- Email: [email protected]
DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE AND METHANOL IN AQUEOUS SOLUTIONS
Proceedings, 10th of IFAC International Symposium on Advanced Control Chemical Processes Proceedings, 10th of IFAC International Symposium on Advanced Control Chemical Processes Shenyang, Liaoning, China, July 25-27, 2018 online Proceedings, 10th 10th IFAC IFAC International International Available Symposium on at www.sciencedirect.com Proceedings, Symposium on Advanced Control of Chemical Processes Shenyang, Liaoning, China, July 25-27, 2018 Proceedings, 10th of IFAC International Symposium on Advanced Control Chemical Processes Advanced Control of China, Chemical Processes Shenyang, Liaoning, July 25-27, 2018 Advanced Control of China, Chemical Processes Shenyang, Shenyang, Liaoning, Liaoning, China, July July 25-27, 25-27, 2018 2018 Shenyang, Liaoning, China, July 25-27, 2018
ScienceDirect
IFAC PapersOnLine 51-18 (2018) 578–583 DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL AND METHANOL IN AQUEOUS SOLUTIONS ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE AND METHANOL IN AQUEOUS SOLUTIONS ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE AND METHANOL IN AQUEOUS SOLUTIONS AND METHANOL IN AQUEOUS SOLUTIONS Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien AND METHANOL IN AQUEOUS SOLUTIONS AND METHANOL IN AQUEOUS SOLUTIONS Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien
Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien Taiwan Department of Chemical Engineering, National Taipei 106, TAIWAN Zi-Jie Ai, Chuan-Yi Chuan-Yi Chung andUniversity, I-Lung Zi-Jie Ai, Chung and I-Lung Chien Chien Taiwan Department of Chemical Engineering, National University, Taipei 106, TAIWAN Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien Department of Chemical Engineering, National Taiwan University, Taipei 106, 106, TAIWAN TAIWAN Department of of Chemical Engineering, Engineering, National Taiwan Taiwan University, Taipei Department Department of Chemical Chemical Engineering, National National Taiwan University, University, Taipei Taipei 106, 106, TAIWAN TAIWAN
Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation Abstract: Poly(oxymethylene) dimethylsolvents ethers (OME) are widely used in the reduction of sootgas. formation in diesel engines, and also as physical for absorbing carbon dioxide from natural In this Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural gas. In this Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation work, a new production process of OME is designed and rigorously studied. In the conventional process, Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural In this Abstract: Poly(oxymethylene) dimethyl ethers (OME) arerigorously widely used in theIn reduction of sootgas. formation work, a new production process of OME is designed and studied. the conventional process, in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural gas. this multiple process sections areasrequired tosolvents first separately produce trioxane andInmethylal, and then toIn react in diesel engines, and also physical for absorbing carbon dioxide from natural gas. In this work, a new production process of OME is designed and rigorously studied. the conventional process, in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural gas. Inreact this multiple process sections are required to first separately produce trioxane and methylal, and then to work, a new production process of OME is designed and rigorously studied. In the conventional process, of these two intermediates to produce OME. In this paper, the plant-wide process via a much simply work, a new production process of OME is designed and rigorously studied. In the conventional process, multiple process sections are required to first separately produce trioxane and methylal, and then to react work, a new production process of OME is designed and rigorously studied. In the conventional process, of these two intermediates torequired produceto OME. In methanol this paper, plant-wide process viaand a much multiple process are first separately produce trioxane and methylal, then tosimply react direct synthesis ofsections OME from formaldehyde and in the aqueous solutions is newly developed. The multiple process sections are required to first separately produce trioxane and methylal, and then react of these two intermediates produce In this paper, the plant-wide via aa much multiple process sections areto required toOME. first and separately produce trioxane andprocess methylal, and then to tosimply react direct synthesis of OME from formaldehyde methanol in aqueous solutions is newly developed. The of these two intermediates to produce OME. In this paper, the plant-wide process via much simply overall process includes an upstream section and a downstream product purification section. After the of these two intermediates to produce OME. In this paper, the plant-wide process via aa much simply direct synthesis of OME from formaldehyde and methanol in aqueous solutions is newly developed. The of these two intermediates to produce OME. In this paper, the plant-wide process via much simply overall process includes an upstream section and a downstream product purification section. After the direct synthesis synthesis of OME OMEoffrom from formaldehyde and methanol methanol in aqueous aqueous solutions is newly newly developed. developed. The steady-state simulation the process is established, the process flowsheet is optimized by minimizing direct of formaldehyde and in solutions is The overall process includes an upstream section and a downstream product purification section. After the direct synthesis of OME from formaldehyde and methanol in aqueous solutions is newly developed. The steady-state simulation of the process is established, the process flowsheet is optimized by minimizing overall process includes an upstream section and aa downstream downstream product purification section. After the total annual costincludes (TAC). The dynamics and control of the proposed OME production process is also overall process an upstream section and product purification section. After the steady-state simulation of the process is established, the process flowsheet is optimized by minimizing overall process includes upstream section and a downstream product purification section. After the total annual cost (TAC). The dynamics andstrategy control ofdeveloped the proposed OME process is also steady-state simulation ofan the process is established, established, the process flowsheet isproduction optimized by to minimizing investigated in this study. A proper control is for the optimized process reject the steady-state simulation of the process is the process flowsheet is optimized by minimizing total annual cost (TAC). dynamics and control of the proposed OME process is also steady-state simulation of The theproper process is established, the process flowsheet isproduction optimized by to minimizing investigated in this study. A control strategy is developed for the optimized process reject the total annual cost (TAC). The dynamics and control of the proposed OME production process is also process disturbances. Copyright © 2018 IFAC total annual cost (TAC). The dynamics and control of the proposed OME production process is also investigated in this study. A proper control for the optimized process to reject the total annual cost (TAC). The dynamics andstrategy controlis ofdeveloped the proposed OME production process is also process disturbances. Copyright © 2018 IFAC investigated in this study. A proper control strategy is developed for the optimized process to reject the investigated in this study. A proper control strategy is developed for the optimized process to reject the process disturbances. Copyright © 2018 IFAC investigated in this study. A proper control strategy is developed for the optimized process to reject the © 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Keywords: Computer-aided design; Process models; Distillation columns; Optimization; Computer process disturbances. Copyright © 2018 IFAC process disturbances. Copyright © 2018 IFAC Keywords: Computer-aided design; Process models; Distillation columns; Optimization; Computer process disturbances. Copyright © 2018 IFAC simulation; Process control design; Process models; Distillation columns; Optimization; Computer Keywords: Computer-aided simulation; Process control design; Keywords: Computer-aided Process models; Distillation columns; Optimization; Computer Keywords: Computer-aided simulation; Process control design; Keywords: Computer-aided design; Process Process models; models; Distillation Distillation columns; columns; Optimization; Optimization; Computer Computer simulation; Process control simulation; Process control simulation; Process control
of the overall process with upstream and downstream 1. INTRODUCTION of the overall processdesign with flowsheet upstream isanddetermined downstream sections. The optimal by 1. INTRODUCTION of the overall process with upstream and downstream sections. The optimal design flowsheet is determined by of the overall overall process with system. upstreamSection and 3downstream downstream 1. INTRODUCTION minimizing TAC of the overall shows the of the process with upstream and 1. INTRODUCTION sections. The optimal design flowsheet is determined by of the overall process with upstream and downstream 1. INTRODUCTION Diesel fuel is highly minimizing TAC of the overall system. Section 3 shows the sections. The optimal optimal design flowsheet is determined by demanded in recent years. However, sections. overall control strategy development of thisdetermined system. The 1. INTRODUCTION The design flowsheet is by Diesel fuel is highly minimizing TAC of overall system. Section 33 shows the sections. The optimal design flowsheet is determined by demanded in recent years. However, overall control strategy development of this system. The there are many drawbacks minimizing TAC of the thesimulations overall system. Section shows the for conventional diesel fuel, such closed-loop dynamic of the proposed control minimizing TAC of the overall system. Section 3 shows the fuel is highly demanded in recent years. However, overall control strategy development of this system. The Diesel minimizing TAC of the overall system. Section 3 shows the there are many drawbacks for conventional diesel fuel, such closed-loop dynamic simulations of the proposed control Diesel fuel is highly demanded in recent years. However, overall control strategy development of this system. The as smog production and low combustion efficiency. In order Diesel fuel is demanded in years. However, strategy in the face of simulations various disturbances are system. illustrated in overall control strategy development of The there are many drawbacks for conventional diesel fuel, such closed-loop dynamic of the proposed control Diesel fuel