Applied Energy xxx (2010) xxx–xxx
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Significant thermal energy reduction in lactic acid production process Iqbal M. Mujtaba a,⇑, Elmahboub A. Edreder b, Mansour Emtir c a
School of Engineering, Design and Technology, University of Bradford, West Yorkshire BD7 1DP, UK Libyan Petroleum Institute, P.O. Box 6431, Tripoli, Libyan Arab Jamahiriya c Department of Chemical Engineering, Academy of Graduate Studies, Tripoli, Libyan Arab Jamahiriya b
a r t i c l e
i n f o
Article history: Received 9 August 2010 Received in revised form 29 September 2010 Accepted 19 November 2010 Available online xxxx Keywords: Lactic acid Hydrolysis Batch reactive distillation Optimisation Reflux ratio policy Energy reduction
a b s t r a c t Lactic acid is widely used as a raw material for the production of biodegradable polymers and in food, chemical and pharmaceutical industries. The global market for lactic acid is expected to reach 259 thousand metric tons by the year 2012. For batch production of lactic acid, the traditional process includes the following steps: (i) esterification of impure lactic acid with methanol in a batch reactor to obtain methyl lactate (ester), (ii) separation of the ester in a batch distillation, (iii) hydrolysis of the ester with water in a batch reactor to produce lactic acid and (iv) separation of lactic acid (in high purity) in a batch distillation. Batch reactive distillation combines the benefit of both batch reactor and batch distillation and enhances conversion and productivity (Taylor and Krishna, 2000 [1]; Mujtaba and Macchietto, 1997 [2]). Therefore, the first and the last two steps of the lactic acid production process can be combined together in batch reactive distillation (Fig. 1) processes. However, distillation (batch or continuous) is an energy intensive process and consumes large amount of thermal energy (via steam). This paper highlights how significant (over 50%) reduction in thermal energy consumption can be achieved for lactic acid production process by carefully controlling the reflux ratio but without compromising the product specification. In this paper, only the simultaneous hydrolysis of methyl lactate ester and the separation of lactic acid using batch reactive distillation is considered. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction Taylor and Krishna [1] provided a comprehensive list of research on the use of reactive distillation (batch and continuous) system for producing many useful products/chemicals via a number of chemical reactions and distillations. 1.1. Lactic acid production process Esterification of impure lactic acid with methanol has been considered by several researchers in the past to obtain methyl lactate ester which is then separated by distillation. The distilled ester is then hydrolyzed into pure lactic acid (Fig. 1). Choi and Hong [3] used two reactors and two batch distillation columns to carry out the esterification and hydrolysis reactions and achieved relatively pure lactic acid. Kim et al. [4] considered a batch reactive distillation with esterification and hydrolysis for the recovery of lactic acid and studied the dynamics of batch reactive distillation of lactic acid in terms of instantaneous rate of esterification reaction. They also compared semi-batch operation with the batch mode with continuous feeding of methanol. Kumar ⇑ Corresponding author. Tel.: +44 1274 233645. E-mail address:
[email protected] (I.M. Mujtaba).
et al. [5] explored and investigated batch reactive distillation strategy involving experimental esterification and hydrolysis reaction for recovery of pure lactic acid. They studied the effect of operating parameters such as feed concentration, mole ratio; catalyst loading, and boilup rate on the recovery of lactic acid. Li et al. [6] considered the esterification of diluted lactic acid with methanol in a reactor followed by hydrolysis of methyl lactate in continuous column to achieve pure lactic acid. Kumar et al. [7] carried out both experiments and simulation of a continuous reactive distillation process for esterification of lactic acid and hydrolysis of methyl lactate to get pure lactic acid.
