【www.zhangdahai.com--其他范文】
Ran An, Shengxin Chen, Shun Hou, Yuting Zhu,, Chunhu Li, Xinbao Zhu, Ruixia Liu,4,*,Weizhong An,*
1 Department of Chemical Engineering, Ocean University of China, Qingdao 266100, China
2 Beijing Key Laboratory of Ionic Liquids Clean Process,CAS Key Laboratory of Green Process and Engineering,State Key Laboratory of Multiphase Complex Systems,Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China
3 College of Chemical Engineering, Nanjing Forestry University, Nanjing 210037, China
4 School of Chemical Engineering, University of Chinese Academy of Sciences, Beijing 100049, China
Keywords:Propylene oxide Reactive distillation Heat integration Propylene glycol monomethyl ether Process simulation
ABSTRACT A double-effect reactive distillation(DERD)process was proposed for the production of propylene glycol methyl ether from propylene oxide and methanol to overcome the shortcoming of low selectivity and high-energy consumption in the tubular plug-flow reactor.A single-column reactive distillation(RD)process was conducted under optimized operating conditions based on sensitivity analysis as a reference.The results demonstrated that the proposed DERD process is able to achieve more than 95% selectivity of the desired product.After that, a design approach of the DERD process with an objective of the minimum operating cost was proposed to achieve further energy savings in the RD process.The proposed DERD configuration can provide a large energy-savings by totally utilization of the overhead vapor steam in the high-pressure RD column.A comparison of the single-column RD process revealed that the proposed DERD process can reduce the operating cost and the total annual cost of 25.3% and 30.7%, respectively, even though the total capital cost of DERD process is larger than that of the RD process.
Propylene glycol methyl ether (PGME) is an important derivative of propylene oxide (PO) and is widely used as an ecofriendly solvent in the production of detergents, cosmetics and emulsifiers[1].Additionally, electronics-grade PGME is also used as a solvent for photoresist processing in the semiconductor industry.In China,the demand for PGME has been steadily increasing since the 2010s,having an average growth rate of 4% per annum.
Currently, PGME is commercially produced by reacting the PO with methanol (MeOH),i.e., the propoxylation of MeOH in the presence of an acid or base catalyst.The reaction system is composed of parallel-consecutive reactions and the PGME isomers can further polymerize with PO to generate unwanted polyether polyols as byproducts.In addition, PGME has two isomers: 2-methoxy-1-propanol (1PO1) and 1-methoxy-2-propanol (1PO2).Accordingly, most current studies of PGME production have focused on the development of catalysts and reactor technologies for the improvement of selectivity.To date, the catalytic production of PGME using basic catalysts has attracted more interest than acid catalysts because of main producti.e.base-catalyzed reaction(1PO2)is less toxic than produced from the acid-catalyzed reaction(1PO1) [2].In addition to that much effort has been devoted to enhancing the selectivity toward the desired product through process intensification.Currently, the most industrially applied reactor technology for continuous PGME production is based on a tubular plug-flow reactor(PFR)[3].This type of reactor minimized the dead zones,temperature gradients,and local hot spots.Besides this, such reactions can be operated at higher temperatures than other designed reactors and achieving higher space time yields[4].However,there are several disadvantages;For example,a high molar feed ratio of MeOH to PO is required to ensure high selectivity for PGME, which leads to significant energy consumption because of recovering the unreacted MeOH content from the reaction mixtureviadistillation.Furthermore, insufficient heat transfers from the reactor tubes impeded the stable reactor operation due to the exothermic reactions.To solve these problems,different reactor technologies such as microreactor have been introduced.Recently, Liuetal.[5] proposed a microtubular circulating reactor with ionic liquid catalysts for highly efficient PGME production.They explained the reaction time can be reduced from 3 h to only 0.33 h, and achieved a 92% yield of PGME as compared to the stirred-tank reactor.The utilization of microreactor has so many advantages such as high heat- and mass-transfer rates and excellent safety, and therefore the risk of reactor runaway has been eliminated.On the other hand, microreactors have several disadvantages, including a tendency for clogging, catalyst loading,highly intricate designs, and the need for complex microchannels fabrication [6,7].Therefore, the application of microreactors for PGME production on an industrial scale has not yet been realized.
Reactive distillation (RD), also known as catalytic distillation(CD),is often subjected to process intensification to unify the reaction and separation processes in a single piece of operational equipment [8].RD has proved to be a wise choice for the mixed irreversible and competitive consecutive reactions, enabling improvement in the selectivity for the desired intermediate product [9,10].Although RD technologies and applications have been widely studied, to date, only Parker patented a RD-based process for PGME production using [11].In their process, a dual homogeneous/heterogeneous catalyst, in which a homogeneous basic catalyst is dissolved in MeOH and the heterogeneous basic catalyst is fixed in the RD column,was adopted to increase the overall activity of the catalyst and prevent the deactivation of the heterogeneous catalyst.The result of above patent demonstrated that applying the RD process could achieve lower capital and operating costs compared with a conventional process (case 1), which consisted of an equilibrium reactor followed by a distillation column.However,the studies about of reaction kinetics,the byproducts(including isomers of dipropylene glycol methyl ether)and selectivity for the desired product were deficient in the process.