is highly highly demanded in recent recent years. However, overall control strategy development of this this system. The as smog production and low combustion efficiency. In order strategy in the face ofsome various disturbances are are illustrated in there are many drawbacks for conventional diesel fuel, such closed-loop dynamic simulations of the proposed control to solve these problems, blending an additive with there are many drawbacks for conventional diesel fuel, such this section. Finally, concluding remarks drawn in closed-loop dynamic simulations of the proposed control as smog and low efficiency. In order strategy in the face of various disturbances are illustrated in there are production many drawbacks forcombustion conventional diesel fuel, such closed-loop dynamic simulations of the proposed control to solve these problems, blending an additive with this section. Finally, some concluding remarks are drawn as smog production and low combustion efficiency. In order strategy in the face of various disturbances are illustrated in conventional diesel fuels may be a feasible way. Among as smog production and low combustion efficiency. In order Section 4. strategy in the face of various disturbances are illustrated in to solve these problems, blending an additive with this section. Finally, concluding remarks drawn as smog and combustion efficiency. In order face ofsome various disturbances are are illustrated in conventional diesel fuelslow may be a feasible way.dimethyl Among Section 4.in the to solveproduction these problems, blending an additive additive with strategy this section. Finally, some concluding remarks are drawn in different diesel additives, poly(oxymethylene) to solve these problems, blending an with this section. Finally, some concluding remarks are drawn in conventional diesel fuels may be a feasible way. Among Section 4. to solve these problems, blending an additive with this section. Finally, some concluding remarks are drawn in different diesel additives, poly(oxymethylene) dimethyl conventional diesel fuels may be beones a feasible feasible way. Among Section 4. ethers (OME) are the promising that can enhance the conventional diesel fuels may a way. Among 2. PROPOSED PROCESS DESIGN Section 4. different diesel additives, poly(oxymethylene) dimethyl conventional diesel fuels maypoly(oxymethylene) beones a feasible way. Among Section 4. ethers (OME) are the promising that can enhance the 2. PROPOSED PROCESS DESIGN different diesel additives, dimethyl oxygenation ofarediesel fuel (Pellegrini, et al., enhance 2012). The different diesel additives, poly(oxymethylene) dimethyl ethers (OME) the promising ones that the 2. PROCESS DESIGN different diesel additives, poly(oxymethylene) dimethyl oxygenation of diesel fuel (Pellegrini, et can al., 2012). The ethers (OME) are the promising ones that can enhance the 2. PROPOSED PROPOSED PROCESS DESIGN properties of OME are close to conventional diesel fuel, so The thermodynamic ethers (OME) are the promising ones that can enhance the behaviour of this system involving 2. PROPOSED PROCESS DESIGN oxygenation of diesel fuel (Pellegrini, et al., 2012). The ethers (OME) are the promising ones that can enhance the 2. PROPOSED PROCESS DESIGN properties of OME are fuel close (Pellegrini, to conventional diesel fuel,fuel so formaldehyde, The thermodynamic behaviour ofis complicated this system because involving oxygenation of diesel et al., 2012). The there is no need of modifications for those original diesel oxygenation of diesel fuel (Pellegrini, et al., 2012). The methanol and water of properties of OME are close to conventional diesel fuel, so The thermodynamic behaviour of this system involving oxygenation of diesel fuel (Pellegrini, etoriginal al., 2012). The there is no need of modifications for those diesel fuel formaldehyde, methanol and water is complicated because of properties of OME are close to conventional diesel fuel, so The thermodynamic behaviour of this system involving engines (Burger, et al., 2010). For OMEs chain outside the properties of OME are close to conventional diesel fuel, so oligomer reactions among them. Formaldehyde will The thermodynamic behaviour of this system involving there is no need of modifications for those original diesel formaldehyde, methanol and water complicated of properties of OME are close toFor conventional diesel fuel,fuel so The thermodynamic behaviour ofis thisFormaldehyde system because involving OMEs chain outside the engines (Burger, et al., 2010). oligomer reactions among them. will there is no need of modifications for those original diesel fuel formaldehyde, methanol and water is complicated because of range of 3 to 5, there may be some safety concerns, because there is need ofetmodifications for those diesel fuel oligomerize with methanol and water, respectively. The methanol and is complicated because of engines (Burger, al., 2010). For chain outside the formaldehyde, oligomer among them. Formaldehyde there isofno no need modifications for OMEs those original original diesel fuel methanol and water water is complicated becausewill of range 3points to 5, of there may be some safety concerns, because oligomerizereactions with methanol and water, respectively. The engines (Burger, et al., 2010). For OMEs chain outside oligomer reactions among them. Formaldehyde will the flash of shorter OME areconcerns, too low, and the formaldehyde, n (n<3) engines (Burger, et al., 2010). For OMEs chain outside oligomers of formaldehyde and water are called oligomer reactions among them. Formaldehyde will range of 3 to 5, there may be some safety because oligomerize with methanol and water, respectively. The engines (Burger, et al., 2010). For OMEs chain outside oligomer reactions among them. Formaldehyde will (n<3) are too low, and the the flash points of shorter OME n oligomers of formaldehyde and water are called range of 3 3ofto tolonger 5, there there may be some some safety concerns, because because oligomerize with methanol and respectively. viscosity OME are(n<3) too high. n (n>5) range of 5, may be safety concerns, poly(oxymethylene) glycols (MG HO-(CH the n, water, 2O)n-H) and The oligomerize with methanol and water, respectively. The the points shorter OME are too low, and the n too oligomers of water are called range of 3of tolonger 5, of there may be some safety concerns, because oligomerize withformaldehyde methanol and nand water, respectively. The are high. viscosity OME n (n>5) poly(oxymethylene) glycols (MG , HO-(CH -H) and the the flash flash points of shorter OME (n<3) are too low, and the 2O)nare n oligomers of formaldehyde and water called the flash points of shorter OME (n<3) are too low, and the of formaldehyde and methanol are n too high. oligomers of formaldehyde and water are and called viscosity longer n (n>5) poly(oxymethylene) glycols (MG ,, HO-(CH the the flash of points of OME shorter OMEare are too low, and the oligomers nand 2O)nare n (n<3) of formaldehyde formaldehyde water called of and methanol are and viscosity longer OME (n>5) are too high. poly(oxymethylene) glycols (MG HO-(CH O)n-H) -H) the There areof two routes fornn the manufacture of OMEs (Burger, poly(oxymethylene) n 2 viscosity of longer OME (n>5) are too high. hemiformals n, HO-(CH 2 O)and n-CHthe 3). glycols (MG , HO-(CH -H) n(HF 2O)n oligomers of formaldehyde and methanol are called viscosity of longer OME (n>5) are too high. n the manufacture of OMEs (Burger, poly(oxymethylene) glycols (MG , HO-(CH the There are two routes for n(HF 2O)n-H) hemiformals O)and n, HO-(CH 2are n-CH 3). oligomers of reactions formaldehyde and methanol called et al., 2013; Schmitz, et al., 2016a). The most common route These oilgomer occur in the system everywhere and oligomers of formaldehyde and methanol are called There are two routes for the manufacture of OMEs (Burger, poly(oxymethylene) hemiformals (HF O) n, HO-(CH 2are n-CH 3). oligomers of formaldehyde and methanol called et al., 2013; Schmitz, et al., 2016a). The most common route These oilgomer reactions occur in the system everywhere and There are two routes for the manufacture of OMEs (Burger, poly(oxymethylene) hemiformals (HF , HO-(CH O) -CH is through methanol and formaldehyde forming of route two poly(oxymethylene) n, HO-(CH2 2supplies n-CH3 3). There are two routes for the manufacture of OMEs (Burger, the oligomers are hemiformals notoccur stable. AspenPlus (HF O) ).a n n et al., 2013; Schmitz, et al., 2016a). The most common These oilgomer reactions in the system everywhere and There are two routes for the manufacture of OMEs (Burger, poly(oxymethylene) hemiformals (HF , HO-(CH O) -CH is through methanol and formaldehyde forming of two n 2supplies n 3). the oligomers are not stable. AspenPlus a et al., 2013; Schmitz, et al., 2016a). The most common route These oilgomer reactions occur in the system everywhere and intermediates, trioxane and methylal, then reacting to the et al., 2013; Schmitz, et al., 2016a). The most common route Formaldehyde package which describes the ternary mixture These oilgomer reactions occur in the system everywhere and is through formaldehyde forming of two the oligomers are not stable. supplies et al., 2013; methanol Schmitz, etand al., 2016a). Thethen most common route These oilgomerpackage reactions occur in theAspenPlus system everywhere andaa intermediates, trioxane and methylal, reacting to the Formaldehyde which describes the ternary mixture is through methanol and formaldehyde forming of two the oligomers are not stable. AspenPlus supplies target OME product (named as route 1). The alternative route is through methanol and formaldehyde forming of two with UNIFAC package model to describe vapor-liquid and vapor-a the oligomers are stable. AspenPlus supplies intermediates, trioxane and methylal, then reacting the Formaldehyde which describes the ternary is through methanol and formaldehyde ofto two the oligomers are not not stable. AspenPlus supplies a target OME from product (named as route Theforming alternative route with UNIFACequilibrium model to describe vapor-liquid andmixture vaporintermediates, trioxane and methylal, then reacting to the Formaldehyde package which describes the ternary mixture is directly formaldehyde and1). methanol in aqueous intermediates, trioxane and methylal, then reacting to the liquid-liquid and treat oligomers as Formaldehyde package which describes the ternary mixture target OME product (named as route 1). The alternative route with UNIFAC model to describe vapor-liquid and vaporintermediates, trioxane and methylal, then reacting to the Formaldehyde package which describes ternary mixture is directly from formaldehyde and methanol in aqueous liquid-liquid equilibrium and treat the oligomers as target OME productas(named (named as 2). routeHowever, 1). The The alternative alternative route with UNIFAC model to describe vapor-liquid and vaporsolution (named routeas the complex target OME