1.2. This work It is clear from the previous researches that most of the work has been focused on experiments or simulation to recover lactic acid. No work has focused on the thermal energy requirement or discussed about the possibility of energy saving in lactic acid production process. Therefore, this is the main focus of this paper. Here, energy minimisation (or savings) is achieved via minimisation of production time which is obtained by optimising reflux ratio in a batch reactive distillation process. Successive Quadratic Programming (SQP) based techniques [8] is used for solving the optimisation problem.
0306-2619/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.apenergy.2010.11.031
Please cite this article in press as: Mujtaba IM et al. Significant thermal energy reduction in lactic acid production process. Appl Energy (2010), doi:10.1016/j.apenergy.2010.11.031
2
I.M. Mujtaba et al. / Applied Energy xxx (2010) xxx–xxx
Nomenclature LD distillate flow rate (kmol/h) Ha, Hc accumulator and condenser holdup respectively (kmol) plate and reboiler holdup respectively (kmol) Hj, HN hL, hV liquid, vapour enthalpy (kJ/kmol) L, V liquid, vapour flow rates in the column (kmol/h) N number of plates QC, QR condenser or reboiler duty (kJ/h) T, P temperature (K), pressure (bar) K vapour–liquid equilibrium constant r reaction rate t batch time (h) R1, R2, etc. Reflux ratio in time intervals 1, 2, etc. t1, t2, etc. length of interval 1, 2, etc. (h)
x, y xa xD Dn D R
Superscripts and subscripts i component number j stage number
(note the latent heat of vaporisation is higher for highboiling component compared to that of a low boiling component). (b) Mode 2 – Constant vapour boilup rate (V): In this mode, the heat supply to the reboiler gradually increases to maintain the constant vapour boilup rate. This is due to increase of the latent heat of vaporisation of the reboiler mixture. (c) Mode 3 – Constant condenser vapour load (Vc): In this mode, the vapour load to the overhead condenser is kept constant. Also, in this mode the heat supply to the reboiler (Qr) gradually increases.
2. Energy consumption Traditionally, the most popular type of batch distillation column is conventional (regular) batch column (CBD) as shown in Fig. 1. It consists of a bottom receiver/reboiler, rectifying column (either a tray or packed column) placed over the reboiler, connected to a total condenser or a partial condenser system and distillate receivers. In this column, the charge is loaded into the reboiler at the beginning of the process and heated to its boiling point. Vapour flows upwards in the column and condenses at the top. After some time, a part of the overhead condensate is withdrawn continuously as distillate, and the other part is returned to the column section as reflux (note: the ratio of the liquid being refluxed to the column and the condenser vapour load defines the internal reflux ratio which is used in the work). The liquid in the reboiler is increasingly depleted of the more volatile components and the boiling point of the mixture gradually increases. Also, as the amount of liquid in the reboiler decreases, the concentration of high-boiling components increases. In batch distillation literature [8,9], it can be found that a batch distillation column operates under the following modes: (a) Mode 1 – Constant reboiler duty (Q): Heat supply (usually in the form of steam) to the reboiler is constant over the batch time (production time). In this mode, the vapour boilup rate gradually decreases as the boiling point gradually increases
liquid or vapour composition (molefraction) accumulated distillate composition (molefraction) instant distillate composition (molefraction) change in moles due to chemical reaction amount of product (kmol) reflux ratio
The total energy requirement over the batch time (i.e. over the production time) for any of these operating modes can be given by:
Mode 1 : Q Total ¼ Q tf Z tf k dt Mode 2 : Q Total ¼ V 0 Z tf Mode 3 : Q Total ¼ Q r dt
ð1Þ ð2Þ ð3Þ
0
where k is the instantaneous value of the latent heat of vaporisation which is a function of reboiler composition which changes as the batch progresses in time. Qr is the instantaneous value of the heat supply needed to maintain constant Vc. Note, Vc is always smaller
Esterfication
Hydrolysis MeOH (Unreacted)
H 2O
MeOH
ML
LA + MeOH
H 2O
LA (Final product) Make up H2O Hydrolysis Lactic acid (LA) + Methanol (MeOH )
Methyl lactate (ML) + Water (H 2O) Esterification
Fig. 1. Lactic acid production process by batch reactive distillation.