To further achieve energy savings and improve economic performance, several process intensification technologies have been proposed in the distillation and RD fields such as heat pumpassisted distillation (HPAD), internally heat-integrated distillation columns (HIDiCs), and divided-wall columns (DWCs) [12–14].Multiple effect distillation (MED) is another promising and practical intensification technique for achieving energy savings.In this method, the condenser of a high-pressure (HP) column is integrated with the reboiler of a low-pressure (LP) column [15,16].These types of integrated arrangements have been proven to yield substantial energy savings in the separation of liquid mixtures, as well as for seawater desalination [17–21].However, to the best of our knowledge, there are very few reports in the literature which discuss the energy-savings problem in RD processes by applying the MED scheme.
In this work, a double-effect reactive distillation (DERD),namely RD process combined with double-effect distillation energy-savings technology, was conceptually proposed for PGME production to decrease the energy consumption.The primary attention is paid to identify the suitable operating parameters of RD process from the perspective of preliminary process design.Particularly,a near-optimal configuration of the RD process is suggested by considering the effects of various parameters on the column performance.The steady-state simulations using Aspen Plus software 11.0 is performed.Then,the concept of double-effect distillation, which has been widely used in conventional distillation fields for energy-saving, then extended to RD process.Moreover,a systematic approach for designing a heat-integrated DERD process is developed.Finally, the energy consumption and economic performance indicated by total annual cost(TAC)for the proposed single-column RD-based process and the double-effect RD process were discussed comprehensively.
2.1.Chemiacl reaction system
To apply RD process to PGME production,an imidazole polymer(divinylbenzene–vinyl imidazole) catalyst (P-DVB-VIM) has been developed, and the kinetics model and parameters have been provided in our previous work[2].Briefly,due to the asymmetric characteristic of the PO, the epoxide ring may open either at the C—O bonds to generate two PGME isomers,i.e., 2-methoxy-1-propanol(1PO1) and 1-methoxy-2-propanol (1PO2).After that, PO can further react with PGMEs(two isomers)and producing the unwanted byproduct dipropylene glycol methyl ether (DPGMEs), which involves also two isomers,i.e., 1-(2-methoxypropoxy)-2-propanol(2PO1) and 1-(2-methoxy-1-methylethoxy)-2-propanol (2PO2)respectively.Based on this, the reactions to form PGMEs and DPGMEs can be expressed by Eqs.(1)–(4).
Based on the reaction mentioned above, the PGME production system consists of parallel-consecutive reactions and all reactions were irreversible.Because of the lower toxicity of 1PO2 than that of other products, it is reasonable to select the former as the desired target product.The thermodynamic properties of the six involved components are listed in Table 1.The heats of four reactions under 120 °C are -88.8, -81.0, -100.4, and -89.7 kJ∙(mol PO)-1, respectively, which are considered as highly exothermic.
Table 1Physical properties of components in the reaction system
2.2.Kinetic model
The rates of the reactions in Eqs.(1)–(4) mentioned above is expressed by Eqs.(5)–(8), respectively.
Here,Ciis concentration of componenti,mol∙L-1.kiis the reaction rate constants of the associated reactions, L∙mol-1∙min-1.The kinetic parameters expressed as a function of temperature using the activation energies and pre-exponential factors are given by the following Arrhenius law forms (Eqs.(9) and (10)):
whereTis the temperature (K); andRis the universal gas constant(8.314 kJ∙kmol-1∙K-1).As shown by Eqs.(9) and (10) that the reactions(1)and(3),the reactions(2)and(4)have approximately equal rate constants (the same pre-exponential factors and activation energies), which is in accordance with the mechanism of propoxylation reported by Derio [22].
For simulating a RD process,the reaction volume on each stage need to know.If the reaction is homogeneous,the reaction volume is commonly expressed by‘‘liquid holdup”because chemical reactions take place in the liquid phase.For heterogeneous or catalytic reactions, the simplest way is to treat the reaction pseudohomogeneously,whereby catalyst diffusion and reaction is lumped into an overall reaction term.In this case,one only needs to specify catalyst mass and activity [23].This approach is adopted in this work.On the other hand,for Aspen Plus simulation purposes(reaction rates are not based on the catalyst mass, the pre-exponential factors unit used for heterogeneous reaction is converted into volume-based form), the catalyst loading expressed by volumebased form rather than liquid holdup is adopted in this work.
