- Email: [email protected]
A hybrid heat integration scheme for bioethanol separation through pressure-swing distillation route
Separation and Purification Technology 142 (2015) 307–315
Contents lists available at ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
A hybrid heat integration scheme for bioethanol separation through pressure-swing distillation route Bandaru Kiran, Amiya K. Jana ⇑ Energy and Process Engineering Laboratory, Department of Chemical Engineering, Indian Institute of Technology—Kharagpur, 721 302, India
a r t i c l e
i n f o
Article history: Received 31 July 2014 Received in revised form 1 January 2015 Accepted 4 January 2015 Available online 8 January 2015 Keywords: Bioethanol dehydration Pressure-swing distillation Internal heat integration Vapor recompression Hybrid configuration
a b s t r a c t In this contribution, a hybrid thermal integration scheme is proposed for a pressure-swing distillation (PSD) column by combining an internally heat integrated distillation column (HIDiC) with vapor recompression column (VRC). The purpose of this article is three fold: first, it designs a HIDiC for a PSD system with a reduced number of internal heat exchangers. Second, it develops a hybrid column configuration by integrating that HIDiC, which is devised based on the thermal driving force existed between the two diabatic columns, and VRC that is devised based on the thermal driving force existed between the top and bottom of the same high pressure column, yielding an ideal HIDiC–VRC configuration. Third, it provides a comprehensive comparison between the HIDiC-alone and its hybrid HIDiC–VRC structure with reference to a conventional standalone PSD column. To evaluate the performance of all these configurations, we estimate the two performance indexes, namely energy consumption and total annual cost, for an example of a bioethanol dehydration system. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction It is estimated that global energy consumption in 2040 will be almost 30% higher than the one in 2010 [1]. Because of the exponential growth in energy consumption, production of petroleum from large oil fields has already started declining at the rate of 4–5% annually [2]. The world production of oil is expected to peak in coming years. A worrying statistic is that global production rate is approaching its maximum level and the world is now finding one new barrel of oil for every four it consumes [3]. With this status, along with searching a new sustainable environment-friendly energy vector, the improvement of energy efficiency of the existing and matured process technologies has attracted the interest of research communities. Declining petroleum reserves, increasing energy demands and environmental problems are rising the interest of finding the renewable alternative energy sources, such as bioethanol and biodiesel. Among the various renewable feedstocks, sugar cane and corn are typically used to feed the human populations. Moreover, the sugar cane plantation requires a huge land, which in turn leads to deforestation that has again a negative impact on the environment. Therefore, the transformation of biomass feedstocks into ethanol has gained the research attention in recent years [4,5]. It ⇑ Corresponding author. Tel.: +91 3222 283918; fax: +91 3222 282250. E-mail address: [email protected] (A.K. Jana). http://dx.doi.org/10.1016/j.seppur.2015.01.003 1383-5866/Ó 2015 Elsevier B.V. All rights reserved.
is interesting to note that the last step of this biological ethanol production route involves the ethanol dehydration that is not an easy separation technique due to the existence of a homogeneous minimum-boiling azeotrope in a width range of pressures. According to the recent regulations, the ethanol used in the transportation sector should feature high purity (P99.5 mol%) [6]. As indicated before, the ethanol/water system forms an azeotrope having an ethanol composition of 87.2 mol% at 1 atm. For separating this homogeneous minimum-boiling azeotropic mixture, there are several options open, including azeotropic distillation, membrane separation and pressure-swing distillation (PSD). Presenting a systematic comparison between all these separation techniques, Mulia-Soto and Flores-Tlacuahuac [6] have proposed the PSD as an attractive proposition for bioethanol dehydration. Like the azeotropic distillation, the PSD system includes two separation units, featuring it as an energy demanding process. Aiming to improve the energy efficiency performance, the application of an internally heat integrated distillation column (HIDiC) approach [7–9] in PSD system is proposed in open literature. Huang et al. [10] have designed the HIDiC column by thermally coupling the bottom of low pressure (LP) column with the top of high pressure (HP) column for dealing with the separation of a minimum-boiling azeotrope. As far as the fractionation of maximum-boiling azeotrope is concerned, the reverse combination of thermal pairing is suggested by them. Through the separation of a binary azeotropic mixture (acetonitrile/water), it has been
308
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
demonstrated [10] that the general HIDiC (or HIDiC-alone) column helps to improve the process design in not only thermodynamic efficiency but also capital investment. Recently, Mulia-Soto and Flores-Tlacuahuac [6] have extended the HIDiC technology to a PSD system by coupling the whole portion of both columns. The authors have used the Aspen flowsheet simulator for process dynamics and control studies. The major complexity involved in operation and design of the HIDiC configuration is associated with the internal heat integration arranged between the whole/major part of the rectifier and that of the stripper. Therefore, in spite of the encouraging outcomes from the standpoint of both capital investment and energy consumption [11], it has not yet found wide applications in the process industries. Keeping this practical concern in mind, Chen et al. [12] have explored the possibility of using a reduced number of vertically arranged internal heat exchangers (internal HE). Subsequently, Harwardt and Marquardt [13] have formally used the optimization technique for HIDiC to establish a systematic design methodology driven by economics to find the optimal number of internal HE. In this article, we have attempted to reduce the number of internal HE in HIDiC without compromising much of its overall performance in the aspect of both energy consumption and cost. In this contribution, we propose a hybrid heat integration scheme for a PSD column by combining a HIDiC and vapor
recompression column (VRC). There are typically two possible approaches to hybridize the general HIDiC column with the VRC scheme. These approaches are based on the way the thermal coupling is made between: (i) the overhead vapor with the bottom liquid of the same HP column, and (ii) the overhead vapor of HP column with the bottom liquid of LP column. When the HIDiC is in ideal form (i.e., no entropy generation) or it does not require any external heat for the LP column reboiler, the first option is applicable for further reduction of utility consumption in the HP column of PSD configuration. Similarly, for the nonideal HIDiC case (i.e., irreversible process), the second option makes sense in hybridizing the HIDiC with VRC. This work investigates the techno-economic feasibility of a hybrid heat integrated structure for a bioethanol PSD column. For the representative system, the HIDiC is proposed to hybridize by thermally coupling the overhead vapor and bottom liquid of the same HP column under the framework of VRC system, yielding an ideal HIDiC–VRC configuration. Furthermore, aiming to make this heat integration technology one step ahead toward its implementation in industrial scale, attempts are made to design the hybrid configuration, in which, the HIDiC column includes a small number of internal HE. The performance of all these options is quantified and analyzed in terms of both the energy savings and total annual cost.
Fig. 1. Schematic representation of a conventional PSD column.
309
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
2. Conventional pressure-swing distillation: Bioethanol dehydration
Table 1 System specifications and steady state conditions. Items
2.1. Working principle and configuration
2.2. Modeling, computer simulation and verification The mathematical model for both the conventional and HIDiC of the bioethanol PSD column is derived based on the following assumptions: Perfect mixing and equilibrium on all trays. Negligible tray vapor holdups. Variable liquid holdup in each tray. NRTL model to predict vapor–liquid equilibrium (VLE). Constant stage pressure drop (0.3 kPa) and efficiency (70%) for all trays. Nonlinear Francis-weir formula for liquid hydraulics.
The modeling equations for a typical separation stage (Fig. 2) [15] are shown in the Appendix A. Obviously, the model consists of ordinary differential equations and algebraic equations that correlate the physical properties, tray hydraulics, phase equilibrium, etc. For simulating the PSD model, the optimal column setup is adopted from the literature [6]. Here, computer programs are developed in MATLAB environment. Table 1 reports the process specifications and shows the verification of our developed simulator. Now, the validation of VLE model is shown in Fig. 3. Here, the NRTL model is simulated to predict the VLE of ethanol/water system. The binary interaction parameter values are adopted from Reid et al. [16]. Experimental data available in literature [6] are used to validate the activity-based VLE model. In Fig. 3, one can readily see that the NRTL equations used in the PSD model provide a reasonably good prediction with a root-mean square (RMS) error of 3.1 102.
Vn yn,i
Ln+1 xn+1, i
V Sn , yn,i
Tray
n
(holdup =
F n , zn,i
mn )
L
Sn , xn , i
Low pressure (LP) column Feed flow rate (kmol/h) Feed composition (ethanol/water) Recycle feed rate (kmol/h) Top composition (ethanol/water) Bottom composition Bottoms rate (kmol/h) Distillate rate (kmol/h) Total number of traysa (excluding condenser and reboiler) Fresh feed stage Recycle feed stage Stage pressure drop (kPa) Bottom stage pressure (kPa) Reboiler duty (kW) High pressure (HP) column Feed flow rate (kmol/h) Feed composition (ethanol/water) Top composition Bottom composition Bottoms rate (kmol/h) Distillate rate (kmol/h) Total number of traysa (excluding condenser and reboiler) Feed stage Stage pressure drop (kPa) Bottom stage pressure (kPa) Reboiler duty (kW)
Present work
100 0.2/0.8 79.57 0.861/0.139 0.005/0.995 80.35 99.21 28
100 0.2/0.8 39.848 0.8565/0.1435 0.001054/0.998946 80.114 59.762 28
7 17 0.0 110.325 3320
7 17 0.3 119.025 1050.37
98.21 0.861/0.139 0.827/0.173 0.997/0.003 19.64 79.57 28
59.762 0.8565/0.1435 0.8272/0.1728 0.99623/0.00377 19.976 39.848 28
16 0.0 1012.78 3010
16 0.3 1012.78 850.57
a Tray numbering has started from bottom up (bottommost tray is 1 and topmost tray is 28).