Please cite this article in press as: Mujtaba IM et al. Significant thermal energy reduction in lactic acid production process. Appl Energy (2010), doi:10.1016/j.apenergy.2010.11.031
3
I.M. Mujtaba et al. / Applied Energy xxx (2010) xxx–xxx
than V and depends on the mass transfer rate and column pressure drop. Whatever the mode of operation is chosen, clearly, the amount of energy savings is directly dependent on the reduction in batch time without compromising the product specification.
to maximise the conversion to lactic acid. Therefore, for a given column configuration (in terms of number of stages), the minimum operation time is obtained by optimising the reflux ratio profile. The optimisation problem can be described as: given
the column configuration, the feed mixture, condenser vapour load, product purity and the amount of bottom product (lactic acid) optimal reflux ratio which governs the operation the operation time
3. Production period determine Generally the production period starts when distillate removal from the process is begun. The operation in the product period and its duration depends on the requirements of the product. This period can be operated under the following conditions: Policy-1: The start-up period is ended when the desired distillate purity is reached. Product take off is started and the product is collected at constant composition by varying the reflux ratio until a specified amount of distillate has been collected. This type of operation is known as ‘‘variable reflux operation’’ or ‘‘constant distillate composition operation’’. In this mode of operation the reflux ratio is such that always produces on-specification material. Policy-2: The total reflux start-up period is ended when the unit reaches it steady state. Product is collected at some constant finite reflux ratio until the accumulated product composition reaches its desired purity. This type of operation is known as ‘‘constant reflux operation’’. Under this operation mode the column is operated on a fixed reflux ratio for the whole fraction (cut), producing better than specification material at the beginning and distillate below specification at the end of the fraction. The above two types of operation policies are referred to as ‘‘conventional’’ method of operation in the literature although none of them addresses the issue of energy consumption or energy savings. Note, the second type of operation is the easiest from control point of view and is the most used in industries [8,9]. 4. Energy reduction As mentioned in Section 2, minimisation of batch time will minimise the total energy consumption of the process. Therefore, a third type of operation (Policy-3) which is a trade off between the above two types of operation (described in Section 3) will be considered in this work where an optimal reflux ratio policy will be chosen so that a given amount of lactic acid product with a given purity can be achieved in minimum time to minimise the energy consumption of the process. Operating Policy-2 will also be considered in this work to demonstrate significant energy savings between Policy-2 and Policy-3. Note, Masoud and Mujtaba [10] utilised operating Policy-3 in Inverted Batch Distillation column to study the impact of this operation policy on the design and energy consumption of process for binary (no-reactive) separations. Note, huge literature is available which demonstrates how reflux ratio minimises batch time for a given design of the column and for a fixed vapour load operation. Readers are directed to Mujtaba [8] and Diwekar [9] for a comprehensive review on this. In this work, the hydrolysis reaction is considered for the lactic acid production. For a given amount of lactic acid product with a range of lactic acid purity (from 0.8 to 0.95 molefraction of lactic acid) minimum time optimisation problems are formulated incorporating a mathematical model for the batch reactive distillation process within gPROMS software [11]. The lactic acid being the heaviest in the reaction mixture, reflux ratio policy plays an important role in removing the light product methanol from the system while ensuring the presence of both reactants in the reaction zone
so as to minimise subject to
equality and inequality constraints (e.g. model equations)
Mathematically, the optimisation problem (OP) can be represented as:
Min
tf
RðtÞ
s:t: Process Model Equations ðequality constraintsÞ
B¼B x3 6 x3 6 þe
ð4Þ
ðinequality constraintÞ ðinequality constraintÞ
where B, x3 are the amount of bottom product (lactic acid) and composition at the final time tf ( denotes that the B and x3 are specified), R(t) is the reflux ratio profile which is optimized. e is the small positive number of the order of 103. The process model in the form of differential algebraic equations (DAE) (see Section 5) acts as equality constraint to the optimisation problem. Single and multiple reflux ratio strategies are used, yielding an optimal reflux ratio policy. Single-reflux ratio policy will represent operating Policy-2 described in Section 3. For multiple reflux ratio policy, the total batch time is divided into multiple intervals. Within each interval, the reflux ratio (assumed constant) and the interval length are optimized. The optimisation problems results in a Non Linear Programming (NLP) problem, which is solved using an SQP-based optimisation technique available within gPROMS [11]. 5. Process model A schematic diagram of batch reactive distillation column is shown in Fig. 2 and a dynamic model for the process is shown in Fig. 3 [12]. Condenser and Accumulator
V1, y1 L1, x1
Vj+1, yj+1 Lj-1, x j-1
Plates
Vj, yj Lj, x j
VN, yN
Reactor /Reboiler
Fig. 2. Conventional batch reactive distillation column.