2.3.Vapor-liquid equilibrium model
The simulation of the production of PGME in the RD process was carried out by Aspen Plus 11.0 with the Radfrac module to describe the multistage vapor–liquid equilibrium (VLE) and reaction kinetics.For a RD process, modeling is required to describe and predict the reaction and separation that occur simultaneously.The components involved in this reaction system are PO, MeOH,PGMEs and DPGMEs.The non-ideality with respect to liquid phase behavior is due to their different molecular weights.On the other hand, due to the RD column operated with a higher pressure, the non-ideality with respect to vapor phase should be also taken into account.Based on these, the vapor–liquid equilibrium model is selected in this work by using UNIQUAC equation for liquid activity coefficient calculations and Redlich–Kwong (RK) equation for vapor fugacity calculations.The rationale for using the UNIQUAC and its validity are based on two aspects.Firstly, UNIQUAC model is applicable to multicomponent mixtures of nonpolar and polar liquids, especially the mixtures have large molecular size difference, as addressed in the studied system.Secondly, the UNIQUAC binary interaction parameters estimated from the experimental data are available in Aspen Plus database.According to the help module of Aspen Plus,UNIQUAC model is recommended to be used for phase equilibrium calculations.In this work, all the model parameters used for vapor–liquid equilibrium calculation were provided by the Aspen Plus database.
3.1.The single column RD process characteristics
The reasons to apply the RD process for PGME production are based on several aspects.First, the reaction condition (pressure and temperature)are mild enough to allow productionviareactive distillation[24].Second,RD can enhance the selectivity,especially when an intermediate product (1PO2) is desired and byproducts formation in the serial-parallel reactions [9,10].Third, the difference in mixture volatilities (Table 1) can offer significant advantages for rapid separation of 1PO2, which is also beneficial to improving the 1PO2 selectivity.Fourth, the PO ring-opening and subsequent polymerization are accompanied by a large amount of thermal heat released, making this system a good candidate for the heat integrated design.Finally, using a solid catalyst, the common problem of all conventional processes faced, such as catalyst separation, liquid waste treating and equipment corrosion,can be avoided.
Fig.1 shows a flowsheet of single column RD process.The RD column includes two feed inlets, a total condenser and a partial reboiler.The lower-boiling reactant (PO) is introduced into the lower part of the column, and higher-boiling reactant (MeOH) is introduced at the top of the column.Because no desired overhead product removed from the top of the RD column,it is reasonable to eliminate a rectifying section.Correspondingly, the RD column is operated under total-reflux mode.The products leaving from the bottom of the RD column involve predominantly the PGMEs,DPGMEs and unconverted MeOH.The mixture needs to be further separated in a distillation column (MS column) to recover unreacted MeOH, where 99.9% unreacted MeOH is obtained as the top product and recycled to the RD column for reuse (mixed with fresh MeOH).Another feature of the flowsheet given in Fig.1 is the inherent opportunity for conserving energy through the use of heat integration.It is noted that the condenser temperature of the RD column is higher than the reboiler temperatures of MS column due to high reaction temperature and thus, the integration between RD column and MS column is possible, as showed in Fig.1.
Fig.1. Flowsheet for PGME production of the RD process.
3.2.Identification of process conditions and parameters
The base conditions and ranges of operating parameters for the RD process used for simulation runs were gathered from our laboratory, which are important both for safety and process analysis.
(1) The fresh PO is feed to the RD column at a rate of 5000 kg∙h-1, yielding roughly 60000 tons (based on 8000 operating hours per year)of PGME products per year.A complete PO conversion is specified as a constraint for RD column.The MeOH introduced in RD column consists of two parts:fresh MeOH and recycled MeOH.The fresh MeOH with a flow rate of 2758 kg∙h-1was fed into stage 2 of the RD column (the fresh MeOH/PO feed molar ratio was one).To achieve nearly complete conversion of PO in the RD column,the molar feed ratio of MeOH to PO in the feed should be greater than one based on their stoichiometric coefficients in the Eqs.(1)–(4).In other words, the feed molar ratio of MeOH to PO introduced into the RD column can be considered an adjustable operating parameter varied by MeOH recycle flowrate distilled from the MS column,which affects the feed molar ratio introduced in RD column and energy consumption of MS column simultaneous.In the following discussion,the molar feed ratio of total MeOH introduced in RD column to PO was claimed feed molar ratio of MeOH to PO.
(2) The temperature of the reactive zone was limited to 100–150 °C based on previous kinetic characteristics over the PDVB-VIM catalyst.Briefly, lower temperatures (below 100°C)resulted in a low conversion of PO in the RD column,whereas higher temperatures accelerate reaction rate and catalyst deactivation, leading to potentially dangerous reactor operating conditions.Within the specified temperature range, the absolute operating pressure of the reactor should be 0.3–1.4 MPa based on the MeOH saturated vapor pressure because the predominant component in the reactive stages is MeOH.