1.0
Experimental VLE Simulated VLE
0.9 0.8
Yethanol (mol fract)
Fig. 1 demonstrates a typical pressure-swing distillation (PSD) column. It is a dual-column process, in which, one column operates at a relatively high pressure than the other one. The purpose of this pressure elevation is to circumvent the azeotropic point. For minimum-boiling systems, high purity product streams are taken out from the bottom of the two distillation columns and the distillate streams are recycled [14]. For the representative system, the ethanol/water azeotrope is not so sensitive to pressure changes. For instance, the azeotropic composition changes from 87.2 mol% ethanol at 1 atm to 79.4 mol% at 10 atm. However, it is observed [6] that this low sensitivity is sufficient enough to achieve a high degree of separation.
Reported Aspen work [6]
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Xethanol (mol fract) Fig. 3. X–Y diagram comparing simulated NRTL model and experimental data taken from literature [6].
It is evident in Table 1 that the PSD system secures the production of ethanol at a purity of more than 99.6 mol%, which is higher than the permissible limit (i.e., 99.5 mol%) recommended to the transportation sector. Moreover, the PSD column also produces a high-purity water (i.e., about 99.9 mol%) from the bottom of LP column, showing a successful and efficient separation of both the constituent components. 3. HIDiC-alone scheme: A conventional approach
Ln xn, i
Vn−1 yn−1, i
Fig. 2. Quantities associated with a typical nth tray.
The HIDiC configuration is built by dividing the distillation tower into two columns, one for each of its rectifying and stripping sections. The stripping column accompanies a reboiler and the rectifying column includes a condenser. A compressor is installed over
310
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
Table 2 Cost estimating formula and parameter value. 1:066 0:802 Column shell: M&S Lc ðcin þ cm cp Þ 280 101:9Dc where Dc is the column diameter (ft), Lc the column height (ft), M&S = 1625.9 [20], and the coefficients cin = 2.18, cm = 3.67, and cp = 1.2 1:55 Lc ðcs þ ct þ cm Þ Column tray: M&S 280 4:7Dc where the coefficients cs = 1, ct = 0, and cm = 1.7 M&S Heat exchanger: 280 101:3A0:65 ðcin þ cm ðct þ cp ÞÞ where the coefficients cin = 2.29, cm = 3.75, ct = 0.1, and cp = 1.35 0:82 ðcin þ ct Þ Compressor: M&S 280 517:9BHP where the coefficient cin = 2.11, and ct = 1.0
the stripper to pressurize the overhead vapor before its entry to the rectifier. Consequently, the rectifying section operates at a relatively high pressure compared to the stripping section, causing a thermal driving force between them. To accomplish the heat transfer from the HP to the LP column, a set of internal heat exchangers (HE) is employed to couple the trays between these two diabatic sections. This internal thermal integration leads to bring the liquid reflux for the rectifier and the vapor boil-up for the stripper, thereby reducing the consumption of thermal utility supplied externally to the condenser and reboiler, respectively. Note that the bottom stream of HP rectifier is further throttled in a pressure reducing valve before recycling back to the LP stripper.
In an adiabatic one-column system, called conventional distillation column (CDiC), a condenser and a reboiler are used to generate external reflux and boil-up, respectively. When the internal heat exchangers in HIDiC are able to alter these external flows, the HIDiC scheme gets its ideal form. Actually, the ideal HIDiC corresponds to a complete reversible operation, in that no entropy is generated and it demarcates the theoretical limit of performance. Obviously, this ideal scheme does not require any reboiler as well as condenser. The model of HIDiC column additionally includes a couple of equations as shown in the Appendix A. These additional modeling equations are basically used to respectively compute the internal liquid and vapor flows through the HP rectifier and the LP stripper, and the heat transfer between these two integrated columns through the internal HE. As explained above, compared to an adiabatic column, its HIDiC counterpart is capable of reducing the consumption of both reboiler duty of LP column and the condenser heat load of HP column. However, the heat integration effectively involves an additional compressor, along with a set of internal HE. Obviously, both of these equipments require an additional capital investment. Moreover, although the heat exchanger set does not require any external medium, but the compressor is driven by externally supplied electricity, which is much more expensive than the thermal utility.
Fig. 4. Schematic representation of the HIDiC-alone of bioethanol PSD column.
311
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
It clearly indicates the necessity of a thorough quantitative analysis to judge the energetic and economic potential of the HIDiC column with reference to a conventional standalone column (i.e., CDiC). In this regard, the estimation of two performance indexes is briefly presented below. 3.1. Performance indexes 3.1.1. Energy consumption The total energy consumed (Q Cons ) by a thermally integrated scheme is determined by adding the reboiler heat load (Q R ) plus three times the compressor duty (Q Comp ). It implies:
Q Cons ¼ Q R þ 3Q Comp
ð1Þ
The factor of three for the compression duty is assumed to convert the compression work into the thermal energy required to produce an equivalent amount of electrical power [17]. It should be highlighted that for the CDiC column, Q Comp ¼ 0. As far as HIDiC is concerned, the following form of correlation is used to estimate the compressor duty [18]:
l l1
h l1 i V nT Pi CR l 1
ð2Þ
In the above equation, the pressure (inlet pressure, P i and outlet pressure, Po ) is in lbf/ft2, and the vapor inflow rate to the compressor, V nT is in ft3/min. Here, the compression ratio (CR) is expressed as:
CR ¼
l=ðl1Þ Po To ¼ Pi Ti
ð3Þ
where T i and T o are the inlet and outlet temperature (K), respectively, with respect to the compressor. We compute the CR from Eq. (3) assuming the requirement of a minimum temperature difference (DT min ) of 20 K for complete condensation of compressed vapor in thermal coupling [18]. The polytropic coefficient, lð¼ C p =C v Þ can be calculated from:
1=ðl 1Þ ¼
X
yi =ðli 1Þ
3.2.1. Selection of total number of internal heat exchangers and their locations As shown in Table 1, the HP column operates at about 10 atm, while the LP column at 1 atm. It clearly indicates that the trays in HP section are hotter than those in LP section. In this regard, Fig. 5a displays a comparative temperature profile before implementing the thermal coupling. Since, there exists a reasonably large thermal driving force along the length of both diabatic sections, one can think of coupling all trays between them. However, as stated previously, the participation of all trays (or most of them) in thermal pairing makes the design and operation of HIDiC a complex one. In fact, this complexity is the main barrier in implementing the HIDiC in industrial practice. Prior to configuring the hybrid HIDiC–VRC scheme, therefore, in this section, we would like to find a small number of internal heat exchangers (HE) and their suitable locations for the HIDiC of PSD column. In order to select the number of heat exchangers, we fix the two criteria: (i) same ethanol productivity and purity (with that of conventional PSD), and (ii) either complete heat reversibility (i.e., no entropy generation) or zero reboiler duty in the LP column. For a
ð4Þ
Here, C p and C v are the heat capacities, and yi the mole fraction of any vapor component i. Knowing the Q Cons for CDiC and its heat integrated counterpart, one can estimate the percent energy savings.
430 420
Tray temperature (K)
Q Comp ¼ 3:03 105
internal heat exchangers. Interestingly, the bioethanol PSD column presented before inherently includes a compressor (CR = 10) and a throttling valve. So, the application of HIDiC to a PSD system (Fig. 4) additionally requires only a couple of heat exchangers to connect the trays between HP and LP columns, indicating a greater possibility of energy efficiency and economic performance improvement over the HIDiC of traditional distillation process.
HP LP
400 390 380 370 360 350 340
3.1.2. Economic evaluation The total annual cost (TAC) that is commonly used to assess the economic feasibility of a heat integrated configuration is expressed as:
2
4
6
8
10 12 14 16 18 20 22 24 26 28 30
Tray number 420
TAC ð$=yrÞ ¼ annual operating cost ðOCÞ
b
410
ð5Þ
The capital cost is computed by summing up the cost of all major equipments. The cost estimating formulas are given in Table 2. The operating cost of the compressor is estimated as suggested by Douglas [18] based on the bhp (=hp/0.9), and a motor efficiency of 0.6. For the sake of simplicity, operating costs are taken to be identical to utility costs, i.e. the number resulting from the summation of electricity (0.084$/kW h), steam (17$/t) and cooling water (0.06$/t) costs for a year having 8000 operating hours [10].