product route 1). route electrolytes in equilibrium the feature of Chemistry. In thisoligomers research, the with UNIFAC model to describe vapor-liquid and is directly from formaldehyde and aqueous liquid-liquid and treat the as target OME product (named as 2). route 1).methanol The alternative route electrolytes with UNIFAC model to of describe vapor-liquid and vaporvaporsolution (named as route However, thein in the feature Chemistry. In this research, the is directly from formaldehyde and methanol in complex aqueous liquid-liquid equilibrium and treat the oligomers as azeotropes among the reactants and the products make it is directly from formaldehyde and methanol in aqueous simulations follow the idea of Formaldehyde package and liquid-liquid equilibrium and treat the oligomers as solution as route 2). However, the electrolytes in the feature of Chemistry. In this research, the is directly(named from formaldehyde and methanol in complex aqueous liquid-liquid equilibrium and treat the oligomers as azeotropes among the reactants and the products make it simulations follow the idea of Formaldehyde package and solution (named as route 2). However, the complex electrolytes in the the feature of Chemistry. Chemistry. Inrelated this research, research, the difficult to get target OME 3-5 product. solution (named as route 2). However, the complex further modify the parameters with the literatures electrolytes in feature of In this the azeotropes among the reactants and the products make it simulations follow the idea of Formaldehyde package and solution (named as route 2). However, the complex electrolytes in the feature of Chemistry. In this research, the difficult to get target OME 3-5 product. further modify the parameters with the related literatures azeotropes among the reactants and the products make it simulations follow the Kuhnert, idea of of Formaldehyde package and and azeotropes among the reactants and (Schmitz, et al., 2016b; et al., 2006). follow the idea package difficult to get target OME 3-5 product. further modify the parameters with related literatures azeotropes among the reactants and the the products products make make it it simulations simulations follow the idea of Formaldehyde Formaldehyde package and Kuhnert, et al., the 2006). et al., 2016b; difficult get target product. 3-5 further modify the parameters with the related literatures In this to paper, the OME process design and control for the (Schmitz, difficult to get target OME 3-5 product. further modify the parameters with the related literatures (Schmitz, et al., 2016b; Kuhnert, et al., 2006). difficult get target 3-5 product. further modify the parameters with the related literatures In this to paper, the OME process design and control for the (Schmitz, et al., 2016b; Kuhnert, et al., 2006). production of OME via route 2 will be investigated. 3-5 In order et to al., realize theKuhnert, simulations result easily, overall (Schmitz, 2016b; et In this paper, the process design and control for the (Schmitz, 2016b; et al., al., 2006). 2006). via route 2 will be investigated. production of shows OME 3-5 In order et to al., realize theKuhnert, simulations result easily, overall In this paper, the process design and control for the Section 2 the proposed design flowsheet In this paper, the process design and control for the composition is used instead true composition. True production of OME via route be investigated. order to realize the simulations result easily, overall In this paper, the 3-5 process design22 will anddesign control for the In Section 2 shows the proposed flowsheet True composition is used instead true composition. production of OME via route will be investigated. 3-5 In order to realize the simulations result easily, overall production of OME via route 2 will be investigated. 3-5 the In order to realize the simulations result easily, overall Section 2 shows proposed design flowsheet composition is used true composition. True production of shows OME3-5 the via route 2 willdesign be investigated. In order to realize the instead simulations result easily, overall Section 2 proposed flowsheet composition is used instead true composition. Section 2 shows the proposed design flowsheet True Copyright ©22018shows IFAC 572 composition is used instead true composition. True Section the proposed design flowsheet 2405-8963 © 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. composition is used instead true composition. True Copyright © 2018 IFAC 572
Peer review©under of International Federation of Automatic Copyright 2018 responsibility IFAC 572Control. Copyright © 572 10.1016/j.ifacol.2018.09.362 Copyright © 2018 2018 IFAC IFAC 572 Copyright © 2018 IFAC 572
2018 IFAC ADCHEM Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
composition presents all the components including the unstable oligomers. Overall composition means the unstable HFs and MGs decomposed into formaldehyde, methanol and water, and there are mathematical relationships between true and overall composition: n~FA nFA i nMGi i nHFi
(1)
n~MeOH nMeOH nHFi
(2)
n~H 2 O nH 2 O nMG i
(3)
adopted from Zhang, et al., (2016) and it is also the stoichiometry ratio to form OME3. The CSTR is packed with catalyst (amberlyst 46) and operated at constant temperature, 363.15 K, and constant pressure of 3 atm, to reach higher conversion and prevents vaporization. The packing voidage is assumed at 50%. The reactions are highly exothermic, so an internal heat exchanger is required to supply enough heat transfer area (Luyben, 2016). Luyben (2016) proposed an equation to estimate the volume of internal heat exchanger and a practical limit that the volume fraction of internal heat exchanger for the whole reactor cannot be higher than 50%. From Fig. 1 and Fig. 2 the residence time at 1.5 min gives 0.99 equilibrium conversion condition, but the internal heat exchanger occupies over 50% volume of the reactor. Conservatively, 15 min residence time is determined and it can reach 0.999 equilibrium conversion condition.
The overall amounts of formaldehyde equals to the true amount of formaldehyde plus the summation of formaldehyde in MGs and HFs. Similarly, we can calculate the overall amounts of methanol and water. The overall amounts of OMEs equal to their true amount and then the overall amounts of total components are shown below: n~total n~FA n~MeOH n~H 2O n~OMEi
The specifications of the two following columns are consequently designed. The distillate (D1R) of the first column (C1) contains the light mixture including MeOH, MAL, OME2, and parts of water, so the top specification of impurity is 100 ppm OME3 and bottom specification is 100 ppm MeOH. The bottom flow (B1) of C1 is sent to the second column (C2) to separate and recycle the heavy mixture including longer OMEs and oligomers of FA and water from bottom (B2R) of C2. To maintain FA at bottom, a certain amount of water is necessary to form the heavy oligomers (MGs), and highly concentrated formaldehyde may cause some transportation problem. The bottom specification of C2 is 10 wt% water and the top is 1000 ppm FA. Under these specifications, the distillate (D2) of C2 is close to the azeotropic composition of water and OME3 and Fig. 3 shows the binary Txy behaviour.
(4)
The reactions for poly(oxymethylene) dimethyl ethers (OMEn, H3CO-(CH2O)n-CH3) synthesis taking place in the reactor can be seen as below: FA + MeOH HF1
(5)
MeOH + HF1 MAL(OME1) + H2O
(6)
MeOH + HFn OMEn + H2O ; 8>n>1
(7)
FA + OME1 OME2 + H2O
(8)
FA + OMEn-1 OMEn + H2O ; 8>n>1
(9)
579
The kinetics of all reactions were assumed to behave according to power law. The kinetics parameters from Schmitz, et al. (2015) are used in this study. The presence of water makes the yield of OME decrease and form complex azeotropes with OMEs. The proposed design of the OME production process is to produce high purity OME3-5 product (99.95 wt%) to match the diesel regulation. The overall process is divided into upstream section and downstream section. The upstream section includes a CSTR and two distillation columns, and the downstream section is a one-decanter-two-stripper system with a purge stream. Besides producing with the CSTR, recycling of the unreacted reactants (FA and MeOH), and shorter and longer OMEs is the main purpose of the two distillation columns in the upstream section. In downstream section, the OME product and high purity water are withdrawn from the bottom of two strippers.
Fig. 1. Relationship between equilibrium conversion condition and residence time.
The fresh feed is formalin and pure methanol. The formalin feed consists of 37 wt% formaldehyde and 63 wt% water and the flow rate is fixed at 4000 kg/hr. Adjusting fresh methanol flow rate makes the mass ratio of formaldehyde and methanol equal to 90:64 feeding to the CSTR. This feed ratio is
Fig. 2. Relationship between volume fraction of internal heat exchanger and residence time. 573
2018 IFAC ADCHEM 580 Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
440
respectively. The optimized total stages of WSTRP is 8 and total stages of OSTRP is 6, and the purge fraction is 10%. Comparing the effect of purge fraction on TAC with and without product loss, the result is shown in Fig. 5 which justifies the selection of purge fraction at 10%.
T-xy diagram for H2O/OME3
Temperature, K
x 1. 0 atm y 1. 0 atm 420
400
The product loss means the production difference between the simulating point with the lowest TAC and others simulating point. A larger purge fraction increases the product loss and a smaller fraction gives higher recycle flows which increase the energy consumption. The production difference needs to be concerned and the unit product cost is assumed to be 0.6148 $/kg (Schmitz, et al., 2016a).
380
360 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Liqu id/vapo r mass f raction, H2O
Fig. 3. Binary Txy behaviour of OME3 and water. The minimum azeotrope of OME3 and water locates on the liquid-liquid splitting region, so the one-decanter-twostripper system can be applied in downstream section. A small amount of lighter component (MeOH) in D2 may be accumulated in the system, so that a MeOH purge stream is necessary. The objective function of optimization is to minimize total annual cost (TAC) and iterative optimization procedure is followed to find the best set of design variables. The overall optimization is carried out sequentially through upstream to downstream section. The design variables in upstream section need to be optimized are: total stages of C1 and C2, feed stage of C1 and C2. The design variables in downstream section need to be optimized are: total stages of the two strippers, and the purge fraction. The total stages of C1 is fixed at 50, since the original optimization displays the total stages of C1 becomes extremely high and not practical at 95. Fixing NC1 at 50 only increases the TAC by 4.3% as compared with the impractical optimized TAC with extremely high C1 column.