Please cite this article in press as: Mujtaba IM et al. Significant thermal energy reduction in lactic acid production process. Appl Energy (2010), doi:10.1016/j.apenergy.2010.11.031
4
I.M. Mujtaba et al. / Applied Energy xxx (2010) xxx–xxx
Internal Plates, j = 2, N-1
Component Mass Balance:
Total Mass Balance:
a) Accumulator:
0 = L j −1 + V j +1 − L j − V j + Δ n j H j
Ha
Component Mass Balance: Hj
dxji dt
b) Condenser Holdup Tank
= Lj−1 xj−1,i +Vj+1 yj+1,i −Lj xji −Vj yji +rjiHj
Hc
Energy Balance:
0 = V 2 hV 2 − (V 2 + Δn1 H c ) h L1 − Qc
Other Equation
Equilibrium:
L1 = R(V2 + Δ n1H c ) , LD = (V2 + Δ n1Hc )( 1 − R)
,
T1 = T1( xD,i , P )
Restrictions:
∑y
dxDi =V2y2,i + r1,i Hc − (V2 + Δ n1 Hc)xDi dt
Energy Balance:
0 = L j −1 hL j −1 + V j +1 hV j +1 − L j hL j − V j hVj
y ji = K ji x ji
dxa = LD ( xD,i − xa,i ) dt
,
h L1 = h L1( xD,i , T1, P)
Reboiler: j = N
ji = 1
Total Mass Balance: Relations defining physical properties: dH N = LN −1 − VN + ΔnN H N dt
K ji = K ji ( y j , x j , T j , P)
Component Mass Balance:
h L j = h L j ( x j , T j , P ) , hV j = hV j ( y j , T j , P )
r ji = r ji (k ji , x ji ) , Δ n j =
∑r
HN
ji
Condenser and Distillate Accumulator: j=1
dxNi = LN −1 ( xN −1,i − xNi ) − VN ( yNi − xNi ) dt + rNi HN −Δ nN HN x Ni
Energy Balance: L
0 = L N−1(hLN−1 − hLN ) −VN (hVN − h N ) + QR
Accumulator Total Mass Balance: Fig. 3. Model equations.
5.1. Chemical reaction and kinetics The hydrolysis reaction of methyl lactate (ester) can be expressed as:
Methyl lactate ð1Þ þ Water ð2Þ Lactic acid ð3Þ þ Methanol ð4Þ
ð5Þ
Boiling pointðKÞ 417:15ð1Þ 373:15ð2Þ 490:15ð3Þ 337:15ð4Þ
A quasi-homogeneous (QH) activity (ai = ci xi) based kinetic model is used [13] and can be written as:
50:91 r ¼ 1:65 105 exp a1 a2 1:16 106 RT 48:52 a3 a4 exp RT
V
hj ¼ ð6Þ
L hj
¼
X X
V
yi hi
V xi ðhi
ð8Þ ki Þ
ð9Þ
where ki is the latent heat of vaporisation (kJ/kmol) of component i.