Tray temperature (K)
capital investment ðCIÞ þ payback period ðhÞ
a
410
HP28 LP28 HP15 LP15 HP8 LP8 HP4 LP4
400 390 380 370 360 350 2
4
6
8
10 12 14 16 18 20 22 24 26 28 30
Tray number 3.2. Development of HIDiC-alone scheme for bioethanol separation In comparison with an adiabatic CDiC, its HIDiC counterpart additionally includes a compressor, a throttling valve and a set of
Fig. 5. Comparative tray temperature profiles of HP and LP columns having: (a) no internal HE, and (b) 4 different sets of internal HE (i.e., 28, 15, 8 and 4) [tray numbering has started from bottom up; reboiler as 1 and condenser as 30 in both the columns].
312
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
Table 3 A comparative performance in terms of energy savings and cost.
a
Item
Conventional PSD
HIDiC of PSD
Ideal HIDiC–VRC
28 HE
15 HE
8 HE
4 HE
Total energy consumption (kW) Energy savings (%)
1919.58
869.39 54.71
869.51 54.70
870.02 54.68
870.065 54.67
328.66 82.88
Operating cost ($/yr) LP Column Condenser Duty (kW) Cost of CW
832.73 341025.88
849.05 347710.05
870.62 356547.26
891.24 364988.06
901.23 369079.24
901.23 369079.24
Reboiler (LP column) Duty (kW) Cost of steam
1050.37 258068.18
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
HP Column Condenser Duty (kW) Cost of CW
1090.12 446467.41
968.02 396432.46
962.88 394328.25
969.42 397005.66
970.62 397496.42
0.0 0.0
Trim-reboiler Duty (kW) Cost of steam
850.58 322831.19
850.58 322738.73
850.58 322505.69
850.58 322510.58
850.58 322526.18
302.98 115715.54
Comp1 Duty (kW) Cost of electricity
6.21 6955.2
6.27 7022.4
6.31 7067.2
6.48 7259.8
6.495 7274.4
6.495 7274.4
Comp2 Duty (kW) Cost of electricity Total operating cost ($/yr) Utility cost savings (%)
1375347.86
1073903.64 21.92
1080448.4 21.44
1091764.1 20.62
1096576.24 20.27
2.065 2312.8 494381.98 64.05
Operating conditions UA (kW/K stage) Ethanol purity, mol% Ethanol productivity (kmol/h)
99.623 19.976
0.55 99.628 19.983
1.043 99.627 19.974
2.3 99.627 19.96
4.135 99.626 19.94
4.135 99.626 19.94
Reflux ratio LP column HP column Stage pressure drop (kPa)
1.218 2.448 0.3
1.218 2.44 0.3
1.221 2.42 0.3
1.224 2.40 0.3
1.226 2.38 0.3
1.226 2.38 0.3
Capital cost ($) LP Column Condenser Reboiler Column Trays
260437.89 250688.90 733268.36 2081624.16
263744.21 0.0 733268.36 2081624.16
268082.53 0.0 733268.36 2081624.16
272190.84 0.0 733268.36 2081624.16
274170.13 0.0 733268.36 2081624.16
274170.13 0.0 733268.36 2081624.16
HP Column Condenser Trim-reboiler Bottom reboiler Column Trays Internal heat exchanger Comp1 Comp2 Total capital cost ($) TACa ($/yr) TAC savings (%)
155190.06 218562.62 512260.01 1235673.88 0.0 50165.81 0.0 5497871.69 3207971.76
143651.63 218562.62 512260.01 1235673.88 1458997.03 50562.92 0.0 6698344.82 3306685.25 ve
143155.55 218562.62 512260.01 1235673.88 1173198.94 50827.28 0.0 6416653.33 3219332.84 ve
143786.60 218562.62 512260.01 1235673.88 898191.32 51947.45 0.0 6147505.23 3140932.51 2.09
143902.11 218562.62 512260.01 1235673.88 738060.10 52046.04 0.0 5989567.41 3092818.71 3.59
0.0 113213.58 247008.99 512260.01 1235673.88 738060.10 52046.04 20337.95 6007663.19 2496936.38 22.16
Payback period = 3 yr.
meaningful comparison between a conventional column and its heat integrated schemes, it is reasonable to force all the schemes to meet both the above criteria. It is worth noticing that in all configurations, two internal heat exchangers (HEs) are targeted to keep fixed, one at the top and other one at the bottom. The former one is arranged between the tops of the two diabatic sections, targeting the HP column to operate in a reflux-free operation mode [19]. On the other hand, the bottom external heat exchanger is accommodated between the bottoms of those sections, enabling the LP column to run in a reboil-free operation mode. Furthermore, the intermediate heat
exchangers are employed to generate internal flows in both the columns. A systematic comparison is presented in Table 3 between various forms of HIDiC, differing with respect to the number of internal heat exchangers used and their locations. In the first case, total 28 internal heat exchangers are employed to connect all 28 trays of both the diabatic sections. Note that both the HP and LP columns have the same number of trays (i.e., 28), excluding the reboiler and condenser. For the case of 15 heat exchangers, each of these heat exchangers of the HIDiC is coupled in every alternate tray, except at the top where two heat exchangers are consecutively
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
Internal vapor flow rate (kmol/s)
0.045
28HE 15HE 8HE 4HE
0.040 0.035 0.030 0.025 0.020 0.015 0.010 0.005 0.000
2
4
6
8
10 12 14 16 18 20 22 24 26 28
Tray number Fig. 6. Internal vapor flow rate profile through the HP column.
connected with 27th as well as 28th trays. In the next, 8 heat exchangers are configured and each of them is connected in every 4th tray. The last scheme includes 4 heat exchangers and these are coupled with following trays of both the columns: 1, 10, 19 and 28. Note that except the top and bottom internal heat exchangers (HE),
313
all other HE and their locations are selected arbitrarily in the four HIDiC configurations stated above. It is evident in Fig. 5b that all four cases have very close thermal driving force between the trays of HP and LP columns. It is further interesting to observe the internal fluid flow profile in the HP or LP columns. As shown the vapor flow rate through the HP column in Fig. 6, there are several peaks depending on their number of internal HE, except a breakthrough associated with the feed tray (i.e., 16). Actually, the peaks are involved with those trays which are connected with the internal HE. It should be noted that the difference existed between the vapor flow rates in the four HIDiC configurations (Fig. 6) causes a difference in separation effect. As shown in Table 3, all four options of HIDiC having a different number of heat exchangers and their locations achieve very close ethanol productivity (19.97 kmol/h) and purity (99.6%) with that of the conventional PSD scheme, indicating the validation of first criterion. As far as second criterion is concerned, all these configurations achieve zero reboiler duty in the LP column by mainly adjusting the UA of internal heat exchanger. It should be noted that in each case, the UA is taken same for all internal heat exchangers. Since all four options meet the second criterion, we obtain a close energy savings (54.7%) for all of them but a wide variation in economic performance. In Table 3, one can readily find that the HIDiC with 4 heat exchangers shows the best performance securing an energy and TAC savings of about 54.7% and 3.59%, respectively.
Fig. 7. Schematic representation of the proposed hybrid HIDiC–VRC of PSD column.
314
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
In the subsequent section, this HIDiC column is used to propose a hybrid heat integration scheme. 4. Hybrid HIDiC–VRC scheme As shown in Fig. 1, the PSD system comprises of two columns; both of which have a reboiler and a condenser. For improving the energetic potential of the HIDiC-alone scheme, there is a scope of further intensification by introducing the VRC in the HIDiC of PSD configuration. This hybridization can be accomplished in two modes, which differ in the way the thermal coupling is made between (i) the overhead vapor (heat source) and the bottom liquid (heat sink) of the same HP column, and (ii) the overhead vapor of HP column (heat source) and the bottom liquid of LP column (heat sink). For the case of HIDiC, which theoretically achieves a condition of zero reboiler duty with respect to the LP column of PSD system, the first option should be adopted. As far as nonideal HIDiC is concerned, the VRC can be suitably hybridized with HIDiC by adopting the second option. It is worth noticing that both these hybrid structures have the potential of improving energy efficiency performance compared to the HIDiC-alone. However, for economic feasibility of both the hybrid HIDiC–VRC columns, it is recommended to perform the cost analysis. At this stage, it should be pointed out that for the hybrid configuration accompanying with the HIDiC (i.e., under first option), the additional capital investment involved for the VRC can be compensated by the thermal utility consumption saved in the HP column of PSD. On the other hand, the economic performance of the hybrid scheme under second option improves with the increase of the degree of heat irreversibility existed in the LP column of nonideal HIDiC. 4.1. Development of a hybrid HIDiC–VRC scheme for bioethanol separation This section deals with the development of a hybrid heat integrated configuration by combining the HIDiC with 4 HE and the VRC scheme. The HIDiC of the representative bioethanol PSD column does not require any external heat for the LP column reboiler, leaving no option open for further coupling between the overhead vapor of HP column and the bottom liquid of LP column (i.e., second option). In this work, therefore, an alternative thermal integration approach is explored between the overhead vapor (at 421.99 K) and bottom liquid (at 422.97 K) of the same HP column under the VRC framework (i.e., first option). As assumed the requirement of a thermal driving force of at least 20 K [18] for complete condensation of the compressed vapor, the inclusion of VRC is inevitable for the sample bioethanol dehydration system. By this way, the HIDiC scheme is hybridized with the VRC. Fig. 7 illustrates the hybrid HIDiC–VRC column. For the VRC system, the compression ratio (CR) is estimated for Comp2 as 1.23 using Eq. (3). This compressed vapor releases latent heat, amounting 547.6 kW [=(850.58–302.98) kW], in the bottom reboiler of HP column for liquid reboiling. Additionally, the HP column of this hybrid configuration does not require the top vapor condensation, thereby involving no operating and capital investment for the respective overhead condenser. It is now obvious that the combined HIDiC–VRC scheme does not involve any reboiler for the LP column and condenser for the HP column, and their respective heat duties. Therefore, the resulting structure can be referred to as the ideal HIDiC–VRC column. This hybrid HIDiC–VRC confirmation, as shown in Table 3, achieves the best performance compared to all other schemes discussed for the representative PSD column, securing a 82.88% savings in utility consumption, a 64.05% savings in utility cost and a 22.16% savings in TAC.