Fig. 5. Minimized TAC at each fixed purge fraction. 3. CONTROL STRATEGY DEVELOPMENT After the optimized design has been developed, an investigation is made to develop to the proper overall control strategy of the system. The following design of control strategy is simulated by using Aspen Plus Dynamics. In this section the overall process is still simulated separately with upstream and downstream sections. The control strategy needs to deal with various disturbance changes with OME product remaining at high purity. The disturbances considered in this study include: throughput changes via feed
From Fig. 4 below, the final optimized number of total stages for C1 is 50 and total stages of C2 is 16. The optimized feed locations for C1 and C2 are at 6th stage and 5th stage,
Purge 320 K 43.45 kg/hr 0.509 H2O 130 ppm FA 0.020 MeOH 0.470 OME3 740 ppm OME4
320 K -0.25 MW
-20.12 MW
Methanol 363.15 K 3 atm 1045.52 kg/hr 1 MeOH
Formalin 363.15 K 3 atm 4000 kg/hr 0.63 H2O 0.37 FA
CSTR -5.74 MW VR 30.6 m3 VHX 3.49 m3 363.15 K 3 atm Reactor Feed 369.5 K 3 atm 58297.4 kg/hr 0.108 H2O 0.266 FA 0.189 MeOH 0.269 MAL 0.123 OME2 0.014 OME3 0.020 OME4 0.008 OME5 0.003 OME6 0.001 OME7 411 ppm OME8
1 atm
6
D1R 324.9 K; 1 atm 34683.8 kg/hr 0.055 H2O 0.287 MeOH 0.452 MAL 0.206 OME2 100 ppm OME3
RR 1.93 D 3.19 m
D2 371.5 K; 1 atm 5045.25 kg/hr 0.558 H2O 1000 ppm FA 183 ppm MeOH 0.434 OME3 0.007 OME4 10 ppm OME5
320 K -2.56 MW
-3.96 MW
Reactor Effluent 363.15 K; 3 atm 58297.4 kg/hr 0.113 H2O 0.240 FA 0.171 MeOH 0.269 MAL 0.123 OME2 0.052 OME3 0.021 OME4 0.008 OME5 0.003 OME6 0.001 OME7 410 ppm OME8
20.21 MW
1 atm
50 B1 403.3 K; 1.39 atm 23613.5 kg/hr 0.198 H2O 0.593 FA 100 ppm MeOH 0.127 OME3 0.051 OME4 0.020 OME5 0.007 OME6 0.003 OME7 0.001 OME8
5
C2
15
RR 1.06 D 1.39 m B2R 419.4 K 1.07 atm 18568.6 kg/hr 5.18 MW 0.1 H O 2 0.754 FA 16 77 ppm MeOH 0.044 OME3 0.063 OME4 0.025 OME5 0.009 OME6 0.004 OME7 0.001 OME8
Fig. 4. Optimized process design flowsheet. 574
OR 377 K 434.45 kg/hr 0.509 H2O 130 ppm FA 0.020 MeOH 0.470 OME3 740 ppm OME4
Duty=0 WR 376.5 K 6042.19 kg/hr 0.54 H2O 53 ppm FA 0.008 MeOH 0.451 OME3 623 ppm OME4
C1
49
320 K -0.18 MW
Decanter
1.2 atm
WSTRP
7
D 0.97 m
2.76 MW 8
OP 320 K; 1 atm 2612.84 kg/hr 0.085 H2O 65 ppm FA 0.003 MeOH 0.904 OME3 0.008 OME4 11 ppm OME5 WP 320 K; 1 atm 8866.19 kg/hr 0.683 H2O 592 ppm FA 0.005 MeOH 0.309 OME3 0.002 OME4 3 ppm OME5
1.2 atm OSTRP
5
WOUT 380.6 K; 1.28 atm 2824 kg/hr 0.988 H2O 0.002 FA 0.004 OME3 0.006 OME4 8 ppm OME5
D 0.47 m
0.32 MW 6 OOUT 437.6 K; 1.26 atm 2178.39 kg/hr 420 ppm H2O 52 ppm FA 28 ppm MeOH 0.9903 OME3 0.0092 OME4 1 ppm OME5
2018 IFAC ADCHEM Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
581
flow rate and composition changes of formaldehyde in formalin. Before developing quality control loops, the inventory control loops have to be designed first. The liquid levels of reactor, decanter, reflux drums and sumps are controlled by manipulating their outlet flow. The temperature of reactor and inlet flow of decanter is controlled by manipulating cooling duty. The top pressures of column and stripper are controlled by manipulating the condenser duties and the top vapor flow, respectively.
Fig. 9. C2_RR open-loop test.
Open-loop and closed-loop sensitivity analyses are carried out to determine the locations of temperature control for each column as quality control loops. The open-loop tests are increasing and decreasing 0.1% manipulating variables, reboiler duty (Qr) and reflux rate (RR). The closed-loop tests are introducing composition disturbance to the system with assuming of perfect control is achieved. The ratio of FA and MeOH feeding to the CSTR is fixed. Fig. 10. C2_Qr open-loop test.
Figs. 6 to 11 are the result of upstream sensitivity tests, and Fig. 12 is the proposed upstream control structure. A notch locates on the stage 5 from the closed-loop test, and for the two manipulated variable of C1 (RR and Qr) stage 5 shows the largest open-loop sensitivity. Intuitively, fixing the reflux ratio and manipulating reboiler duty to control the temperature of stage 5 are determined.
Fig. 11. C2 closed-loop test. LC1 PC1
FC1
1 atm
TCr MeOH
6 X
Formalin
C1
Fig. 6. C1_RR open-loop test.
MeOH
FC2 sp X
FA
PC2
T5 T5C
LCr 49
LC3 1 atm
50
5
T4
T4C
C2
LC2
T15
T15C
15 16
LC4
Fig. 12. The proposed control structure of upstream section. There are two locations, stage 4 and 15, with smaller temperature deviation from the closed-loop test. The stage 15 and 16 give larger open-loop sensitivity of reboiler duty, so the stage 15 is controlled. The stage 4 shows enough openloop sensitivity from reflux rate, so the reflux rate is chosen to control the temperature of stage 4.
Fig. 7. C1_Qr open-loop test.
The dead time of 1 min for each temperature measurement is assumed in the simulation study. The upstream dynamic responses of important variables are shown in Fig. 13 and Fig. 14. The most important output variable is the MeOH containing in D2. Then the new steady state of the upstream section is considered as the disturbances introduced into the downstream section.
Fig. 8. C1 closed-loop test. 575
2018 IFAC ADCHEM 582 Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
Fig. 15 is the result of downstream sensitivity tests, and Fig. 16 is the proposed downstream control structure with the purge fraction is fixed at base case value. The temperature of stage 5 of OSTRP and the temperature of stage 7 of WSTRP are controlled for smaller closed-loop temperature difference and enough open-loop sensitivity. The downstream dynamic responses of important variables are shown in Fig. 17 and Fig. 18. This control strategy is able to reach a new steady state quickly with acceptable small purity deviations. The largest OME purity deviation is only 6×10-4% with positive throughput change.
Fig. 14. The upstream dynamic responses of the feed composition disturbances.
Fig. 13. The upstream dynamic responses of the throughput changes. Fig. 15. Results of the downstream sensitivity test. FC1
320 K TC1 sp FC2
X
0.1
Decanter Duty=0
TC3 320 K
TC2 320 K
OLC
PC2
WLC
PC1 1.2 atm
1.2 atm OSTRP
WSTRP
T5 T7
WT7C
7
OT5C
5 8
6
LC1
LC2
Fig. 16. The control structure of downstream section. 576
2018 IFAC ADCHEM Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
583
outlet stream of the process is the water by-product at high purity of 98.8 wt%. Another methanol purge stream is also necessary to prevent accumulation of light component in the downstream section. In the process optimization study, small loss of OME product in this purge stream is also included in the TAC calculations. Due to page limitation, a further energy-saving design of dividing-wall column to combine the two columns in the upstream section is not described in this paper The control structures of upstream and downstream sections are studied separately because there is no recycle stream from downstream to upstream. The proposed control strategy is able to hold product streams at high purity despite various feed flow rate and feed composition disturbance changes. REFERENCES
Burger, J.; Siegert, M.; Strofer, E.; Hasse, H. (2010). Poly (oxymethylene) dimethyl ethers as components of tailored diesel fuel: Properties, synthesis and purification concepts. Fuel, 89, 3315-3319. Burger, J.; Strofer, E.; Hasse, H. (2013). Production process for diesel fuel components poly(oxymethylene) dimethyl ethers from methane-based products by hierarchical optimization with varying model depth. Chemical Engineering Research and Design, 91, 2648-2662. Kuhnert, C., Albert, M., Breyer, S., Hahnenstein, I., Hasse, H., & Maurer, G. (2006). Phase equilibrium in formaldehyde containing multicomponent mixtures: experimental results for fluid phase equilibria of (formaldehyde+(water or methanol)+ methylal)) and (formaldehyde+ water+ methanol+ methylal) and comparison with predictions. Industrial & engineering chemistry research, 45(14), 5155-5164. Luyben, W. L. (2016). Economic trade-offs in acrylic acid reactor design. Computers & Chemical Engineering, 93, 118-127. Pellegrini, L.; Marchionna, M.; Patrini. R. (2012). Combustion behavior and emission performance of neat and blended polyoxymethylene dimethyl ethers in a light-duty diesel engine. SAE Technical Paper, 2012-011053. Schmitz, N., Burger, J., Ströfer, E., & Hasse, H. (2016a). From methanol to the oxygenated diesel fuel poly (oxymethylene) dimethyl ether: An assessment of the production costs. fuel, 185, 67-72. Schmitz, N., Burger, J., & Hasse, H. (2015) Reaction kinetics of the formation of poly (oxymethylene) dimethyl ethers from formaldehyde and methanol in aqueous solutions. Industrial & Engineering Chemistry Research, 54(50), 12553-12560. Schmitz, N., Friebel, A., von Harbou, E., Burger, J., & Hasse, H. (2016b) Liquid-liquid equilibrium in binary and ternary mixtures containing formaldehyde, water, methanol, methylal, and poly (oxymethylene) dimethyl ethers. Fluid Phase Equilibria, 425, 127-135. Zhang, X., Oyedun, A. O., Kumar, A., Oestreich, D., Arnold, U., & Sauer, J. (2016). An optimized process design for oxymethylene ether production from woody-biomassderived syngas. Biomass and Bioenergy, 90, 7-14.
Fig. 17. The downstream dynamic responses of the throughput changes.