(3) A maximum concentration of PO in the reflux liquid was set as a constraint for RD column simulation and optimization.As previously mentioned, a higher concentration of PO in the RD column has the potential to a hazard of runaway reactions.Because the aim of this work is to develop a reactor having industrial applications, the complete conversion of PO is essential.Because there is almost no PO in the stripping zone and PO conversion in the whole column is nearly 100%, the concentration of PO in the reflux liquid (CPO) as a constraint was used instead of the PO conversion.Specifically, theCPOis limited below 0.5% (mass).
(4) The ambient temperature water (25°C to 35 °C)is specified as coolant, the steams with different levels are used as a heating medium.The cost relations of economics and equipment sizing are summarized in Table 2.In this work, an‘‘equilibrium stage” or ‘‘theoretical tray” based assumption was adopted to simulate and design a RD based process.therefor the column hardware designs (the type of packing in packed column or tray sizes in tray column) are not considered in details.In other words, the way to simulate and design RD process does not include column internals opti-mization.Based on these, the specific parameters including HETP of packing and heat-transfer coefficient of trays were not considered in this manuscript.
Table 2Basis for the economic evaluation [24]
(5) For the MS column,a total stage number of 30,an operating pressure of 0.1 MPa,and a MeOH recovery of 99.9%are specified.The reflux ratio could be determined according to the specific MeOH recovery requirement.On the basis of these specification,sensitivity analysis is then carried out to determine the suitable feed location of MS column.It was noteworthy that, in order to perfume a sensitivity analysis for feed location of MS column, the objective parameters of RD column (PO conversion and 1PO2 selectivity) should be also considered since there exists a circulating stream between the two columns.
3.3.Parameters analysis
Based on the basic process conditions and constraints mentioned above, the key design parameters were analyzed based on sensitivity analysis approach, which have significant effects on the column performance.For the RD column, the selected operating parameters are pressure (P), total number of stage (N), feed location of PO (NF), catalyst loading (V), and boilup ratio (BR);For the MS column,the selected operating parameters are the distillate flow rate(D)and reflux ratio(RR).Note that,the feed ratio of MeOH to PO (nMeOH/nPO) of RD column was varied as the MeOH recycle flow rate distillate from the MS column, which was analyzed in the following discussion.The initial operating parameters for the RD column were identified:P=0.8 MPa,N=42,NFPO=30,nMeOH/nPO=1.57, BR=3.0,V=500 L; The initial operating parameters for the MS column were identified:P=0.1 MPa, RR=2.4,N=30, NFPO=12.It is noted that this approach could obtain the designed result which is close to the optimal design value,although this sensitivity analysis of the design variables does not give the global optimal point.
Fig.2. Effects of the pressure on the PO concentration in the reflux liquid, 1PO2 selectivity and reaction temperature.Other conditions of RD column: N=42,NFPO=30, nMeOH/nPO=1.57, BR=3.0, V=500 L.
3.3.1.EffectofRDpressure
In the RD column,a higher operating pressure leads to a higher reaction temperature, because the boiling point of the liquid mixture depending on the column pressure.Consequently,the PO conversion, 1PO2 selectivity and reboiler temperature could be affected.In addition,the operating pressure has a significant influence on the catalyst stability(limited by deactivation temperature of 150°C in Section 3.2)and reboiler temperature(limited by temperature level of heating medium).Thus,the effect of the pressure on PO concentration in the reflux liquid,1PO2 selectivity and reaction temperature in the absolute pressure range of 0.7–1.4 MPa was examined.As shown in Fig.2, the temperature in the RD column increased with pressure,which affected the reaction rate and PO conversion, leading to lower PO concentration in reflux liquid.However, the 1PO2 selectivity decreased from 95.64% to 95.1% as the operating pressure increases from 0.7 to 1.4 MPa.This result indicated that the operating pressure has a little effect on the 1PO2 selectivity, which could be attributed to the main reaction and side reactions being accelerated simultaneously as the increase of reaction temperature.In particular, a high reaction temperature tends to cause safety issues,such as temperature runaway, thus the operating pressure should not be too high.Therefore, based on the results of the sensitivity analysis, the most appropriate operating pressure for the RD process was 0.85 MPa.