5. Conclusions The use of internal heat exchangers connecting all or most of the trays between two diabatic sections is remained as a major challenge in implementing the HIDiC technology. Keeping this practical concern in mind, it is attempted to design the HIDiC column with a reduced number of internal heat exchangers without compromising much of its overall performance in the aspect of both energy consumption and cost. In this contribution, we propose a hybrid heat integration scheme for a bioethanol PSD column by combining a HIDiC with reduced number of internal HE and VRC. Since the HIDiC of the sample system requires zero reboiler duty with respect to its LP column, it is not really meaningful to have further intensification between two diabatic sections. However, there exists a possibility of further reduction of utility consumption in the HP column of PSD configuration. With this objective, we introduce a thermal integration between the overhead vapor (heat source) and bottom liquid (heat sink) in the same HP column under the framework of VRC scheme. By this way, the VRC is proposed to hybridize with the HIDiC in a PSD system, eventually yielding an ideal HIDiC– VRC configuration. For the representative bioethanol dehydration system, this hybrid HIDiC–VRC configuration outperforms the HIDiC-alone and conventional PSD system, showing a promising performance from the standpoint of energy (82.88%), utility cost (64.05%) and TAC savings (22.16%). Appendix A A.1. Modeling equations of a distillation tray (CDiC and HIDiC) As depicted in Fig. 2, the nth tray of a distillation column is fed with a liquid feed mixture. Side streams are withdrawn as a liquid as well as a vapor. The modeling equations can be derived in the following forms [15]: Total mole balance
_ n ¼ Lnþ1 þ V n1 þ F n ðLn þ SLn Þ ðV n þ SVn Þ m
ðA1Þ
Component mole balance
_ n x_ n;i ¼ Lnþ1 xnþ1;i þ V n1 yn1;i þ F n zn;i ðLn þ SLn Þxn;i m ðV n þ SVn Þyn;i
ðA2Þ
Energy balance
_ n H_ Ln ¼ Lnþ1 HLnþ1 þ V n1 HVn1 þ F n HFn ðLn þ SLn ÞHLn m ðV n þ SVn ÞHVn Q n
ðA3Þ
Equilibrium
yn;i ¼ kn;i xn;i ¼ cn;i
P 0n;i xn;i Pt
ðA4Þ
Summation NC X xn;i ¼ 1
ðA5aÞ
i¼1 NC X yn;i ¼ 1
ðA5bÞ
i¼1
Here, x denotes the liquid phase composition, y the vapor phase composition, z the feed composition, L the liquid flow rate, V the vapor flow rate, k the vapor–liquid equilibrium constant, H the enthalpy, N C the total number of components, P t the total pressure, P0 the vapor pressure, Q the heat loss, F the feed flow rate, S the flow rate of side stream, and c the activity coefficient. The
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
subscript/superscript n indicates the tray index, i the component index, F the feed, L the liquid stream and V the vapor stream. In the computer simulation, algebraic form of equations [15] has been solved to compute the vapor and liquid enthalpies. Note that if there are no side draws, then we consider SL ¼ SV ¼ 0. Negligible heat loss from a stage to the surroundings (Q n ¼ 0) is also assumed. Now, it is indeed straightforward to extend this modeling approach to all other trays. A.2. Internal liquid and vapor flows (heat integrated schemes) Supposing the thermal coupling between jth stage of LP column with jth stage of HP column through an internal heat exchanger (HE), we have:
Q HE;j ¼ UAðT j;HP T j;LP Þ
ðA6Þ
where Q HE refers to the rate of heat transfer through the internal HE. Now, the internal flow rates in the LP and HP columns are given as: LP column
Q HE;j V LP j ¼ PN c ð i¼1 yi ki Þj
ðA7Þ
HP column
Q HE;j LHP ¼ PN c j ð i¼1 xi ki Þj
ðA8Þ
Here k represents the latent heat. References [1] I. López, J. Andreu, S. Ceballos, I. Martínez de Alegría, I. Kortabarria, Review of wave energy technologies and the necessary power-equipment, Renew. Sustain. Energy Rev. 27 (2013) 413–434. [2] P.S. Nigam, A. Singh, Production of liquid biofuels from renewable resources, Prog. Energy Combust. Sci. 37 (2011) 52–68.
315
[3] K. Aleklett, C.J. Campbell, The peak and decline of world oil and gas production, Miner. Energy 18 (2003) 35–42. [4] C.A. Cardona, O.J. Sánchez, Fuel ethanol production: process design trends and integration opportunities, Biores. Technol. 98 (2007) 2415–2457. [5] J. Regalbuto, An NSF perspective on next generation hydrocarbon biorefineries, Comput. Chem. Eng. 34 (2010) 1393–1396. [6] J.F. Mulia-Soto, A. Flores-Tlacuahuac, Modeling, simulation and control of an internally heat integrated pressure-swing distillation process for bioethanol separation, Comput. Chem. Eng. 35 (2011) 1532–1546. [7] X. Liu, J. Qian, Modeling, control, and optimization of ideal internal thermally coupled distillation columns, Chem. Eng. Technol. 23 (2000) 235–241. [8] Y. Zhu, X. Liu, Dynamics and control of high purity heat integrated distillation columns, Ind. Eng. Chem. Res. 44 (2005) 8806–8814. [9] B. Kiran, A.K. Jana, Introducing vapor recompression mechanism in heatintegrated distillation column: impact of internal energy driven intermediate and bottom reboiler, AIChE J. 61 (2015) 118–131. [10] K. Huang, L. Shan, Q. Zhu, J. Qian, Adding rectifying/stripping section type heat integration to a pressure-swing distillation (PSD) process, Appl. Therm. Eng. 8 (2008) 923–932. [11] K. Naito, M. Nakaiwa, K. Huang, A. Endo, K. Aso, T. Nakanishi, T. Nakamura, H. Noda, T. Takamatsu, Operation of a bench-scale ideal heat integrated distillation column (HIDiC): an experimental study, Comput. Chem. Eng. 24 (2000) 495–499. [12] H. Chen, K. Huang, S. Wang, A novel simplified configuration for an ideal heatintegrated distillation column (ideal HIDiC), Sep. Purif. Technol. 73 (2010) 230–242. [13] A. Harwardt, W. Marquardt, Heat-integrated distillation columns: vapor recompression or internal heat integration?, AIChE J 58 (2012) 3740–3750. [14] W.L. Luyben, Comparison of extractive distillation and pressure-swing distillation for acetone/chloroform separation, Comput. Chem. Eng. 50 (2013) 1–7. [15] A.K. Jana, Chemical Process Modelling and Computer Simulation, PrenticeHall, New Delhi, 2011. [16] R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, McGraw-Hill, New York, 1987. [17] K. Iwakabe, M. Nakaiwa, K. Huang, T. Nakanishi, A. Røsjorde, T. Ohmori, A. Endo, T. Yamamoto, Energy saving in multicomponent separation using an internally heat-integrated distillation column (HIDiC), Appl. Therm. Eng. 26 (2006) 1362–1368. [18] J.M. Douglas, Conceptual Design of Chemical Processes, McGraw-Hill, New York, 1988. [19] X. Zhang, K. Huang, H. Chen, S. Wang, Comparing three configurations of the externally heat-integrated double distillation columns (EHIDDiCs), Comput. Chem. Eng. 35 (2011) 2017–2033. [20] Economic Indicators, Chem Eng August 2011, pp. 56.