Fig. 18. The downstream dynamic responses of the feed composition disturbances. 4. CONCLUSIONS In this study, the design and control of a simply OME production process via direct synthesis from formaldehyde and methanol in aqueous solutions has been investigated. To the best of our knowledge, no paper in open literature can be found to study the rigorous design and also control strategy of this process. The proposed design of the process contains upstream and downstream sections requiring only limited pieces of process equipments. High purity OME product at 99.95 wt% can be obtained from the bottom of the product stripper. The other 577
ScienceDirect
IFAC PapersOnLine 51-18 (2018) 578–583 DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL DESIGN AND CONTROL OF POLY(OXYMETHYLENE) DIMETHYL AND METHANOL IN AQUEOUS SOLUTIONS ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE AND METHANOL IN AQUEOUS SOLUTIONS ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE ETHERS PRODUCTION PROCESS DIRECTLY FROM FORMALDEHYDE AND METHANOL IN AQUEOUS SOLUTIONS AND METHANOL IN AQUEOUS SOLUTIONS Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien AND METHANOL IN AQUEOUS SOLUTIONS AND METHANOL IN AQUEOUS SOLUTIONS Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien
Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien Taiwan Department of Chemical Engineering, National Taipei 106, TAIWAN Zi-Jie Ai, Chuan-Yi Chuan-Yi Chung andUniversity, I-Lung Zi-Jie Ai, Chung and I-Lung Chien Chien Taiwan Department of Chemical Engineering, National University, Taipei 106, TAIWAN Zi-Jie Ai, Chuan-Yi Chung and I-Lung Chien Department of Chemical Engineering, National Taiwan University, Taipei 106, 106, TAIWAN TAIWAN Department of of Chemical Engineering, Engineering, National Taiwan Taiwan University, Taipei Department Department of Chemical Chemical Engineering, National National Taiwan University, University, Taipei Taipei 106, 106, TAIWAN TAIWAN
Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation Abstract: Poly(oxymethylene) dimethylsolvents ethers (OME) are widely used in the reduction of sootgas. formation in diesel engines, and also as physical for absorbing carbon dioxide from natural In this Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural gas. In this Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation work, a new production process of OME is designed and rigorously studied. In the conventional process, Abstract: Poly(oxymethylene) dimethyl ethers (OME) are widely used in the reduction of soot formation in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural In this Abstract: Poly(oxymethylene) dimethyl ethers (OME) arerigorously widely used in theIn reduction of sootgas. formation work, a new production process of OME is designed and studied. the conventional process, in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural gas. this multiple process sections areasrequired tosolvents first separately produce trioxane andInmethylal, and then toIn react in diesel engines, and also physical for absorbing carbon dioxide from natural gas. In this work, a new production process of OME is designed and rigorously studied. the conventional process, in diesel engines, and also as physical solvents for absorbing carbon dioxide from natural gas. Inreact this multiple process sections are required to first separately produce trioxane and methylal, and then to work, a new production process of OME is designed and rigorously studied. In the conventional process, of these two intermediates to produce OME. In this paper, the plant-wide process via a much simply work, a new production process of OME is designed and rigorously studied. In the conventional process, multiple process sections are required to first separately produce trioxane and methylal, and then to react work, a new production process of OME is designed and rigorously studied. In the conventional process, of these two intermediates torequired produceto OME. In methanol this paper, plant-wide process viaand a much multiple process are first separately produce trioxane and methylal, then tosimply react direct synthesis ofsections OME from formaldehyde and in the aqueous solutions is newly developed. The multiple process sections are required to first separately produce trioxane and methylal, and then react of these two intermediates produce In this paper, the plant-wide via aa much multiple process sections areto required toOME. first and separately produce trioxane andprocess methylal, and then to tosimply react direct synthesis of OME from formaldehyde methanol in aqueous solutions is newly developed. The of these two intermediates to produce OME. In this paper, the plant-wide process via much simply overall process includes an upstream section and a downstream product purification section. After the of these two intermediates to produce OME. In this paper, the plant-wide process via aa much simply direct synthesis of OME from formaldehyde and methanol in aqueous solutions is newly developed. The of these two intermediates to produce OME. In this paper, the plant-wide process via much simply overall process includes an upstream section and a downstream product purification section. After the direct synthesis synthesis of OME OMEoffrom from formaldehyde and methanol methanol in aqueous aqueous solutions is newly newly developed. developed. The steady-state simulation the process is established, the process flowsheet is optimized by minimizing direct of formaldehyde and in solutions is The overall process includes an upstream section and a downstream product purification section. After the direct synthesis of OME from formaldehyde and methanol in aqueous solutions is newly developed. The steady-state simulation of the process is established, the process flowsheet is optimized by minimizing overall process includes an upstream section and aa downstream downstream product purification section. After the total annual costincludes (TAC). The dynamics and control of the proposed OME production process is also overall process an upstream section and product purification section. After the steady-state simulation of the process is established, the process flowsheet is optimized by minimizing overall process includes upstream section and a downstream product purification section. After the total annual cost (TAC). The dynamics andstrategy control ofdeveloped the proposed OME process is also steady-state simulation ofan the process is established, established, the process flowsheet isproduction optimized by to minimizing investigated in this study. A proper control is for the optimized process reject the steady-state simulation of the process is the process flowsheet is optimized by minimizing total annual cost (TAC). dynamics and control of the proposed OME process is also steady-state simulation of The theproper process is established, the process flowsheet isproduction optimized by to minimizing investigated in this study. A control strategy is developed for the optimized process reject the total annual cost (TAC). The dynamics and control of the proposed OME production process is also process disturbances. Copyright © 2018 IFAC total annual cost (TAC). The dynamics and control of the proposed OME production process is also investigated in this study. A proper control for the optimized process to reject the total annual cost (TAC). The dynamics andstrategy controlis ofdeveloped the proposed OME production process is also process disturbances. Copyright © 2018 IFAC investigated in this study. A proper control strategy is developed for the optimized process to reject the investigated in this study. A proper control strategy is developed for the optimized process to reject the process disturbances. Copyright © 2018 IFAC investigated in this study. A proper control strategy is developed for the optimized process to reject the © 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. Keywords: Computer-aided design; Process models; Distillation columns; Optimization; Computer process disturbances. Copyright © 2018 IFAC process disturbances. Copyright © 2018 IFAC Keywords: Computer-aided design; Process models; Distillation columns; Optimization; Computer process disturbances. Copyright © 2018 IFAC simulation; Process control design; Process models; Distillation columns; Optimization; Computer Keywords: Computer-aided simulation; Process control design; Keywords: Computer-aided Process models; Distillation columns; Optimization; Computer Keywords: Computer-aided simulation; Process control design; Keywords: Computer-aided design; Process Process models; models; Distillation Distillation columns; columns; Optimization; Optimization; Computer Computer simulation; Process control simulation; Process control simulation; Process control
of the overall process with upstream and downstream 1. INTRODUCTION of the overall processdesign with flowsheet upstream isanddetermined downstream sections. The optimal by 1. INTRODUCTION of the overall process with upstream and downstream sections. The optimal design flowsheet is determined by of the overall overall process with system. upstreamSection and 3downstream downstream 1. INTRODUCTION minimizing TAC of the overall shows the of the process with upstream and 1. INTRODUCTION sections. The optimal design flowsheet is determined by of the overall process with upstream and downstream 1. INTRODUCTION Diesel fuel is highly minimizing TAC of the overall system. Section 3 shows the sections. The optimal optimal design flowsheet is determined by demanded in recent years. However, sections. overall control strategy development of thisdetermined system. The 1. INTRODUCTION The design flowsheet is by Diesel fuel is highly minimizing TAC of overall system. Section 33 shows the sections. The optimal design flowsheet is determined by demanded in recent years. However, overall control strategy development of this system. The there are many drawbacks minimizing TAC of the thesimulations overall system. Section shows the for conventional diesel fuel, such closed-loop dynamic of the proposed control minimizing TAC of the overall system. Section 3 shows the fuel is highly demanded in recent years. However, overall control strategy development of this system. The Diesel minimizing TAC of the overall system. Section 3 shows the there are many drawbacks for conventional diesel fuel, such closed-loop dynamic simulations of the proposed control Diesel fuel is highly demanded in recent years. However, overall control strategy development of this system. The as smog production and low combustion efficiency. In order Diesel fuel is demanded in years. However, strategy in the face of simulations various disturbances are system. illustrated in overall control strategy development of The there are many drawbacks for conventional diesel fuel, such closed-loop dynamic of the proposed control Diesel fuel is highly highly demanded