5.2. Vapour–liquid equilibrium (VLE) K-values (VLE constants) are computed from Eq. (7) where ci is computed from UNIQUAC equation. The vapour pressure (Psat) of pure components is calculated by using Antoine’s equation. The UNIQUAC binary interaction parameters and Antoine parameters were taken from Sanz et al. [14].
K ¼ ci Psat i =P
The Antoine parameters are given in Table 1. Volume and area parameters for the UNIQUAC equation were taken from the data bank of HYSYS and are also given Table in 1. The UNIQUAC binary interaction parameters are given in Table 2. Vapour phase enthalpies are calculated using empirical equations from Holland [15]. The physical and thermodynamic properties data and enthalpy equations for all pure components are given in Tables 3 and 4. The liquid and vapour enthalpies (hL, hV) which constitute the energy balance equations in the process model (Fig. 3) are usually expressed as a function of liquid/vapour molefractions, temperature and pressure. The liquid phase enthalpies are calculated by subtracting the latent heat of vaporisation from the vapour enthalpies.
ð7Þ
Table 1 Antoine parameters and area and volume parameters. Component
A
B
C
r
q
Methyl lactate (1) Water (2) Lactic acid (3) Methanol (4)
7.24147 7.0436 7.51107 7.21274
2016.46 1636.909 1965.7 1588.63
32.104 48.230 91.021 32.5988
5.95005 0.92000 5.27432 1.4311
5.01723 1.39970 4.47617 1.4320
Please cite this article in press as: Mujtaba IM et al. Significant thermal energy reduction in lactic acid production process. Appl Energy (2010), doi:10.1016/j.apenergy.2010.11.031
5
I.M. Mujtaba et al. / Applied Energy xxx (2010) xxx–xxx Table 2 Binary interaction parameters for UNIQUAC equation.
Methanol–water Methanol–methyl lactate Methanol–lactic acid Water–methyl lactate Water–lactic acid Methyl lactate–lactic acid
Table 5 Summary of optimisation results using reflux Policy-2.
Aij/K
Aji/K
192.6 866.6 322.59 20.05 84.80 367.14
325.0 164.4 17.14 325.31 26.1 302.09
Table 3 Physical and thermodynamic properties for pure components.
Tc (K) M.wt k1 (kJ/kmol) Tb (K)
Methyl lactate (1)
Water (2)
Lactic acid (3)
Methanol (4)
584.0 104.11 38,177 417.95
647.3 18.02 40,651 373.15
627.0 90.08 54,670 490.0
512.6 32.04 35,290 337.8
*
Purity of lactic acid, x3 (molefraction)
Minimum batch time, tf (h)
Optimum reflux ratio
Conversion (%)
Total energy, QTotal (m kJ)
0.80 0.90 0.95*
14.88 46.04 –*
0.933 0.973 –*
77.7 86.9 –*
1.722 5.429 –*
No results obtained.
Table 6 Summary of optimisation results using reflux Policy-3 (two intervals). x3
t1, R1
tf, R2
Conversion (%)
QTotal (m kJ)
0.80 0.90 0.950
9.54, 0.914 8.66, 0.922 10.55, 0.935
13.72, 0.957 23.95, 0.979 44.73, 0.990
77.7 87.9 92.5
1.588 2.822 5.318
Note: t1 = length of interval 1, R1, R2 = reflux ratio in intervals 1 and 2.
