorption heat, HD, is the energy required for CO2 desorption reactions; and stripping heat, QC, is the energy required for H2O and NH3 vaporisation. Table 2 shows the distribution of the regeneration energy consumption calculated by Eq(11) in cases 1 – 4. In cases 2 – 4, regeneration energy decreased when PZ concentration was increased under the same operating conditions. The desorption heat remained almost the same because of similar CO2 loadings. Due to the similar composition of solvents in all cases, sensible heat is mainly determined by the amount of CO2 desorbed by the stripper, which is determined by the amount of CO2 absorbed in the absorption process. This indicates that increasing CO2 capacity of the solvent is the key to reducing sensible heat. CO2 capacity is increased by adding PZ, which is in favour of reducing the sensible heat. Figure 5 shows that H2O and NH3 concentration decreased when PZ concentration increased in cases 2 – 4. Since temperature and total vapour flow rates were similar among these cases, adding PZ decreased the stripping heat, as shown in Table 2. Comparing case 1 and 2 considers the situation when the desorbed CO2 flow rate is fixed. The regeneration energy of case 2 was slightly lower than that of case 1. The sensible heat and desorption heat of both cases were similar, due to the similar compositions of rich solvents. The lower 1877 regeneration duty in case 2 is attributed to the smaller stripping heat, which is due to the lower NH3 and water concentrations in the vapour phase as shown in Figure 5. Table 2: Distribution of the regeneration energy consumption Case no. NH3/PZ molality HS (MJ/kg CO2) HD (MJ/kg CO2) -QC (MJ/kg CO2) Regeneration energy (MJ/kg CO2) 1 4/0 1.89 1.59 0.61 4.09 2 4/0.4 1.84 1.61 0.52 3.97 3 4/0.2 2.04 1.61 0.76 4.41 4 4/0 2.27 1.60 0.95 4.82 5. Conclusions Detailed analysis of the simulation results reveals that PZ addition can effectively accelerate the absorption rate in the NH3-based CO2 capture process, and reduce the height of absorber required to achieve a high CO2 removal efficiency. The temperature along the column was higher in the presence of PZ due to its larger heat of absorption. Reactions happened along the packing of the absorber were studied. Additional NH3 was released into the solution via the reaction between PZ carbamate and NH3 carbamate at the upper stages of the absorber. However, NH3 loss also increased due to higher temperature and higher free NH3 concentration. With PZ as a promoter in the system, the capital costs of the absorption process can be largely reduced and the energy requirement for solvent regeneration is decreased by reducing stripping heat for the conditions evaluated. Acknowledgments The authors gratefully acknowledge financial support from CSIRO Energy and Curtin University, and National Science and Technology Support Program of China (No. 2015BAC04B01 and 2015BAC04B02), Ruize Lu is also grateful for a Tsinghua University short-term visiting scholarship. References Aspen Technology, 2010, Aspen physical property system: physical property methods, Aspen Technology, Cambridge, MA, USA. Bishnoi S., Rochelle G.T., 2002, Absorption of carbon dioxide in aqueous piperazine/methyldiethanolamine, AIChE Journal, 48, 2788–2799. Fang M., Xiang Q., Zhou X., Ma Q., Luo Z., 2014, Experimental study on CO2 absorption into aqueous ammonia- based blended absorbents, Energy Procedia, 61, 2284–2288. Liu J., Wang S., Svendsen H.F., Idrees M.U., Kim I., Chen C., 2012, Heat of absorption of CO2 in aqueous ammonia, piperazine solutions and their mixtures, International Journal of Greenhouse Gas Control, 9, 148– 159. Li K., Yu H., Feron P., Tade M., Wardhaugh L., 2015, Technical and energy performance of an advanced, aqueous ammonia-based CO2 capture technology for a 500 MW coal-fired power station, Environmental Science & Technology, 49, 10243–10252. Li L., Han W., Yu H., Tang H., 2013, CO2 absorption by piperazine promoted aqueous ammonia solution: absorption kinetics and ammonia loss, Greenhouse Gases: Science and Technology, 3(3), 231–245. Molina C.T., Bouallou C., 2015, Assessment of different methods of CO2 capture in post-combustion using ammonia as solvent, Journal of Cleaner Production, 103, 463–468. Mumford K.A., Wu Y., Smith K.H., Stevens G.W., 2015, Review of solvent based carbon-dioxide capture technologies, Frontiers of Chemical Science and Engineering, 9(2), 125–141. Pinsent B.R.W., Pearson L., Roughton F.J.W., 1956, The kinetics of combination of carbon dioxide with ammonia, Transactions of the Faraday Society, 52, 1594–1598. Yu J., Wang S., Yu H., 2016, Experimental studies and rate-based simulations of CO2 absorption with aqueous ammonia and piperazine blended solutions, International Journal of Greenhouse Gas Control, 50, 135–146. Zhang M., Guo Y., 2013, Process simulations of large-scale CO2 capture in coal-fired power plants using aqueous ammonia solution, International Journal of Greenhouse Gas Control, 16, 61–71. 1878