Contents lists available at ScienceDirect
Separation and Purification Technology journal homepage: www.elsevier.com/locate/seppur
A hybrid heat integration scheme for bioethanol separation through pressure-swing distillation route Bandaru Kiran, Amiya K. Jana ⇑ Energy and Process Engineering Laboratory, Department of Chemical Engineering, Indian Institute of Technology—Kharagpur, 721 302, India
a r t i c l e
i n f o
Article history: Received 31 July 2014 Received in revised form 1 January 2015 Accepted 4 January 2015 Available online 8 January 2015 Keywords: Bioethanol dehydration Pressure-swing distillation Internal heat integration Vapor recompression Hybrid configuration
a b s t r a c t In this contribution, a hybrid thermal integration scheme is proposed for a pressure-swing distillation (PSD) column by combining an internally heat integrated distillation column (HIDiC) with vapor recompression column (VRC). The purpose of this article is three fold: first, it designs a HIDiC for a PSD system with a reduced number of internal heat exchangers. Second, it develops a hybrid column configuration by integrating that HIDiC, which is devised based on the thermal driving force existed between the two diabatic columns, and VRC that is devised based on the thermal driving force existed between the top and bottom of the same high pressure column, yielding an ideal HIDiC–VRC configuration. Third, it provides a comprehensive comparison between the HIDiC-alone and its hybrid HIDiC–VRC structure with reference to a conventional standalone PSD column. To evaluate the performance of all these configurations, we estimate the two performance indexes, namely energy consumption and total annual cost, for an example of a bioethanol dehydration system. Ó 2015 Elsevier B.V. All rights reserved.
1. Introduction It is estimated that global energy consumption in 2040 will be almost 30% higher than the one in 2010 [1]. Because of the exponential growth in energy consumption, production of petroleum from large oil fields has already started declining at the rate of 4–5% annually [2]. The world production of oil is expected to peak in coming years. A worrying statistic is that global production rate is approaching its maximum level and the world is now finding one new barrel of oil for every four it consumes [3]. With this status, along with searching a new sustainable environment-friendly energy vector, the improvement of energy efficiency of the existing and matured process technologies has attracted the interest of research communities. Declining petroleum reserves, increasing energy demands and environmental problems are rising the interest of finding the renewable alternative energy sources, such as bioethanol and biodiesel. Among the various renewable feedstocks, sugar cane and corn are typically used to feed the human populations. Moreover, the sugar cane plantation requires a huge land, which in turn leads to deforestation that has again a negative impact on the environment. Therefore, the transformation of biomass feedstocks into ethanol has gained the research attention in recent years [4,5]. It ⇑ Corresponding author. Tel.: +91 3222 283918; fax: +91 3222 282250. E-mail address: [email protected] (A.K. Jana). http://dx.doi.org/10.1016/j.seppur.2015.01.003 1383-5866/Ó 2015 Elsevier B.V. All rights reserved.
is interesting to note that the last step of this biological ethanol production route involves the ethanol dehydration that is not an easy separation technique due to the existence of a homogeneous minimum-boiling azeotrope in a width range of pressures. According to the recent regulations, the ethanol used in the transportation sector should feature high purity (P99.5 mol%) [6]. As indicated before, the ethanol/water system forms an azeotrope having an ethanol composition of 87.2 mol% at 1 atm. For separating this homogeneous minimum-boiling azeotropic mixture, there are several options open, including azeotropic distillation, membrane separation and pressure-swing distillation (PSD). Presenting a systematic comparison between all these separation techniques, Mulia-Soto and Flores-Tlacuahuac [6] have proposed the PSD as an attractive proposition for bioethanol dehydration. Like the azeotropic distillation, the PSD system includes two separation units, featuring it as an energy demanding process. Aiming to improve the energy efficiency performance, the application of an internally heat integrated distillation column (HIDiC) approach [7–9] in PSD system is proposed in open literature. Huang et al. [10] have designed the HIDiC column by thermally coupling the bottom of low pressure (LP) column with the top of high pressure (HP) column for dealing with the separation of a minimum-boiling azeotrope. As far as the fractionation of maximum-boiling azeotrope is concerned, the reverse combination of thermal pairing is suggested by them. Through the separation of a binary azeotropic mixture (acetonitrile/water), it has been
308
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
demonstrated [10] that the general HIDiC (or HIDiC-alone) column helps to improve the process design in not only thermodynamic efficiency but also capital investment. Recently, Mulia-Soto and Flores-Tlacuahuac [6] have extended the HIDiC technology to a PSD system by coupling the whole portion of both columns. The authors have used the Aspen flowsheet simulator for process dynamics and control studies. The major complexity involved in operation and design of the HIDiC configuration is associated with the internal heat integration arranged between the whole/major part of the rectifier and that of the stripper. Therefore, in spite of the encouraging outcomes from the standpoint of both capital investment and energy consumption [11], it has not yet found wide applications in the process industries. Keeping this practical concern in mind, Chen et al. [12] have explored the possibility of using a reduced number of vertically arranged internal heat exchangers (internal HE). Subsequently, Harwardt and Marquardt [13] have formally used the optimization technique for HIDiC to establish a systematic design methodology driven by economics to find the optimal number of internal HE. In this article, we have attempted to reduce the number of internal HE in HIDiC without compromising much of its overall performance in the aspect of both energy consumption and cost. In this contribution, we propose a hybrid heat integration scheme for a PSD column by combining a HIDiC and vapor
recompression column (VRC). There are typically two possible approaches to hybridize the general HIDiC column with the VRC scheme. These approaches are based on the way the thermal coupling is made between: (i) the overhead vapor with the bottom liquid of the same HP column, and (ii) the overhead vapor of HP column with the bottom liquid of LP column. When the HIDiC is in ideal form (i.e., no entropy generation) or it does not require any external heat for the LP column reboiler, the first option is applicable for further reduction of utility consumption in the HP column of PSD configuration. Similarly, for the nonideal HIDiC case (i.e., irreversible process), the second option makes sense in hybridizing the HIDiC with VRC. This work investigates the techno-economic feasibility of a hybrid heat integrated structure for a bioethanol PSD column. For the representative system, the HIDiC is proposed to hybridize by thermally coupling the overhead vapor and bottom liquid of the same HP column under the framework of VRC system, yielding an ideal HIDiC–VRC configuration. Furthermore, aiming to make this heat integration technology one step ahead toward its implementation in industrial scale, attempts are made to design the hybrid configuration, in which, the HIDiC column includes a small number of internal HE. The performance of all these options is quantified and analyzed in terms of both the energy savings and total annual cost.
Fig. 1. Schematic representation of a conventional PSD column.
309
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
2. Conventional pressure-swing distillation: Bioethanol dehydration
Table 1 System specifications and steady state conditions. Items
2.1. Working principle and configuration
2.2. Modeling, computer simulation and verification The mathematical model for both the conventional and HIDiC of the bioethanol PSD column is derived based on the following assumptions: Perfect mixing and equilibrium on all trays. Negligible tray vapor holdups. Variable liquid holdup in each tray. NRTL model to predict vapor–liquid equilibrium (VLE). Constant stage pressure drop (0.3 kPa) and efficiency (70%) for all trays. Nonlinear Francis-weir formula for liquid hydraulics.
The modeling equations for a typical separation stage (Fig. 2) [15] are shown in the Appendix A. Obviously, the model consists of ordinary differential equations and algebraic equations that correlate the physical properties, tray hydraulics, phase equilibrium, etc. For simulating the PSD model, the optimal column setup is adopted from the literature [6]. Here, computer programs are developed in MATLAB environment. Table 1 reports the process specifications and shows the verification of our developed simulator. Now, the validation of VLE model is shown in Fig. 3. Here, the NRTL model is simulated to predict the VLE of ethanol/water system. The binary interaction parameter values are adopted from Reid et al. [16]. Experimental data available in literature [6] are used to validate the activity-based VLE model. In Fig. 3, one can readily see that the NRTL equations used in the PSD model provide a reasonably good prediction with a root-mean square (RMS) error of 3.1 102.
Vn yn,i
Ln+1 xn+1, i
V Sn , yn,i
Tray
n
(holdup =
F n , zn,i
mn )
L
Sn , xn , i
Low pressure (LP) column Feed flow rate (kmol/h) Feed composition (ethanol/water) Recycle feed rate (kmol/h) Top composition (ethanol/water) Bottom composition Bottoms rate (kmol/h) Distillate rate (kmol/h) Total number of traysa (excluding condenser and reboiler) Fresh feed stage Recycle feed stage Stage pressure drop (kPa) Bottom stage pressure (kPa) Reboiler duty (kW) High pressure (HP) column Feed flow rate (kmol/h) Feed composition (ethanol/water) Top composition Bottom composition Bottoms rate (kmol/h) Distillate rate (kmol/h) Total number of traysa (excluding condenser and reboiler) Feed stage Stage pressure drop (kPa) Bottom stage pressure (kPa) Reboiler duty (kW)
Present work
100 0.2/0.8 79.57 0.861/0.139 0.005/0.995 80.35 99.21 28
100 0.2/0.8 39.848 0.8565/0.1435 0.001054/0.998946 80.114 59.762 28
7 17 0.0 110.325 3320
7 17 0.3 119.025 1050.37
98.21 0.861/0.139 0.827/0.173 0.997/0.003 19.64 79.57 28
59.762 0.8565/0.1435 0.8272/0.1728 0.99623/0.00377 19.976 39.848 28
16 0.0 1012.78 3010
16 0.3 1012.78 850.57
a Tray numbering has started from bottom up (bottommost tray is 1 and topmost tray is 28).