3.3.2.EffectsofthemolarfeedratioofMeOHtoPO
The feed ratio of MeOH/PO is a very critical parameter, which affect the performance and operating cost of two columns.Because the feed molar ratio of MeOH to PO was affected by MeOH recycle flowrate distilled from the MS column, the analysis of feed molar ratio of MeOH to PO was also considered as that of the MeOH distillate flowrate of the MS column.The effect of the MeOH/PO feed ratio on the column performance was investigated varying from 1.3 to 1.7,and the results were shown in Table 3.As can been seen,with the increase of MeOH/PO feed ratio, the following changes occur:Firstly,the overhead temperature(TC)and the average value of reaction temperature (TR) were almost kept constant, resulting from the fact that MeOH is the dominant component in the reactive sections of column (almost more than 90%, see the simulated concentration profiles in Fig.6).Secondly,the reboiler temperature(TB)significant decreased because a significant amount of untransformed MeOH (whose boiling point is lower than those of the PGMEs products)entered the reboiler.Thirdly,the 1PO2 selectivity increased, which is because when the reaction system consists of sets of parallel consecutive reactions, the increase in the MeOH/PO feed ratio leads to a decrease in the PO concentration in the RD column,which enhances the selectivity for the desired intermediate product.Fourthly, PO concentration increased in the reflux liquid,which is attributed to a higher MeOH feed flow rate results in a decrease in the residence time of PO in the reactive sections,thereby increasing PO concentration at the top of the column.Overall, an excess of MeOH in this reaction system is beneficial for improving 1PO2 selectivity,but also lead to a high-energy consumption due to recovery of the excess MeOH content from the reaction mixture.Taking these factors together, an operational MeOH/PO feed ratio of 1.5 was suggested for this RD process.
3.3.3.Effectsoftheboilupratio
Because the proposed RD column is operated in total-reflux mode and the product is only removed from the column bottom,the boilup ratio was set as an operating parameter substitute for the reflux ratio to allow the flow rate inside the RD column to be adjusted.The boilup ratio affects both reaction and separation performance of the RD process.In addition, an excessive boilup ratio leads to operating problems in the RD column, for example, column flooding.Consequently, a suitable boilup ratio must be identified to ensure good separation performance, a low reboiler duty,and sufficient reaction residence time in the RD column.Table 4 described the obtained results on varying the boilup ratio from 2.0 to 4.0 on the RD column performance.As seen from Table 4,firstly, the 1PO2 selectivity and top PO concentration increased with the boilup ratio increased.This is because higher boiler ratio can lead to a higher internal flow rate.On one hand,MeOH concentration and molar ratio of MeOH to PO in the column was increased, the separation efficiency of the products and unreacted MeOH (which is forced to recycle back to the column) will beimproved.On the other hand, a high internal flow rate in the column means that the PO have a short residence time in the reactive stages, which results in their incomplete conversion.Secondly, as the boilup ratio increased, the condenser temperature decreased and the reactive section temperature remained almost unchanged,but the reboiler temperature increased, which is attributed to the higher-boiling products that are enriched in the column bottom and the PO enriched at the column top.Beside this, the increase in the boilup ratio results in a higher reboiler duty and higher energy consumption.Considering these results together, it suggested that an appropriate boilup ratio for the proposed RD process was 2.5.
Table 3Effects of the molar feed ratio of MeOH to PO on the column performance.Other conditions of RD column: P=0.8 MPa, N=42, NFPO=30, BR=3.0, V=500 L
Table 4Effects of the boilup ratio on the column performance.Other conditions of RD column: P=0.8 MPa, N=42, NFPO=30, nMeOH/nPO=1.57, V=500 L.
Fig.3. Effects of the total number of trays on the PO concentration in the reflux liquid and 1PO2 selectivity.Other conditions of RD column:P=0.8 MPa,NFPO=30,nMeOH/nPO=1.57, BR=3.0, V=500 L.
3.3.4.Effectsoftotalnumberoftrays
The total number of trays will improve the separation ability of the column, but the excess of trays increases more equipment investment.As shown in Fig.3,a higher number of trays was beneficial to improve 1PO2 purity and reduce PO concentration.The 1PO2 selectivity increased obviously, when the number of trays increased from 30 to 33, and then it increased slightly when the number of trays was more than 33.The top PO concentration decreased from 11.04%to 0.496%as the number of trays increased from 30 to 44,and remain negligibly changed when the number of trays exceed 44.According to the analysis, the suggested total number of trays was 44.
Fig.4. Effects of the feed location of PO on the PO concentration in the reflux liquid and 1PO2 selectivity.Other conditions of RD column: P=0.8 MPa, N=42, nMeOH/nPO=1.57, BR=3.0, V=500 L.
3.3.5.EffectsoffeedlocationofPO
The PO feed location is an important parameter of the RD column since it affects directly the number of stages of the reaction zone and the stripping section.In this work,we assumed that reactive zone is between MeOH and PO feed location,thus the location of the reactive zone is changed by PO feed location.As can be seen in Fig.4, when the PO feed location is moved down along column,the 1PO2 selectivity decrease.This is because of higher PO concentration at the bottom of the column, resulting in a decrease in the MeOH/PO ratio on the stage and a decrease of 1PO2 selectivity.The top PO concentration is decreased from 1.50% to a minimum of 0.498% and then the top PO concentration up to 5.15% as the PO feed location is moved down along column.This was caused by reducing in the number of stripping stages and a decreasing the MeOH/PO ratio in each stage.According to the analysis,the favored PO feed location was at stage 32.