in recent recent years. However, overall control strategy development of this this system. The as smog production and low combustion efficiency. In order strategy in the face ofsome various disturbances are are illustrated in there are many drawbacks for conventional diesel fuel, such closed-loop dynamic simulations of the proposed control to solve these problems, blending an additive with there are many drawbacks for conventional diesel fuel, such this section. Finally, concluding remarks drawn in closed-loop dynamic simulations of the proposed control as smog and low efficiency. In order strategy in the face of various disturbances are illustrated in there are production many drawbacks forcombustion conventional diesel fuel, such closed-loop dynamic simulations of the proposed control to solve these problems, blending an additive with this section. Finally, some concluding remarks are drawn as smog production and low combustion efficiency. In order strategy in the face of various disturbances are illustrated in conventional diesel fuels may be a feasible way. Among as smog production and low combustion efficiency. In order Section 4. strategy in the face of various disturbances are illustrated in to solve these problems, blending an additive with this section. Finally, concluding remarks drawn as smog and combustion efficiency. In order face ofsome various disturbances are are illustrated in conventional diesel fuelslow may be a feasible way.dimethyl Among Section 4.in the to solveproduction these problems, blending an additive additive with strategy this section. Finally, some concluding remarks are drawn in different diesel additives, poly(oxymethylene) to solve these problems, blending an with this section. Finally, some concluding remarks are drawn in conventional diesel fuels may be a feasible way. Among Section 4. to solve these problems, blending an additive with this section. Finally, some concluding remarks are drawn in different diesel additives, poly(oxymethylene) dimethyl conventional diesel fuels may be beones a feasible feasible way. Among Section 4. ethers (OME) are the promising that can enhance the conventional diesel fuels may a way. Among 2. PROPOSED PROCESS DESIGN Section 4. different diesel additives, poly(oxymethylene) dimethyl conventional diesel fuels maypoly(oxymethylene) beones a feasible way. Among Section 4. ethers (OME) are the promising that can enhance the 2. PROPOSED PROCESS DESIGN different diesel additives, dimethyl oxygenation ofarediesel fuel (Pellegrini, et al., enhance 2012). The different diesel additives, poly(oxymethylene) dimethyl ethers (OME) the promising ones that the 2. PROCESS DESIGN different diesel additives, poly(oxymethylene) dimethyl oxygenation of diesel fuel (Pellegrini, et can al., 2012). The ethers (OME) are the promising ones that can enhance the 2. PROPOSED PROPOSED PROCESS DESIGN properties of OME are close to conventional diesel fuel, so The thermodynamic ethers (OME) are the promising ones that can enhance the behaviour of this system involving 2. PROPOSED PROCESS DESIGN oxygenation of diesel fuel (Pellegrini, et al., 2012). The ethers (OME) are the promising ones that can enhance the 2. PROPOSED PROCESS DESIGN properties of OME are fuel close (Pellegrini, to conventional diesel fuel,fuel so formaldehyde, The thermodynamic behaviour ofis complicated this system because involving oxygenation of diesel et al., 2012). The there is no need of modifications for those original diesel oxygenation of diesel fuel (Pellegrini, et al., 2012). The methanol and water of properties of OME are close to conventional diesel fuel, so The thermodynamic behaviour of this system involving oxygenation of diesel fuel (Pellegrini, etoriginal al., 2012). The there is no need of modifications for those diesel fuel formaldehyde, methanol and water is complicated because of properties of OME are close to conventional diesel fuel, so The thermodynamic behaviour of this system involving engines (Burger, et al., 2010). For OMEs chain outside the properties of OME are close to conventional diesel fuel, so oligomer reactions among them. Formaldehyde will The thermodynamic behaviour of this system involving there is no need of modifications for those original diesel formaldehyde, methanol and water complicated of properties of OME are close toFor conventional diesel fuel,fuel so The thermodynamic behaviour ofis thisFormaldehyde system because involving OMEs chain outside the engines (Burger, et al., 2010). oligomer reactions among them. will there is no need of modifications for those original diesel fuel formaldehyde, methanol and water is complicated because of range of 3 to 5, there may be some safety concerns, because there is need ofetmodifications for those diesel fuel oligomerize with methanol and water, respectively. The methanol and is complicated because of engines (Burger, al., 2010). For chain outside the formaldehyde, oligomer among them. Formaldehyde there isofno no need modifications for OMEs those original original diesel fuel methanol and water water is complicated becausewill of range 3points to 5, of there may be some safety concerns, because oligomerizereactions with methanol and water, respectively. The engines (Burger, et al., 2010). For OMEs chain outside oligomer reactions among them. Formaldehyde will the flash of shorter OME areconcerns, too low, and the formaldehyde, n (n<3) engines (Burger, et al., 2010). For OMEs chain outside oligomers of formaldehyde and water are called oligomer reactions among them. Formaldehyde will range of 3 to 5, there may be some safety because oligomerize with methanol and water, respectively. The engines (Burger, et al., 2010). For OMEs chain outside oligomer reactions among them. Formaldehyde will (n<3) are too low, and the the flash points of shorter OME n oligomers of formaldehyde and water are called range of 3 3ofto tolonger 5, there there may be some some safety concerns, because because oligomerize with methanol and respectively. viscosity OME are(n<3) too high. n (n>5) range of 5, may be safety concerns, poly(oxymethylene) glycols (MG HO-(CH the n, water, 2O)n-H) and The oligomerize with methanol and water, respectively. The the points shorter OME are too low, and the n too oligomers of water are called range of 3of tolonger 5, of there may be some safety concerns, because oligomerize withformaldehyde methanol and nand water, respectively. The are high. viscosity OME n (n>5) poly(oxymethylene) glycols (MG , HO-(CH -H) and the the flash flash points of shorter OME (n<3) are too low, and the 2O)nare n oligomers of formaldehyde and water called the flash points of shorter OME (n<3) are too low, and the of formaldehyde and methanol are n too high. oligomers of formaldehyde and water are and called viscosity longer n (n>5) poly(oxymethylene) glycols (MG ,, HO-(CH the the flash of points of OME shorter OMEare are too low, and the oligomers nand 2O)nare n (n<3) of formaldehyde formaldehyde water called of and methanol are and viscosity longer OME (n>5) are too high. poly(oxymethylene) glycols (MG HO-(CH O)n-H) -H) the There areof two routes fornn the manufacture of OMEs (Burger, poly(oxymethylene) n 2 viscosity of longer OME (n>5) are too high. hemiformals n, HO-(CH 2 O)and n-CHthe 3). glycols (MG , HO-(CH -H) n(HF 2O)n oligomers of formaldehyde and methanol are called viscosity of longer OME (n>5) are too high. n the manufacture of OMEs (Burger, poly(oxymethylene) glycols (MG , HO-(CH the There are two routes for n(HF 2O)n-H) hemiformals O)and n, HO-(CH 2are n-CH 3). oligomers of reactions formaldehyde and methanol called et al., 2013; Schmitz, et al., 2016a). The most common route These oilgomer occur in the system everywhere and oligomers of formaldehyde and methanol are called There are two routes for the manufacture of OMEs (Burger, poly(oxymethylene) hemiformals (HF O) n, HO-(CH 2are n-CH 3). oligomers of formaldehyde and methanol called et al., 2013; Schmitz, et al., 2016a). The most common route These oilgomer reactions occur in the system everywhere and There are two routes for the manufacture of OMEs (Burger, poly(oxymethylene) hemiformals (HF , HO-(CH O) -CH is through methanol and formaldehyde forming of route two poly(oxymethylene) n, HO-(CH2 2supplies n-CH3 3). There are two routes for the manufacture of OMEs (Burger, the oligomers are hemiformals notoccur stable. AspenPlus (HF O) ).a n n et al., 2013; Schmitz, et al., 2016a). The most common These oilgomer reactions in the system everywhere and There are two routes for the manufacture of OMEs (Burger, poly(oxymethylene) hemiformals (HF , HO-(CH O) -CH is through methanol and formaldehyde forming of two n 2supplies n 3). the oligomers are not stable. AspenPlus a et al., 2013; Schmitz, et al., 2016a). The most common route These oilgomer reactions occur in the system everywhere and intermediates, trioxane and methylal, then reacting to the et al., 2013; Schmitz, et al., 2016a). The most common route Formaldehyde package which describes the ternary mixture These oilgomer reactions occur in the system everywhere and is through formaldehyde forming of two the oligomers are not stable. supplies et al., 2013; methanol Schmitz, etand al., 2016a). Thethen most common route These oilgomerpackage reactions occur in theAspenPlus system everywhere andaa intermediates, trioxane and methylal, reacting to the Formaldehyde which describes the ternary mixture is through methanol and formaldehyde forming of two the oligomers are not stable. AspenPlus supplies target OME product (named as route 1). The alternative route is through methanol and formaldehyde forming of two with UNIFAC package model to describe vapor-liquid and vapor-a the oligomers are stable. AspenPlus supplies intermediates, trioxane and methylal, then reacting the Formaldehyde which describes the ternary is through methanol and formaldehyde ofto two the oligomers are not not stable. AspenPlus supplies a target OME from product (named as route Theforming alternative route with UNIFACequilibrium model to describe vapor-liquid andmixture vaporintermediates, trioxane and methylal, then reacting to the Formaldehyde package which describes the ternary mixture is directly formaldehyde and1). methanol in aqueous intermediates, trioxane and methylal, then reacting to the liquid-liquid and treat oligomers as Formaldehyde package which describes the ternary mixture target OME product (named as route 1). The alternative route with UNIFAC model to describe vapor-liquid and vaporintermediates, trioxane and methylal, then reacting to the Formaldehyde package which describes ternary mixture is directly from formaldehyde and methanol in aqueous liquid-liquid equilibrium and treat the oligomers as target OME productas(named (named as 2). routeHowever, 1). The The alternative alternative route with UNIFAC model to describe vapor-liquid and vaporsolution (named routeas the complex target OME product route 