1.0
Experimental VLE Simulated VLE
0.9 0.8
Yethanol (mol fract)
Fig. 1 demonstrates a typical pressure-swing distillation (PSD) column. It is a dual-column process, in which, one column operates at a relatively high pressure than the other one. The purpose of this pressure elevation is to circumvent the azeotropic point. For minimum-boiling systems, high purity product streams are taken out from the bottom of the two distillation columns and the distillate streams are recycled [14]. For the representative system, the ethanol/water azeotrope is not so sensitive to pressure changes. For instance, the azeotropic composition changes from 87.2 mol% ethanol at 1 atm to 79.4 mol% at 10 atm. However, it is observed [6] that this low sensitivity is sufficient enough to achieve a high degree of separation.
Reported Aspen work [6]
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Xethanol (mol fract) Fig. 3. X–Y diagram comparing simulated NRTL model and experimental data taken from literature [6].
It is evident in Table 1 that the PSD system secures the production of ethanol at a purity of more than 99.6 mol%, which is higher than the permissible limit (i.e., 99.5 mol%) recommended to the transportation sector. Moreover, the PSD column also produces a high-purity water (i.e., about 99.9 mol%) from the bottom of LP column, showing a successful and efficient separation of both the constituent components. 3. HIDiC-alone scheme: A conventional approach
Ln xn, i
Vn−1 yn−1, i
Fig. 2. Quantities associated with a typical nth tray.
The HIDiC configuration is built by dividing the distillation tower into two columns, one for each of its rectifying and stripping sections. The stripping column accompanies a reboiler and the rectifying column includes a condenser. A compressor is installed over
310
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
Table 2 Cost estimating formula and parameter value. 1:066 0:802 Column shell: M&S Lc ðcin þ cm cp Þ 280 101:9Dc where Dc is the column diameter (ft), Lc the column height (ft), M&S = 1625.9 [20], and the coefficients cin = 2.18, cm = 3.67, and cp = 1.2 1:55 Lc ðcs þ ct þ cm Þ Column tray: M&S 280 4:7Dc where the coefficients cs = 1, ct = 0, and cm = 1.7 M&S Heat exchanger: 280 101:3A0:65 ðcin þ cm ðct þ cp ÞÞ where the coefficients cin = 2.29, cm = 3.75, ct = 0.1, and cp = 1.35 0:82 ðcin þ ct Þ Compressor: M&S 280 517:9BHP where the coefficient cin = 2.11, and ct = 1.0
the stripper to pressurize the overhead vapor before its entry to the rectifier. Consequently, the rectifying section operates at a relatively high pressure compared to the stripping section, causing a thermal driving force between them. To accomplish the heat transfer from the HP to the LP column, a set of internal heat exchangers (HE) is employed to couple the trays between these two diabatic sections. This internal thermal integration leads to bring the liquid reflux for the rectifier and the vapor boil-up for the stripper, thereby reducing the consumption of thermal utility supplied externally to the condenser and reboiler, respectively. Note that the bottom stream of HP rectifier is further throttled in a pressure reducing valve before recycling back to the LP stripper.
In an adiabatic one-column system, called conventional distillation column (CDiC), a condenser and a reboiler are used to generate external reflux and boil-up, respectively. When the internal heat exchangers in HIDiC are able to alter these external flows, the HIDiC scheme gets its ideal form. Actually, the ideal HIDiC corresponds to a complete reversible operation, in that no entropy is generated and it demarcates the theoretical limit of performance. Obviously, this ideal scheme does not require any reboiler as well as condenser. The model of HIDiC column additionally includes a couple of equations as shown in the Appendix A. These additional modeling equations are basically used to respectively compute the internal liquid and vapor flows through the HP rectifier and the LP stripper, and the heat transfer between these two integrated columns through the internal HE. As explained above, compared to an adiabatic column, its HIDiC counterpart is capable of reducing the consumption of both reboiler duty of LP column and the condenser heat load of HP column. However, the heat integration effectively involves an additional compressor, along with a set of internal HE. Obviously, both of these equipments require an additional capital investment. Moreover, although the heat exchanger set does not require any external medium, but the compressor is driven by externally supplied electricity, which is much more expensive than the thermal utility.
Fig. 4. Schematic representation of the HIDiC-alone of bioethanol PSD column.
311
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
It clearly indicates the necessity of a thorough quantitative analysis to judge the energetic and economic potential of the HIDiC column with reference to a conventional standalone column (i.e., CDiC). In this regard, the estimation of two performance indexes is briefly presented below. 3.1. Performance indexes 3.1.1. Energy consumption The total energy consumed (Q Cons ) by a thermally integrated scheme is determined by adding the reboiler heat load (Q R ) plus three times the compressor duty (Q Comp ). It implies:
Q Cons ¼ Q R þ 3Q Comp
ð1Þ
The factor of three for the compression duty is assumed to convert the compression work into the thermal energy required to produce an equivalent amount of electrical power [17]. It should be highlighted that for the CDiC column, Q Comp ¼ 0. As far as HIDiC is concerned, the following form of correlation is used to estimate the compressor duty [18]:
l l1
h l1 i V nT Pi CR l 1
ð2Þ
In the above equation, the pressure (inlet pressure, P i and outlet pressure, Po ) is in lbf/ft2, and the vapor inflow rate to the compressor, V nT is in ft3/min. Here, the compression ratio (CR) is expressed as:
CR ¼
l=ðl1Þ Po To ¼ Pi Ti
ð3Þ
where T i and T o are the inlet and outlet temperature (K), respectively, with respect to the compressor. We compute the CR from Eq. (3) assuming the requirement of a minimum temperature difference (DT min ) of 20 K for complete condensation of compressed vapor in thermal coupling [18]. The polytropic coefficient, lð¼ C p =C v Þ can be calculated from:
1=ðl 1Þ ¼
X
yi =ðli 1Þ
3.2.1. Selection of total number of internal heat exchangers and their locations As shown in Table 1, the HP column operates at about 10 atm, while the LP column at 1 atm. It clearly indicates that the trays in HP section are hotter than those in LP section. In this regard, Fig. 5a displays a comparative temperature profile before implementing the thermal coupling. Since, there exists a reasonably large thermal driving force along the length of both diabatic sections, one can think of coupling all trays between them. However, as stated previously, the participation of all trays (or most of them) in thermal pairing makes the design and operation of HIDiC a complex one. In fact, this complexity is the main barrier in implementing the HIDiC in industrial practice. Prior to configuring the hybrid HIDiC–VRC scheme, therefore, in this section, we would like to find a small number of internal heat exchangers (HE) and their suitable locations for the HIDiC of PSD column. In order to select the number of heat exchangers, we fix the two criteria: (i) same ethanol productivity and purity (with that of conventional PSD), and (ii) either complete heat reversibility (i.e., no entropy generation) or zero reboiler duty in the LP column. For a
ð4Þ
Here, C p and C v are the heat capacities, and yi the mole fraction of any vapor component i. Knowing the Q Cons for CDiC and its heat integrated counterpart, one can estimate the percent energy savings.
430 420
Tray temperature (K)
Q Comp ¼ 3:03 105
internal heat exchangers. Interestingly, the bioethanol PSD column presented before inherently includes a compressor (CR = 10) and a throttling valve. So, the application of HIDiC to a PSD system (Fig. 4) additionally requires only a couple of heat exchangers to connect the trays between HP and LP columns, indicating a greater possibility of energy efficiency and economic performance improvement over the HIDiC of traditional distillation process.
HP LP
400 390 380 370 360 350 340
3.1.2. Economic evaluation The total annual cost (TAC) that is commonly used to assess the economic feasibility of a heat integrated configuration is expressed as:
2
4
6
8
10 12 14 16 18 20 22 24 26 28 30
Tray number 420
TAC ð$=yrÞ ¼ annual operating cost ðOCÞ
b
410
ð5Þ
The capital cost is computed by summing up the cost of all major equipments. The cost estimating formulas are given in Table 2. The operating cost of the compressor is estimated as suggested by Douglas [18] based on the bhp (=hp/0.9), and a motor efficiency of 0.6. For the sake of simplicity, operating costs are taken to be identical to utility costs, i.e. the number resulting from the summation of electricity (0.084$/kW h), steam (17$/t) and cooling water (0.06$/t) costs for a year having 8000 operating hours [10].