3.3.6.Effectsofcatalystloading
The catalyst loading is a very important parameter to reactive distillation.An adequate catalyst loading on the reactive trays is essential to realize the complete conversion of reactants and meet hydraulic limitations.The total catalyst loadings placed in the column with a volume of 15 m3is estimated firstly,aiming at achieving a complete conversion of reactant PO in RD column.The estimation is based on the space velocity of reaction obtained fromprevious kinetic studies (the catalyst volume required to accomplish a conversion of PO close to 0.99 in a RD column was approximately 500 L).Then, an identical catalyst loading value of each stage was specified for simulation.The effect of varying the catalyst loading on the PO concentration in the reflux liquid and 1PO2 selectivity was investigated in the range of 300–650 L.As shown in Fig.5,the PO concentration decreased as the catalyst load increased.However, as the catalyst loading increased, the residence time increase and the more 1PO2 react with PO to form byproduct 2PO.Although a further increase of the catalyst loading can result in a small decrease of PO concentration in the reflux liquid, there will be sacrificed by a loss of selectivity.Correspondingly, the side reaction increased and the 1PO2 selectivity decreased.When the catalyst loading was increased from 450 L to 700 L, then PO concentration decreased from 4.21% to 0.40%.Therefore, a catalyst loading of 550 L on each reactive tray was finally selected.
Fig.5. Effects of the catalyst load on the PO concentration in the reflux liquid and 1PO2 selectivity.Other conditions of RD column: P=0.8 MPa, N=42, NFPO=30,nMeOH/nPO=1.57, BR=3.0.
3.3.7.EffectsoffeedlocationforMScolumn
Table 5 shows the effects of the feed location of MS column on the MS and RD column performance.As can be seen in Table 5,when the PO feed location was moved down along column from 4 to 8, the MeOH recovery was increased from 98.61% to 99.99%.For RD column, the increase in the feed location increased top PO concentration, selectivity and reboiler duty because of the higher molar ratio of MeOH/PO.When the feed location was moved down to 12,the all parameters were unchanged.According to the analysis, the PO feed location was set at stage 8.
3.4.Near-optimal single column RD process
Based on the results of simulation studies following the sensitivity analysis, the appropriate operating parameters for the RD process were identified as shown in Fig.1 and Table 6.The vapor from the top of RD column was sent to the reboiler of MS column as its heat steam,and was partially condensed by an auxiliary condenser at the same time.In this section,the near-optimal design of the single-column RD process is discussed and analyzed,aiming to provide a base case to verify the economic feasibility of alternative heat-integrated process (see Section 5).
Table 5Effects of the feed location on the MeOH recovery in the MS column.P=0.1 MPa, RR=2.4, N=30
Table 6Simulation results obtained using the final design
Fig.6 displays the composition profile in the liquid phase of RD column under the specified operating conditions.In the reactive section(from 2rd to 32nd stages),MeOH is the predominant component (more than 90%), the PO and 1PO2 concentration went through a maximum in the feed location, and side products are almost invisible.Based on this, the concept of ‘‘internal recycle of reactants” in the reactive/catalytic distillation process was suggested [10].Even though the fresh feed molar ratio of MeOH/PO of 1.47:1 is low, a ratio more than 70:1 within the reactive stages is obtained to enhancing the selectivity of 1PO2,because the unreacted MeOH is forced to recycle back internally into the reaction stages.In the stripping section,the concentration of MeOH reaches its peak in stage 40 and sharply decreases between stages 40 and 44.These results obtained from the 1PO2 are the heaviest component accumulates and dilute of the system under the effects of separation.
The stage temperature of the column was determined from the competing effect between the component.As shown in Fig.7, the smooth temperature profiles above stage 30 reflected a slight change in the composition of the mixture in this section, while the temperature increased rapidly below stage 30 because of a sharp increase of 1PO2 (the heavy enrichment) concentration under the effects of separation.In addition, the temperature of the reactive zone(the 2nd to 32nd stages)was in accordance with the temperature range of 100–150 °C, which is satisfied the maximum allowable temperature limit of the catalyst.
Fig.8 shows the reaction rate profiles along the RD column,indicating that the reactions mainly proceed between stages 2 and 32.The fastest reaction rate was observed at stage 32.However comparatively slow reaction rate was seen at other stages because the PO concentration close to the feed stage is the largest.