6. Case study 6.1. Specifications Hydrolysis reaction and separation of lactic acid is carried out in a 10 stages column (including condenser and reboiler). The column is operated with energy consumption Mode 3 with condenser vapour load (Vc) of 2.5 (kmol/h). The total column holdup is 4% of the initial feed. Fifty percent of the total holdup is taken as the condenser hold up and the rest is equally divided in the plates. The initial charge to the reboiler is 5 kmol. The feed composition hMethyl Lactate (1), Water (2), Lactic acid (3), Methanol (4)i is: h0.5, 0.5, 0.0, 0.0i. It is assumed that the mixture in the reboiler, in the plates and in the condenser are at the boiling point at time t = 0 (beginning of the process). There is no pre-assigned period of total reflux operation. The column only runs at total reflux if it is demanded by the optimiser. The total amount of desired lactic acid product (in the reboiler) is 2.5 kmol with purity of lactic acid varying from 0.8 to 0.95 (molefraction). Two types of reflux ratio policies are adopted to show how these policies affect the energy consumption of the process without compromising the product specifications. These are: (i) traditional constant reflux ratio Policy-2; (ii) time optimal reflux ratio Policy-3. Note, however, in Policy-2 instead of selecting a random constant reflux ratio, the reflux ratio is optimized to minimise the batch time. In Policy-3, the batch time is discretised into 2–4 intervals and the reflux ratio is assumed to be constant in each intervals. For each interval, both the reflux ratio and the interval length are optimized to minimise the overall batch time. 6.2. Results and discussions Table 5 summarises the optimisation results for Policy-2, in terms of constant (but optimal) reflux ratio, conversion of methyl lactate to lactic acid, minimum operating time and the total energy consumption for a range of product purity. Tables 6–8 summarises
the optimisation results for Policy-3 where time dependent reflux ratio is used. The results of Tables 5–8 show that as the demand on lactic acid purity increases (from 0.8 to 0.95 molefraction), high reflux ratio operation, longer batch time, increased energy consumption are required to keep the reactants together to convert to more of the desired product (lactic acid). Interestingly, with Policy-2, product purity of 0.95 molefraction was not at all achievable with very high reflux ratio and very long batch time such as 200 h. Single constant reflux ratio was not able to simultaneously remove methanol from the system and keep the reactants (methyl lactate and water) together to have more conversion to lactic acid. However, with reflux ratio Policy-3 (Tables 6–8), product purity of 0.95 was achievable. Low reflux ratio in the first interval was able to remove the methanol from the system quickly while the higher reflux ratios in subsequent intervals helped keeping the reactants together in the systems for further reaction. Fig. 4 shows the total energy consumption profile for different product purity and for different reflux ratio policies. Cleary with the increase of product purity, the total energy consumption increases exponentially (Policy-3, with 2–4 intervals). For example, with increase in product purity from 0.8 to 0.9 molefraction (12.5% increase), the increase in total energy consumption with Policy-3 (two intervals) is 77.7%. However, with increase in product purity from 0.9 to 0.95 molefraction (5.5% increase), the increase in total energy consumption with Policy-3 (two intervals) is 88.4%. This is due to having comparatively higher reflux ratio and longer batch time to ensure further reaction to satisfy product purity (compare the reflux ratio profile at different purity in Table 6 and also in Tables 7 and 8). Also note, for high product purity (0.95 molefraction), the column required to operate almost near to total reflux ratio (R = 1) for a longer period in order to keep the reactants together for further reaction and in order to transport lighter components up the column from the reboiler.
Table 4 Vapour enthalpy equations for all pure components. V
h1 ¼ 104:11 ð0:0 2:9834 101 T þ 3:4219 103 T 2 2:43350 106 T 3 þ 1:0223 109 T 4 1:7705 1013 T 5 Þ V
h2 ¼ 2:32 ð0:1545871 105 þ 0:8022526 10 ðT 1:8Þ 0:4745722 103 ðT 1:8Þ2 0:6878047 106 ðT 1:8Þ3 0:1439752 109 ðT 1:8Þ4 Þ V
h3 ¼ 90:0784 ð0:0 3:77592 102 T þ 2:41939 103 T 2 1:38409 106 T 3 þ 4:8395 1010 T 4 7:68398 1014 T 5 Þ V
h4 ¼ 2:32 ð0:1174119 105 þ 0:7121495 10 ðT 1:8Þ þ 0:5579442 102 ðT 1:8Þ2 0:4506170 106 ðT 1:8Þ3 0:2091904 1010 ðT 1:8Þ4 Þ where hV is in kJ/kmol and T is in °K.