1). route electrolytes in equilibrium the feature of Chemistry. In thisoligomers research, the with UNIFAC model to describe vapor-liquid and is directly from formaldehyde and aqueous liquid-liquid and treat the as target OME product (named as 2). route 1).methanol The alternative route electrolytes with UNIFAC model to of describe vapor-liquid and vaporvaporsolution (named as route However, thein in the feature Chemistry. In this research, the is directly from formaldehyde and methanol in complex aqueous liquid-liquid equilibrium and treat the oligomers as azeotropes among the reactants and the products make it is directly from formaldehyde and methanol in aqueous simulations follow the idea of Formaldehyde package and liquid-liquid equilibrium and treat the oligomers as solution as route 2). However, the electrolytes in the feature of Chemistry. In this research, the is directly(named from formaldehyde and methanol in complex aqueous liquid-liquid equilibrium and treat the oligomers as azeotropes among the reactants and the products make it simulations follow the idea of Formaldehyde package and solution (named as route 2). However, the complex electrolytes in the the feature of Chemistry. Chemistry. Inrelated this research, research, the difficult to get target OME 3-5 product. solution (named as route 2). However, the complex further modify the parameters with the literatures electrolytes in feature of In this the azeotropes among the reactants and the products make it simulations follow the idea of Formaldehyde package and solution (named as route 2). However, the complex electrolytes in the feature of Chemistry. In this research, the difficult to get target OME 3-5 product. further modify the parameters with the related literatures azeotropes among the reactants and the products make it simulations follow the Kuhnert, idea of of Formaldehyde package and and azeotropes among the reactants and (Schmitz, et al., 2016b; et al., 2006). follow the idea package difficult to get target OME 3-5 product. further modify the parameters with related literatures azeotropes among the reactants and the the products products make make it it simulations simulations follow the idea of Formaldehyde Formaldehyde package and Kuhnert, et al., the 2006). et al., 2016b; difficult get target product. 3-5 further modify the parameters with the related literatures In this to paper, the OME process design and control for the (Schmitz, difficult to get target OME 3-5 product. further modify the parameters with the related literatures (Schmitz, et al., 2016b; Kuhnert, et al., 2006). difficult get target 3-5 product. further modify the parameters with the related literatures In this to paper, the OME process design and control for the (Schmitz, et al., 2016b; Kuhnert, et al., 2006). production of OME via route 2 will be investigated. 3-5 In order et to al., realize theKuhnert, simulations result easily, overall (Schmitz, 2016b; et In this paper, the process design and control for the (Schmitz, 2016b; et al., al., 2006). 2006). via route 2 will be investigated. production of shows OME 3-5 In order et to al., realize theKuhnert, simulations result easily, overall In this paper, the process design and control for the Section 2 the proposed design flowsheet In this paper, the process design and control for the composition is used instead true composition. True production of OME via route be investigated. order to realize the simulations result easily, overall In this paper, the 3-5 process design22 will anddesign control for the In Section 2 shows the proposed flowsheet True composition is used instead true composition. production of OME via route will be investigated. 3-5 In order to realize the simulations result easily, overall production of OME via route 2 will be investigated. 3-5 the In order to realize the simulations result easily, overall Section 2 shows proposed design flowsheet composition is used true composition. True production of shows OME3-5 the via route 2 willdesign be investigated. In order to realize the instead simulations result easily, overall Section 2 proposed flowsheet composition is used instead true composition. Section 2 shows the proposed design flowsheet True Copyright ©22018shows IFAC 572 composition is used instead true composition. True Section the proposed design flowsheet 2405-8963 © 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved. composition is used instead true composition. True Copyright © 2018 IFAC 572
Peer review©under of International Federation of Automatic Copyright 2018 responsibility IFAC 572Control. Copyright © 572 10.1016/j.ifacol.2018.09.362 Copyright © 2018 2018 IFAC IFAC 572 Copyright © 2018 IFAC 572
2018 IFAC ADCHEM Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
composition presents all the components including the unstable oligomers. Overall composition means the unstable HFs and MGs decomposed into formaldehyde, methanol and water, and there are mathematical relationships between true and overall composition: n~FA nFA i nMGi i nHFi
(1)
n~MeOH nMeOH nHFi
(2)
n~H 2 O nH 2 O nMG i
(3)
adopted from Zhang, et al., (2016) and it is also the stoichiometry ratio to form OME3. The CSTR is packed with catalyst (amberlyst 46) and operated at constant temperature, 363.15 K, and constant pressure of 3 atm, to reach higher conversion and prevents vaporization. The packing voidage is assumed at 50%. The reactions are highly exothermic, so an internal heat exchanger is required to supply enough heat transfer area (Luyben, 2016). Luyben (2016) proposed an equation to estimate the volume of internal heat exchanger and a practical limit that the volume fraction of internal heat exchanger for the whole reactor cannot be higher than 50%. From Fig. 1 and Fig. 2 the residence time at 1.5 min gives 0.99 equilibrium conversion condition, but the internal heat exchanger occupies over 50% volume of the reactor. Conservatively, 15 min residence time is determined and it can reach 0.999 equilibrium conversion condition.
The overall amounts of formaldehyde equals to the true amount of formaldehyde plus the summation of formaldehyde in MGs and HFs. Similarly, we can calculate the overall amounts of methanol and water. The overall amounts of OMEs equal to their true amount and then the overall amounts of total components are shown below: n~total n~FA n~MeOH n~H 2O n~OMEi
The specifications of the two following columns are consequently designed. The distillate (D1R) of the first column (C1) contains the light mixture including MeOH, MAL, OME2, and parts of water, so the top specification of impurity is 100 ppm OME3 and bottom specification is 100 ppm MeOH. The bottom flow (B1) of C1 is sent to the second column (C2) to separate and recycle the heavy mixture including longer OMEs and oligomers of FA and water from bottom (B2R) of C2. To maintain FA at bottom, a certain amount of water is necessary to form the heavy oligomers (MGs), and highly concentrated formaldehyde may cause some transportation problem. The bottom specification of C2 is 10 wt% water and the top is 1000 ppm FA. Under these specifications, the distillate (D2) of C2 is close to the azeotropic composition of water and OME3 and Fig. 3 shows the binary Txy behaviour.
(4)
The reactions for poly(oxymethylene) dimethyl ethers (OMEn, H3CO-(CH2O)n-CH3) synthesis taking place in the reactor can be seen as below: FA + MeOH HF1
(5)
MeOH + HF1 MAL(OME1) + H2O
(6)
MeOH + HFn OMEn + H2O ; 8>n>1
(7)
FA + OME1 OME2 + H2O
(8)
FA + OMEn-1 OMEn + H2O ; 8>n>1
(9)
579
The kinetics of all reactions were assumed to behave according to power law. The kinetics parameters from Schmitz, et al. (2015) are used in this study. The presence of water makes the yield of OME decrease and form complex azeotropes with OMEs. The proposed design of the OME production process is to produce high purity OME3-5 product (99.95 wt%) to match the diesel regulation. The overall process is divided into upstream section and downstream section. The upstream section includes a CSTR and two distillation columns, and the downstream section is a one-decanter-two-stripper system with a purge stream. Besides producing with the CSTR, recycling of the unreacted reactants (FA and MeOH), and shorter and longer OMEs is the main purpose of the two distillation columns in the upstream section. In downstream section, the OME product and high purity water are withdrawn from the bottom of two strippers.
Fig. 1. Relationship between equilibrium conversion condition and residence time.
The fresh feed is formalin and pure methanol. The formalin feed consists of 37 wt% formaldehyde and 63 wt% water and the flow rate is fixed at 4000 kg/hr. Adjusting fresh methanol flow rate makes the mass ratio of formaldehyde and methanol equal to 90:64 feeding to the CSTR. This feed ratio is
Fig. 2. Relationship between volume fraction of internal heat exchanger and residence time. 573
2018 IFAC ADCHEM 580 Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
440
respectively. The optimized total stages of WSTRP is 8 and total stages of OSTRP is 6, and the purge fraction is 10%. Comparing the effect of purge fraction on TAC with and without product loss, the result is shown in Fig. 5 which justifies the selection of purge fraction at 10%.
T-xy diagram for H2O/OME3
Temperature, K
x 1. 0 atm y 1. 0 atm 420
400
The product loss means the production difference between the simulating point with the lowest TAC and others simulating point. A larger purge fraction increases the product loss and a smaller fraction gives higher recycle flows which increase the energy consumption. The production difference needs to be concerned and the unit product cost is assumed to be 0.6148 $/kg (Schmitz, et al., 2016a).
380
360 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Liqu id/vapo r mass f raction, H2O
Fig. 3. Binary Txy behaviour of OME3 and water. The minimum azeotrope of OME3 and water locates on the liquid-liquid splitting region, so the one-decanter-twostripper system can be applied in downstream section. A small amount of lighter component (MeOH) in D2 may be accumulated in the system, so that a MeOH purge stream is necessary. The objective function of optimization is to minimize total annual cost (TAC) and iterative optimization procedure is followed to find the best set of design variables. The overall optimization is carried out sequentially through upstream to downstream section. The design variables in upstream section need to be optimized are: total stages of C1 and C2, feed stage of C1 and C2. The design variables in downstream section need to be optimized are: total stages of the two strippers, and the purge fraction. The total stages of C1 is fixed at 50, since the original optimization displays the total stages of C1 becomes extremely high and not practical at 95. Fixing NC1 at 50 only increases the TAC by 4.3% as compared with the impractical optimized TAC with extremely high C1 column.