The simulations and analysis mentioned above have revealed that integrating chemical reactions and product separation into an RD column for PGME production offers several benefits due to process intensification.According to the results of the sensitive analysis, a feasible operating parameter of the reactive distillation column has been acquired to minimize reboiler duty.However,it is a challenge to further reduce energy consumption in the RD system.Note that, the choice of a suitable heat-integrated scheme needs to take the characteristics of the specific system into consideration.For example, the mechanical vapor recompression heat pump (VRC), pressurizing the overhead vapor at a higher temperature by a compressor supply heat required to the reboiler, is not suitable for this system based on two aspects.One factor is the economy of process.Because the potential of energy-savings using heat pump technology is for the separation of the close-boiling or lower relative volatilities mixtures, a small compression ratio and compressor duty is required.However,as indicated in Fig.1,a temperature difference between the top and bottom of the present RD column is more than 40 °C, which requires a higher compressing ratio or a higher electrical power of quality will be required,which brings the more capital cost.Another limitation is the danger of this process.Because of the characteristics of PO (a flammable and explosive gas), the potential hazard of the PO exposed in a compressor will limit application for the mechanical vapor recompression process.Considering the characteristics of the reaction system under study, it is suitable for the application of a MED scheme to achieve further energy savings.
4.1.Process design and optimization
To improve the economic benefits of PGME production by RD,a heat-integrated double-effect reactive distillation (DERD) process is proposed.According to different feed location,the design configuration can be classified into three types:light-split-reverse (LSR),light-split-forward (LSF), and feed-split (FS).As for the FS design configuration,it is the one that split the feed into two feed streamsand then go into both HP and LP columns [25].Fig.9 shows the basic concept and schematically proposed the DERD system, the design flowsheet is set to FS,has been used frequently in industry.For the DERD system, the original RD column is improved to a high-pressure reactive distillation (HPRD) column and a lowpressure reactive distillation (LPRD) column, which are thermally linked with each other and operated at different pressures so that the overhead vapor from the HPRD column acts as a heat source for the LPRD column.For the sake of simplicity, the following conditions and limitations were applied in the design and simulation of the DERD process.
Fig.6. Composition profiles in the liquid phase along the RD column.
(1) The reactants(PO and MeOH)were fed into the two columns in split-flow mode,i.e.,a parallel-flow double-effect RD process, as indicated in Fig.9.
Fig.7. Temperature profile along the RD column.
Fig.8. Profiles of component generation rates along the RD column.
(2) The feed flow rate of fresh PO and MeOH were the same as the single-column RD process as shown in Fig.9.It was also assumed in this design phase that the PO feed flow rate to each RD column is equal(i.e.,2500 kg∙h-1)for simplification purpose.In such arrangement,the catalyst loadings on each stage could be determined with a volume of 300 L,according to the PO feed flow rate as discussed in Section 3.2.Another specification for the design of double-effect reaction distillation process are: the number of total stages of RD column and MS column, the feed locations of the RD columns and MS column and the pressure of the MS column, all of which are the same as those in the single-column RD process.
Fig.9. Process diagram for the double-effect reactive distillation process.
(3) The utility conditions and the minimum temperature difference for heat exchange were the same as those of the single RD process, as detailed in Table 2.A minimum approach temperature (Δtmin) of 10 °C is specified for heat transfer from the condenser of HPRD column to the reboiler of LPRD and MS column.The temperatures of the reactive stages in each column were also limited to 100–150 °C based on the previous kinetic investigations.
(4) For comparison purposes,the maximum concentration of PO in the reflux liquid in each column was constrained to be less than 0.5% by mass.In addition, an overall selectivity for 1PO2 greater than 95% calculated from the mixed bottom products of two columns was set as a constraint, despite the different selectivity obtained in each column.
Based on the specified conditions and parameters given above,the adjustable parameters for designing DERD system can be determined through a free degree analysis.Specifically,for RD columns,they are operating pressures,MeOH/PO molar feed ratio and boilup ratio; For MS column, they are reflux ratio and distillate flow rate,the latter is also the circulating stream flow rate between two columns.Modification of these variables will be performed in order to obtain a feasible DERD process design.The implementation of heat integration for the DERD process is based on the appropriate redistribution of energy and the adjustment of operating conditions in each column.The feasibility of the heat integration scheme is based on these three factors.First,a minimum temperature difference between the top of HPRD column and the bottom of LPRD and MS column of 10°C is required to ensure heat transfer between the two columns.Second, the overhead vapor from the HPRD column is used to provide the total heat required in the LPRD and MS column.Third, the process constraints and limitations described in the previous section must be satisfied.Based on this, a procedure to perform suggested the heat-integrated DERD process was developed, as shown in Fig.10.First, a double-column RD flowsheet without heat integration was employed, which RD column both operated at the parameters as the same as the single-column RD design given in Fig.1.Second, a sequential iterative design procedure was performed by varying the operating pressure (PLPandPHP), MeOH recycling flow rate (D), boilup ratio (BRLPand BRHP)and reflux ratio(RRMS) to satisfy the feasible heat integration conditions and the specified purity requirements.Briefly, thePLPandPHPare adjusted to vary the temperature and reaction rate of the tray; The BRLPand BRHPare adjusted to change the heat duty in each unit together to improve the 1PO2 selectivity; TheDis adjusted to alter the temperature profiles and the 1PO2 selectivity in each RD columns(which is particularly sensitive to the reboiler temperature of LPRD column).