Please cite this article in press as: Mujtaba IM et al. Significant thermal energy reduction in lactic acid production process. Appl Energy (2010), doi:10.1016/j.apenergy.2010.11.031
6
I.M. Mujtaba et al. / Applied Energy xxx (2010) xxx–xxx
Table 7 Summary of optimisation results using reflux Policy-3 (three intervals). x3
t1, R1
t2, R2
tf, R3
Conversion (%)
QTotal (m kJ)
0.800 0.900 0.950
2.33, 0.813 4.86, 0.886 6.24, 0.909
2.00, 0.907 3.74, 0.950 11.09, 0.975
11.27, 0.945 21.90, 0.980 37.82, 0.992
77.8 88.1 92.9
1.305 2.581 4.497
Table 8 Summary of optimisation results using reflux Policy-3 (four intervals). t1, R1
t2, R2
t3, R3
tf , R 4
Conversion (%)
QTotal (m kJ)
0.800 0.900 0.950
0.500, 1.00 0.420, 1.00 0.540, 1.00
2.11, 0.800 2.70, 0.835 2.93, 0.848
2.63, 0.918 3.88, 0.937 8.79, 0.964
10.80, 0.935 20.09, 0.976 34.17, 0.989
77.9 88.3 93.3
1.250 2.368 4.063
Total Energy Consumption (mkJ)
x3
Fig. 4. Total energy consumption profile.
For the same product purity, significant reduction in batch time and total energy consumption are possible by selecting appropriate operation strategy in terms of reflux ratio policies. For example, for product purity 0.9 molefraction, a batch time reduction of 56.4% and an energy reduction of 56.4% are possible by having four interval reflux ratio policy (Table 8) compared to one interval reflux ratio policy (Table 5). For the same product purity, even a two interval reflux ratio policy (Table 6) reduces 48% energy consumption. For high product purity such as 0.95 molefraction, reduction in energy consumption of 24% is possible when four interval reflux ratio policy (Table 8) is adopted compared to two interval reflux ratio policy (Table 6). Note, one interval reflux ratio policy (Table 5) does not even achieve the product with the desired specifications. Unlike esterification reaction in conventional batch reactive distillation (Fig. 1) where the reaction product (Methyl lactate ester) is the lightest, the hydrolysis reaction considered in this work produces the reaction product (lactic acid) which is the heaviest in the mixture. The column has to always operate at high reflux ratio so that both the reactants (methyl lactate and water) are available in the reaction zone (reboiler and stages). Low reflux ratio operation will separate the reactants from the system and will thus lower the conversion. Multi-reflux ratio operation (Policy-3) enjoys more freedom to balance between the conversion and product purity.