Tray temperature (K)
capital investment ðCIÞ þ payback period ðhÞ
a
410
HP28 LP28 HP15 LP15 HP8 LP8 HP4 LP4
400 390 380 370 360 350 2
4
6
8
10 12 14 16 18 20 22 24 26 28 30
Tray number 3.2. Development of HIDiC-alone scheme for bioethanol separation In comparison with an adiabatic CDiC, its HIDiC counterpart additionally includes a compressor, a throttling valve and a set of
Fig. 5. Comparative tray temperature profiles of HP and LP columns having: (a) no internal HE, and (b) 4 different sets of internal HE (i.e., 28, 15, 8 and 4) [tray numbering has started from bottom up; reboiler as 1 and condenser as 30 in both the columns].
312
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
Table 3 A comparative performance in terms of energy savings and cost.
a
Item
Conventional PSD
HIDiC of PSD
Ideal HIDiC–VRC
28 HE
15 HE
8 HE
4 HE
Total energy consumption (kW) Energy savings (%)
1919.58
869.39 54.71
869.51 54.70
870.02 54.68
870.065 54.67
328.66 82.88
Operating cost ($/yr) LP Column Condenser Duty (kW) Cost of CW
832.73 341025.88
849.05 347710.05
870.62 356547.26
891.24 364988.06
901.23 369079.24
901.23 369079.24
Reboiler (LP column) Duty (kW) Cost of steam
1050.37 258068.18
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
0.0 0.0
HP Column Condenser Duty (kW) Cost of CW
1090.12 446467.41
968.02 396432.46
962.88 394328.25
969.42 397005.66
970.62 397496.42
0.0 0.0
Trim-reboiler Duty (kW) Cost of steam
850.58 322831.19
850.58 322738.73
850.58 322505.69
850.58 322510.58
850.58 322526.18
302.98 115715.54
Comp1 Duty (kW) Cost of electricity
6.21 6955.2
6.27 7022.4
6.31 7067.2
6.48 7259.8
6.495 7274.4
6.495 7274.4
Comp2 Duty (kW) Cost of electricity Total operating cost ($/yr) Utility cost savings (%)
1375347.86
1073903.64 21.92
1080448.4 21.44
1091764.1 20.62
1096576.24 20.27
2.065 2312.8 494381.98 64.05
Operating conditions UA (kW/K stage) Ethanol purity, mol% Ethanol productivity (kmol/h)
99.623 19.976
0.55 99.628 19.983
1.043 99.627 19.974
2.3 99.627 19.96
4.135 99.626 19.94
4.135 99.626 19.94
Reflux ratio LP column HP column Stage pressure drop (kPa)
1.218 2.448 0.3
1.218 2.44 0.3
1.221 2.42 0.3
1.224 2.40 0.3
1.226 2.38 0.3
1.226 2.38 0.3
Capital cost ($) LP Column Condenser Reboiler Column Trays
260437.89 250688.90 733268.36 2081624.16
263744.21 0.0 733268.36 2081624.16
268082.53 0.0 733268.36 2081624.16
272190.84 0.0 733268.36 2081624.16
274170.13 0.0 733268.36 2081624.16
274170.13 0.0 733268.36 2081624.16
HP Column Condenser Trim-reboiler Bottom reboiler Column Trays Internal heat exchanger Comp1 Comp2 Total capital cost ($) TACa ($/yr) TAC savings (%)
155190.06 218562.62 512260.01 1235673.88 0.0 50165.81 0.0 5497871.69 3207971.76
143651.63 218562.62 512260.01 1235673.88 1458997.03 50562.92 0.0 6698344.82 3306685.25 ve
143155.55 218562.62 512260.01 1235673.88 1173198.94 50827.28 0.0 6416653.33 3219332.84 ve
143786.60 218562.62 512260.01 1235673.88 898191.32 51947.45 0.0 6147505.23 3140932.51 2.09
143902.11 218562.62 512260.01 1235673.88 738060.10 52046.04 0.0 5989567.41 3092818.71 3.59
0.0 113213.58 247008.99 512260.01 1235673.88 738060.10 52046.04 20337.95 6007663.19 2496936.38 22.16
Payback period = 3 yr.
meaningful comparison between a conventional column and its heat integrated schemes, it is reasonable to force all the schemes to meet both the above criteria. It is worth noticing that in all configurations, two internal heat exchangers (HEs) are targeted to keep fixed, one at the top and other one at the bottom. The former one is arranged between the tops of the two diabatic sections, targeting the HP column to operate in a reflux-free operation mode [19]. On the other hand, the bottom external heat exchanger is accommodated between the bottoms of those sections, enabling the LP column to run in a reboil-free operation mode. Furthermore, the intermediate heat
exchangers are employed to generate internal flows in both the columns. A systematic comparison is presented in Table 3 between various forms of HIDiC, differing with respect to the number of internal heat exchangers used and their locations. In the first case, total 28 internal heat exchangers are employed to connect all 28 trays of both the diabatic sections. Note that both the HP and LP columns have the same number of trays (i.e., 28), excluding the reboiler and condenser. For the case of 15 heat exchangers, each of these heat exchangers of the HIDiC is coupled in every alternate tray, except at the top where two heat exchangers are consecutively
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
Internal vapor flow rate (kmol/s)
0.045
28HE 15HE 8HE 4HE
0.040 0.035 0.030 0.025 0.020 0.015 0.010 0.005 0.000
2
4
6
8
10 12 14 16 18 20 22 24 26 28
Tray number Fig. 6. Internal vapor flow rate profile through the HP column.
connected with 27th as well as 28th trays. In the next, 8 heat exchangers are configured and each of them is connected in every 4th tray. The last scheme includes 4 heat exchangers and these are coupled with following trays of both the columns: 1, 10, 19 and 28. Note that except the top and bottom internal heat exchangers (HE),
313
all other HE and their locations are selected arbitrarily in the four HIDiC configurations stated above. It is evident in Fig. 5b that all four cases have very close thermal driving force between the trays of HP and LP columns. It is further interesting to observe the internal fluid flow profile in the HP or LP columns. As shown the vapor flow rate through the HP column in Fig. 6, there are several peaks depending on their number of internal HE, except a breakthrough associated with the feed tray (i.e., 16). Actually, the peaks are involved with those trays which are connected with the internal HE. It should be noted that the difference existed between the vapor flow rates in the four HIDiC configurations (Fig. 6) causes a difference in separation effect. As shown in Table 3, all four options of HIDiC having a different number of heat exchangers and their locations achieve very close ethanol productivity (19.97 kmol/h) and purity (99.6%) with that of the conventional PSD scheme, indicating the validation of first criterion. As far as second criterion is concerned, all these configurations achieve zero reboiler duty in the LP column by mainly adjusting the UA of internal heat exchanger. It should be noted that in each case, the UA is taken same for all internal heat exchangers. Since all four options meet the second criterion, we obtain a close energy savings (54.7%) for all of them but a wide variation in economic performance. In Table 3, one can readily find that the HIDiC with 4 heat exchangers shows the best performance securing an energy and TAC savings of about 54.7% and 3.59%, respectively.
Fig. 7. Schematic representation of the proposed hybrid HIDiC–VRC of PSD column.
314
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
In the subsequent section, this HIDiC column is used to propose a hybrid heat integration scheme. 4. Hybrid HIDiC–VRC scheme As shown in Fig. 1, the PSD system comprises of two columns; both of which have a reboiler and a condenser. For improving the energetic potential of the HIDiC-alone scheme, there is a scope of further intensification by introducing the VRC in the HIDiC of PSD configuration. This hybridization can be accomplished in two modes, which differ in the way the thermal coupling is made between (i) the overhead vapor (heat source) and the bottom liquid (heat sink) of the same HP column, and (ii) the overhead vapor of HP column (heat source) and the bottom liquid of LP column (heat sink). For the case of HIDiC, which theoretically achieves a condition of zero reboiler duty with respect to the LP column of PSD system, the first option should be adopted. As far as nonideal HIDiC is concerned, the VRC can be suitably hybridized with HIDiC by adopting the second option. It is worth noticing that both these hybrid structures have the potential of improving energy efficiency performance compared to the HIDiC-alone. However, for economic feasibility of both the hybrid HIDiC–VRC columns, it is recommended to perform the cost analysis. At this stage, it should be pointed out that for the hybrid configuration accompanying with the HIDiC (i.e., under first option), the additional capital investment involved for the VRC can be compensated by the thermal utility consumption saved in the HP column of PSD. On the other hand, the economic performance of the hybrid scheme under second option improves with the increase of the degree of heat irreversibility existed in the LP column of nonideal HIDiC. 4.1. Development of a hybrid HIDiC–VRC scheme for bioethanol separation This section deals with the development of a hybrid heat integrated configuration by combining the HIDiC with 4 HE and the VRC scheme. The HIDiC of the representative bioethanol PSD column does not require any external heat for the LP column reboiler, leaving no option open for further coupling between the overhead vapor of HP column and the bottom liquid of LP column (i.e., second option). In this work, therefore, an alternative thermal integration approach is explored between the overhead vapor (at 421.99 K) and bottom liquid (at 422.97 K) of the same HP column under the VRC framework (i.e., first option). As assumed the requirement of a thermal driving force of at least 20 K [18] for complete condensation of the compressed vapor, the inclusion of VRC is inevitable for the sample bioethanol dehydration system. By this way, the HIDiC scheme is hybridized with the VRC. Fig. 7 illustrates the hybrid HIDiC–VRC column. For the VRC system, the compression ratio (CR) is estimated for Comp2 as 1.23 using Eq. (3). This compressed vapor releases latent heat, amounting 547.6 kW [=(850.58–302.98) kW], in the bottom reboiler of HP column for liquid reboiling. Additionally, the HP column of this hybrid configuration does not require the top vapor condensation, thereby involving no operating and capital investment for the respective overhead condenser. It is now obvious that the combined HIDiC–VRC scheme does not involve any reboiler for the LP column and condenser for the HP column, and their respective heat duties. Therefore, the resulting structure can be referred to as the ideal HIDiC–VRC column. This hybrid HIDiC–VRC confirmation, as shown in Table 3, achieves the best performance compared to all other schemes discussed for the representative PSD column, securing a 82.88% savings in utility consumption, a 64.05% savings in utility cost and a 22.16% savings in TAC.