Fig.10. Simplified flowchart of the simulation and design for DERD process.
4.2.DERD process evaluation
The main results of the simulation of the heat-integrated DERD process are summarized in Fig.9 and Table 6.The overall condensing duty of overhead vapor from HPRD(saturated liquid reflux)and the heat duty required in the LPRD and MS column reboiler were matched, which indicate that heat transfer between the two columns is feasible; that is, the overhead vapor of the HPRD column can be totally used as the heat source for the reboiler of the LPRD and MS column without requiring an auxiliary condenser.Furthermore,the energy consumption of the DERD process(1.56 MW)was lower than that of the RD process (3.00 MW), which indicated DERD can reduce total energy requirement by 48.0%.In addition,the PO concentrations in the reflux liquid in both columns were lower than 0.5% by mass, indicating that the proposed DERD process can achieve the complete conversion of PO.
Under the identified operating conditions given in Table 6, the differences of parameter profiles along with the HPRD and LPRD are further illustrated.Fig.11 gives the MeOH composition profile in the liquid phase of the two RD column in DERD process.MeOH mass fraction in the stripping section of the LPRD column is higher than that of HPRD column because of a higher feed molar ratio in the LPRD.Fig.12 shows the temperature profile along the column.The stage temperature of the HPRD column is always higher than that of the LPRD because the pressure of the HPRD column is higher than that of LPRD column.Fig.13 displays the reaction rate profiles of PO along with the LPRD and HPRD columns.
Based on the results mentioned above, the use of heatintegrated DERD processes for the production of PGME has great potential for reducing operating costs.However, because additional equipment has to supplement in a DERD system, an economic assessment in terms of the total annual cost (TAC) is essential.The TAC can be calculated using Eq.(11).
Fig.11. MeOH profile in the liquid phase along the HPRD and LPRD columns.
Fig.12. Temperature profiles along the HPRD and LPRD columns.
The TAC consists of annual operating costs (AOC) and the total capital cost(TCC).The AOC includes steam and cooling water costs.The TCC includes the column cost and the heat exchanger(reboiler and condenser) cost.The costs of the other equipment such as pumps and pipes were ignored because these values are much lower than the main costs[26].The results of the economic calculations for the two proposed processed were summarized in Table 2.
From Table 6,the DERD process exhibits much better economic performance compared with RD process.The total operating cost of DERD process is 7.51×105USD∙a-1,which is 25.3%lower than that of the RD process.This is because for the DRED process, the heat input for the reboilers in the LPRD column and MS column completely relies on the overhead vapor of the HPRD column, so that the operating cost only contains the steam cost of HPRD column and the cooling water cost of LPRD and MS column.Meanwhile,for the single-column RD process, the overhead vapor of RD column cannot be totally used as the heat source for the reboiler of the MS column.This is the main reason for the decrease in the operating cost of the DERD process.In addition, the total capital cost of DERD process of 7.56 × 105USD∙a-1is higher than that of RD process as a result of the cost of the additional RD column,which increased by 8.9%.Despite the disadvantages of the additional column cost, the TAC of the DERD process is still less than that of the RD process due to the significant decrease of AOC,resulting in a TAC saving of 30.7%.In summary, the DERD process for PGME production is feasible energetically and economically.
This work proposed an energy-saving DERD process for PGME production from PO and MeOH.A near-optimal operating conditions for the single-column RD process, including pressure, molar feed ratio of MeOH to PO,boilup ratio,total number of trays,catalyst loading and feed location, were obtained based on sensitivity analysis.The results showed that the RD process can achieve complete conversion of PO and 95.0% selectivity of PGME, which identify the feasibility and advantage of the proposed RD process for PGME production.Then, the RD process combined with doubleeffect distillation technology was developed to further improved economic performance of RD process, and the process design of the DERD process with heat integration was proposed.The results indicated that DERD process greatly reduces the operating cost by 25.3% due to complete utilization of vapor steam of high-pressure reactive distillation column, even though the total capital cost of DERD process is 8.9%higher than that of the RD process.As a result,the DERD process reduces the TAC by 30.7%,compared with the RD process.The proposed approach shows promising applications and great potential in savings in TAC.For the proposed configurations,it is also necessary to analyze environmental, control and safety,which will be investigated in future work.
Fig.13. Profiles of PO reaction rates along the HPRD and LPRD columns.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was supported by the National Nature Science Foundation of China (21878315 and 21808223), National Key Research and Development Program of China (2017YFA0206803), Innovation Academy for Green Manufacture, CAS, (IAGM2020C17), K.C.Wong Education Foundation (GJTD-2018-04).The authors are grateful for the assistance from teachers Hui Wu, Ling Wang and Na Zhou of Analysis and Test Center, Institution of Process Engineering, Chinese Academy of Sciences.
本文来源：http://www.zhangdahai.com/shiyongfanwen/qitafanwen/2023/0727/631075.html