Fig. 5. Minimized TAC at each fixed purge fraction. 3. CONTROL STRATEGY DEVELOPMENT After the optimized design has been developed, an investigation is made to develop to the proper overall control strategy of the system. The following design of control strategy is simulated by using Aspen Plus Dynamics. In this section the overall process is still simulated separately with upstream and downstream sections. The control strategy needs to deal with various disturbance changes with OME product remaining at high purity. The disturbances considered in this study include: throughput changes via feed
From Fig. 4 below, the final optimized number of total stages for C1 is 50 and total stages of C2 is 16. The optimized feed locations for C1 and C2 are at 6th stage and 5th stage,
Purge 320 K 43.45 kg/hr 0.509 H2O 130 ppm FA 0.020 MeOH 0.470 OME3 740 ppm OME4
320 K -0.25 MW
-20.12 MW
Methanol 363.15 K 3 atm 1045.52 kg/hr 1 MeOH
Formalin 363.15 K 3 atm 4000 kg/hr 0.63 H2O 0.37 FA
CSTR -5.74 MW VR 30.6 m3 VHX 3.49 m3 363.15 K 3 atm Reactor Feed 369.5 K 3 atm 58297.4 kg/hr 0.108 H2O 0.266 FA 0.189 MeOH 0.269 MAL 0.123 OME2 0.014 OME3 0.020 OME4 0.008 OME5 0.003 OME6 0.001 OME7 411 ppm OME8
1 atm
6
D1R 324.9 K; 1 atm 34683.8 kg/hr 0.055 H2O 0.287 MeOH 0.452 MAL 0.206 OME2 100 ppm OME3
RR 1.93 D 3.19 m
D2 371.5 K; 1 atm 5045.25 kg/hr 0.558 H2O 1000 ppm FA 183 ppm MeOH 0.434 OME3 0.007 OME4 10 ppm OME5
320 K -2.56 MW
-3.96 MW
Reactor Effluent 363.15 K; 3 atm 58297.4 kg/hr 0.113 H2O 0.240 FA 0.171 MeOH 0.269 MAL 0.123 OME2 0.052 OME3 0.021 OME4 0.008 OME5 0.003 OME6 0.001 OME7 410 ppm OME8
20.21 MW
1 atm
50 B1 403.3 K; 1.39 atm 23613.5 kg/hr 0.198 H2O 0.593 FA 100 ppm MeOH 0.127 OME3 0.051 OME4 0.020 OME5 0.007 OME6 0.003 OME7 0.001 OME8
5
C2
15
RR 1.06 D 1.39 m B2R 419.4 K 1.07 atm 18568.6 kg/hr 5.18 MW 0.1 H O 2 0.754 FA 16 77 ppm MeOH 0.044 OME3 0.063 OME4 0.025 OME5 0.009 OME6 0.004 OME7 0.001 OME8
Fig. 4. Optimized process design flowsheet. 574
OR 377 K 434.45 kg/hr 0.509 H2O 130 ppm FA 0.020 MeOH 0.470 OME3 740 ppm OME4
Duty=0 WR 376.5 K 6042.19 kg/hr 0.54 H2O 53 ppm FA 0.008 MeOH 0.451 OME3 623 ppm OME4
C1
49
320 K -0.18 MW
Decanter
1.2 atm
WSTRP
7
D 0.97 m
2.76 MW 8
OP 320 K; 1 atm 2612.84 kg/hr 0.085 H2O 65 ppm FA 0.003 MeOH 0.904 OME3 0.008 OME4 11 ppm OME5 WP 320 K; 1 atm 8866.19 kg/hr 0.683 H2O 592 ppm FA 0.005 MeOH 0.309 OME3 0.002 OME4 3 ppm OME5
1.2 atm OSTRP
5
WOUT 380.6 K; 1.28 atm 2824 kg/hr 0.988 H2O 0.002 FA 0.004 OME3 0.006 OME4 8 ppm OME5
D 0.47 m
0.32 MW 6 OOUT 437.6 K; 1.26 atm 2178.39 kg/hr 420 ppm H2O 52 ppm FA 28 ppm MeOH 0.9903 OME3 0.0092 OME4 1 ppm OME5
2018 IFAC ADCHEM Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
581
flow rate and composition changes of formaldehyde in formalin. Before developing quality control loops, the inventory control loops have to be designed first. The liquid levels of reactor, decanter, reflux drums and sumps are controlled by manipulating their outlet flow. The temperature of reactor and inlet flow of decanter is controlled by manipulating cooling duty. The top pressures of column and stripper are controlled by manipulating the condenser duties and the top vapor flow, respectively.
Fig. 9. C2_RR open-loop test.
Open-loop and closed-loop sensitivity analyses are carried out to determine the locations of temperature control for each column as quality control loops. The open-loop tests are increasing and decreasing 0.1% manipulating variables, reboiler duty (Qr) and reflux rate (RR). The closed-loop tests are introducing composition disturbance to the system with assuming of perfect control is achieved. The ratio of FA and MeOH feeding to the CSTR is fixed. Fig. 10. C2_Qr open-loop test.
Figs. 6 to 11 are the result of upstream sensitivity tests, and Fig. 12 is the proposed upstream control structure. A notch locates on the stage 5 from the closed-loop test, and for the two manipulated variable of C1 (RR and Qr) stage 5 shows the largest open-loop sensitivity. Intuitively, fixing the reflux ratio and manipulating reboiler duty to control the temperature of stage 5 are determined.
Fig. 11. C2 closed-loop test. LC1 PC1
FC1
1 atm
TCr MeOH
6 X
Formalin
C1
Fig. 6. C1_RR open-loop test.
MeOH
FC2 sp X
FA
PC2
T5 T5C
LCr 49
LC3 1 atm
50
5
T4
T4C
C2
LC2
T15
T15C
15 16
LC4
Fig. 12. The proposed control structure of upstream section. There are two locations, stage 4 and 15, with smaller temperature deviation from the closed-loop test. The stage 15 and 16 give larger open-loop sensitivity of reboiler duty, so the stage 15 is controlled. The stage 4 shows enough openloop sensitivity from reflux rate, so the reflux rate is chosen to control the temperature of stage 4.
Fig. 7. C1_Qr open-loop test.
The dead time of 1 min for each temperature measurement is assumed in the simulation study. The upstream dynamic responses of important variables are shown in Fig. 13 and Fig. 14. The most important output variable is the MeOH containing in D2. Then the new steady state of the upstream section is considered as the disturbances introduced into the downstream section.
Fig. 8. C1 closed-loop test. 575
2018 IFAC ADCHEM 582 Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
Fig. 15 is the result of downstream sensitivity tests, and Fig. 16 is the proposed downstream control structure with the purge fraction is fixed at base case value. The temperature of stage 5 of OSTRP and the temperature of stage 7 of WSTRP are controlled for smaller closed-loop temperature difference and enough open-loop sensitivity. The downstream dynamic responses of important variables are shown in Fig. 17 and Fig. 18. This control strategy is able to reach a new steady state quickly with acceptable small purity deviations. The largest OME purity deviation is only 6×10-4% with positive throughput change.
Fig. 14. The upstream dynamic responses of the feed composition disturbances.
Fig. 13. The upstream dynamic responses of the throughput changes. Fig. 15. Results of the downstream sensitivity test. FC1
320 K TC1 sp FC2
X
0.1
Decanter Duty=0
TC3 320 K
TC2 320 K
OLC
PC2
WLC
PC1 1.2 atm
1.2 atm OSTRP
WSTRP
T5 T7
WT7C
7
OT5C
5 8
6
LC1
LC2
Fig. 16. The control structure of downstream section. 576
2018 IFAC ADCHEM Shenyang, Liaoning, China, July 25-27, 2018 Zi Jie Ai et al. / IFAC PapersOnLine 51-18 (2018) 578–583
583
outlet stream of the process is the water by-product at high purity of 98.8 wt%. Another methanol purge stream is also necessary to prevent accumulation of light component in the downstream section. In the process optimization study, small loss of OME product in this purge stream is also included in the TAC calculations. Due to page limitation, a further energy-saving design of dividing-wall column to combine the two columns in the upstream section is not described in this paper The control structures of upstream and downstream sections are studied separately because there is no recycle stream from downstream to upstream. The proposed control strategy is able to hold product streams at high purity despite various feed flow rate and feed composition disturbance changes. REFERENCES
Burger, J.; Siegert, M.; Strofer, E.; Hasse, H. (2010). Poly (oxymethylene) dimethyl ethers as components of tailored diesel fuel: Properties, synthesis and purification concepts. Fuel, 89, 3315-3319. Burger, J.; Strofer, E.; Hasse, H. (2013). Production process for diesel fuel components poly(oxymethylene) dimethyl ethers from methane-based products by hierarchical optimization with varying model depth. Chemical Engineering Research and Design, 91, 2648-2662. Kuhnert, C., Albert, M., Breyer, S., Hahnenstein, I., Hasse, H., & Maurer, G. (2006). Phase equilibrium in formaldehyde containing multicomponent mixtures: experimental results for fluid phase equilibria of (formaldehyde+(water or methanol)+ methylal)) and (formaldehyde+ water+ methanol+ methylal) and comparison with predictions. Industrial & engineering chemistry research, 45(14), 5155-5164. Luyben, W. L. (2016). Economic trade-offs in acrylic acid reactor design. Computers & Chemical Engineering, 93, 118-127. Pellegrini, L.; Marchionna, M.; Patrini. R. (2012). Combustion behavior and emission performance of neat and blended polyoxymethylene dimethyl ethers in a light-duty diesel engine. SAE Technical Paper, 2012-011053. Schmitz, N., Burger, J., Ströfer, E., & Hasse, H. (2016a). From methanol to the oxygenated diesel fuel poly (oxymethylene) dimethyl ether: An assessment of the production costs. fuel, 185, 67-72. Schmitz, N., Burger, J., & Hasse, H. (2015) Reaction kinetics of the formation of poly (oxymethylene) dimethyl ethers from formaldehyde and methanol in aqueous solutions. Industrial & Engineering Chemistry Research, 54(50), 12553-12560. Schmitz, N., Friebel, A., von Harbou, E., Burger, J., & Hasse, H. (2016b) Liquid-liquid equilibrium in binary and ternary mixtures containing formaldehyde, water, methanol, methylal, and poly (oxymethylene) dimethyl ethers. Fluid Phase Equilibria, 425, 127-135. Zhang, X., Oyedun, A. O., Kumar, A., Oestreich, D., Arnold, U., & Sauer, J. (2016). An optimized process design for oxymethylene ether production from woody-biomassderived syngas. Biomass and Bioenergy, 90, 7-14.
Fig. 17. The downstream dynamic responses of the throughput changes.
Fig. 18. The downstream dynamic responses of the feed composition disturbances. 4. CONCLUSIONS In this study, the design and control of a simply OME production process via direct synthesis from formaldehyde and methanol in aqueous solutions has been investigated. To the best of our knowledge, no paper in open literature can be found to study the rigorous design and also control strategy of this process. The proposed design of the process contains upstream and downstream sections requiring only limited pieces of process equipments. High purity OME product at 99.95 wt% can be obtained from the bottom of the product stripper. The other 577