5. Conclusions The use of internal heat exchangers connecting all or most of the trays between two diabatic sections is remained as a major challenge in implementing the HIDiC technology. Keeping this practical concern in mind, it is attempted to design the HIDiC column with a reduced number of internal heat exchangers without compromising much of its overall performance in the aspect of both energy consumption and cost. In this contribution, we propose a hybrid heat integration scheme for a bioethanol PSD column by combining a HIDiC with reduced number of internal HE and VRC. Since the HIDiC of the sample system requires zero reboiler duty with respect to its LP column, it is not really meaningful to have further intensification between two diabatic sections. However, there exists a possibility of further reduction of utility consumption in the HP column of PSD configuration. With this objective, we introduce a thermal integration between the overhead vapor (heat source) and bottom liquid (heat sink) in the same HP column under the framework of VRC scheme. By this way, the VRC is proposed to hybridize with the HIDiC in a PSD system, eventually yielding an ideal HIDiC– VRC configuration. For the representative bioethanol dehydration system, this hybrid HIDiC–VRC configuration outperforms the HIDiC-alone and conventional PSD system, showing a promising performance from the standpoint of energy (82.88%), utility cost (64.05%) and TAC savings (22.16%). Appendix A A.1. Modeling equations of a distillation tray (CDiC and HIDiC) As depicted in Fig. 2, the nth tray of a distillation column is fed with a liquid feed mixture. Side streams are withdrawn as a liquid as well as a vapor. The modeling equations can be derived in the following forms [15]: Total mole balance
_ n ¼ Lnþ1 þ V n1 þ F n ðLn þ SLn Þ ðV n þ SVn Þ m
ðA1Þ
Component mole balance
_ n x_ n;i ¼ Lnþ1 xnþ1;i þ V n1 yn1;i þ F n zn;i ðLn þ SLn Þxn;i m ðV n þ SVn Þyn;i
ðA2Þ
Energy balance
_ n H_ Ln ¼ Lnþ1 HLnþ1 þ V n1 HVn1 þ F n HFn ðLn þ SLn ÞHLn m ðV n þ SVn ÞHVn Q n
ðA3Þ
Equilibrium
yn;i ¼ kn;i xn;i ¼ cn;i
P 0n;i xn;i Pt
ðA4Þ
Summation NC X xn;i ¼ 1
ðA5aÞ
i¼1 NC X yn;i ¼ 1
ðA5bÞ
i¼1
Here, x denotes the liquid phase composition, y the vapor phase composition, z the feed composition, L the liquid flow rate, V the vapor flow rate, k the vapor–liquid equilibrium constant, H the enthalpy, N C the total number of components, P t the total pressure, P0 the vapor pressure, Q the heat loss, F the feed flow rate, S the flow rate of side stream, and c the activity coefficient. The
B. Kiran, A.K. Jana / Separation and Purification Technology 142 (2015) 307–315
subscript/superscript n indicates the tray index, i the component index, F the feed, L the liquid stream and V the vapor stream. In the computer simulation, algebraic form of equations [15] has been solved to compute the vapor and liquid enthalpies. Note that if there are no side draws, then we consider SL ¼ SV ¼ 0. Negligible heat loss from a stage to the surroundings (Q n ¼ 0) is also assumed. Now, it is indeed straightforward to extend this modeling approach to all other trays. A.2. Internal liquid and vapor flows (heat integrated schemes) Supposing the thermal coupling between jth stage of LP column with jth stage of HP column through an internal heat exchanger (HE), we have:
Q HE;j ¼ UAðT j;HP T j;LP Þ
ðA6Þ
where Q HE refers to the rate of heat transfer through the internal HE. Now, the internal flow rates in the LP and HP columns are given as: LP column
Q HE;j V LP j ¼ PN c ð i¼1 yi ki Þj
ðA7Þ
HP column
Q HE;j LHP ¼ PN c j ð i¼1 xi ki Þj
ðA8Þ
Here k represents the latent heat. References [1] I. López, J. Andreu, S. Ceballos, I. Martínez de Alegría, I. Kortabarria, Review of wave energy technologies and the necessary power-equipment, Renew. Sustain. Energy Rev. 27 (2013) 413–434. [2] P.S. Nigam, A. Singh, Production of liquid biofuels from renewable resources, Prog. Energy Combust. Sci. 37 (2011) 52–68.
315
[3] K. Aleklett, C.J. Campbell, The peak and decline of world oil and gas production, Miner. Energy 18 (2003) 35–42. [4] C.A. Cardona, O.J. Sánchez, Fuel ethanol production: process design trends and integration opportunities, Biores. Technol. 98 (2007) 2415–2457. [5] J. Regalbuto, An NSF perspective on next generation hydrocarbon biorefineries, Comput. Chem. Eng. 34 (2010) 1393–1396. [6] J.F. Mulia-Soto, A. Flores-Tlacuahuac, Modeling, simulation and control of an internally heat integrated pressure-swing distillation process for bioethanol separation, Comput. Chem. Eng. 35 (2011) 1532–1546. [7] X. Liu, J. Qian, Modeling, control, and optimization of ideal internal thermally coupled distillation columns, Chem. Eng. Technol. 23 (2000) 235–241. [8] Y. Zhu, X. Liu, Dynamics and control of high purity heat integrated distillation columns, Ind. Eng. Chem. Res. 44 (2005) 8806–8814. [9] B. Kiran, A.K. Jana, Introducing vapor recompression mechanism in heatintegrated distillation column: impact of internal energy driven intermediate and bottom reboiler, AIChE J. 61 (2015) 118–131. [10] K. Huang, L. Shan, Q. Zhu, J. Qian, Adding rectifying/stripping section type heat integration to a pressure-swing distillation (PSD) process, Appl. Therm. Eng. 8 (2008) 923–932. [11] K. Naito, M. Nakaiwa, K. Huang, A. Endo, K. Aso, T. Nakanishi, T. Nakamura, H. Noda, T. Takamatsu, Operation of a bench-scale ideal heat integrated distillation column (HIDiC): an experimental study, Comput. Chem. Eng. 24 (2000) 495–499. [12] H. Chen, K. Huang, S. Wang, A novel simplified configuration for an ideal heatintegrated distillation column (ideal HIDiC), Sep. Purif. Technol. 73 (2010) 230–242. [13] A. Harwardt, W. Marquardt, Heat-integrated distillation columns: vapor recompression or internal heat integration?, AIChE J 58 (2012) 3740–3750. [14] W.L. Luyben, Comparison of extractive distillation and pressure-swing distillation for acetone/chloroform separation, Comput. Chem. Eng. 50 (2013) 1–7. [15] A.K. Jana, Chemical Process Modelling and Computer Simulation, PrenticeHall, New Delhi, 2011. [16] R.C. Reid, J.M. Prausnitz, B.E. Poling, The Properties of Gases and Liquids, McGraw-Hill, New York, 1987. [17] K. Iwakabe, M. Nakaiwa, K. Huang, T. Nakanishi, A. Røsjorde, T. Ohmori, A. Endo, T. Yamamoto, Energy saving in multicomponent separation using an internally heat-integrated distillation column (HIDiC), Appl. Therm. Eng. 26 (2006) 1362–1368. [18] J.M. Douglas, Conceptual Design of Chemical Processes, McGraw-Hill, New York, 1988. [19] X. Zhang, K. Huang, H. Chen, S. Wang, Comparing three configurations of the externally heat-integrated double distillation columns (EHIDDiCs), Comput. Chem. Eng. 35 (2011) 2017–2033. [20] Economic Indicators, Chem Eng August 2011, pp. 56.