7. Conclusions The opportunity for thermal energy reduction in lactic acid production in batch reactive distillation process is investigated here. For a given mode of heat supply to the distillation column, minimizing the production time (i.e. batch time) without compromising the product specifications offers potential reduction in thermal energy consumption. For a given column configuration
(i.e. number of stages), reflux ratio policy determines the product amount, quality and batch time. A dynamic optimisation problem incorporating a process model is formulated to minimise the batch time subject to constraints on the amount and purity of lactic acid. Piecewise constant reflux ratio profile (with single and multiple time intervals) is considered for exploring the potentials for thermal energy reduction. A series of minimum time optimisation problems was solved with different values of lactic acid purity ranging from 0.8 to 0.95 molefraction and the impact of time dependant reflux ratio policy on the thermal energy consumption are analyzed. It is observed that about 56% reduction in thermal energy is possible for certain product specification by operating the column with multi-reflux ratio strategy instead of operating the column with traditional single-reflux ratio strategy. It is also observed that 95% purity of lactic acid was not achievable with single-reflux ratio operation policy as the policy was not able to keep the reactants together for a longer period in the column to have further conversion to lactic acid to satisfy the product purity. Multi-reflux operation policies enjoyed additional freedom to have the balance between the conversion and the product purity while reducing significant amount of energy consumption. Finally, the methodology to calculate the energy requirement and to explore the scope of energy savings as presented in this work is general and is applicable to other reaction system. Also energy minimisation is achieved in this work via minimisation of production time. Readers are directed to other methods of energy optimisation, for example, Marshman et al. [16] achieved energy optimisation for a pulp and paper mill via total cost minimisation. References [1] Taylor R, Krishna R. Modelling of reactive distillation – review. Chem Eng Sci 2000;55:5183–229. [2] Mujtaba IM, Macchietto S. Efficient optimisation of batch distillation with chemical reaction using polynomial curve fitting techniques. Ind Eng Chem Res 1997;36:2287–96. [3] Choi J, Hong WH. Recovery of lactic acid by batch distillation with chemical reaction using ion exchange resin. J Chem Eng Jpn 1999;32:184–9. [4] Kim JY, Kim YJ, Hong WH, Wozny G. Recovery process of lactic acid using two distillation columns. Biotechnol Bioprocess Eng 2000;5:196–201. [5] Kumar R, Mahajani SM, Nanavati H, Noronha SB. Recovery of lactic acid by batch reactive distillation. J Chem Technol Biotechnol 2006;81:1141–50. [6] Li MA, Yang Z, Jichu Y. Purification of lactic acid by heterogeneous catalytic distillation using ion-exchange resins. Chin J Chem Eng 2005;13:24–31. [7] Kumar R, Nanavati H, Noronha SB, Mahajani SM. A continuous process for the recovery of lactic acid by reactive distillation. J Chem Technol Biotechnol 2006;81:1767–77. [8] Mujtaba IM. Batch distillation: design and operation. London: Imperial College Press; 2004. [9] Diwekar UM. Batch distillation. Simulation, optimal design and control. New York: Taylor & Francis; 1995. [10] Masoud AZ, Mujtaba IM. Effect of operating decisions on the design and energy consumption of inverted batch distillation column. Chem Prod Proc Mod 2009;4(1):1–10 [Article 35].
Please cite this article in press as: Mujtaba IM et al. Significant thermal energy reduction in lactic acid production process. Appl Energy (2010), doi:10.1016/j.apenergy.2010.11.031
I.M. Mujtaba et al. / Applied Energy xxx (2010) xxx–xxx [11] gPROMS. Introductory user guide. Process System Enterprise Ltd. (PSE); 2004. [12] Edreder EA, Mujtaba IM, Emtir MM. Optimisation of design, operation and scheduling of batch reactive distillation process with strict product specification and fixed product demand using gPROMS. Comput Aided Chem Eng 2009;26:411–5. [13] Sanz MT, Murga R, Beltran S, Cabezas JL, Coca J. Kinetic study for the reactive system of lactic acid esterification with methanol: methyl lactate hydrolysis. Ind Eng Chem Res 2004;43:2049–53.
7
[14] Sanz MT, Beltran S, Calvo B, Cabezas JL. Vapor liquid equilibria of the mixtures involved in the esterification of lactic acid with methanol. J Chem Eng Data 2003;48:1446–52. [15] Holland CD. Fundamentals of multicomponent distillation. New York: McGraw-Hill; 1981. [16] Marshman DJ, Chmelyk T, Sidhu MS, Gopaluni RB, Dumont GA. Energy optimization in a pulp and paper mill cogeneration facility. Appl Energy 2010;87:3514–25.
Please cite this article in press as: Mujtaba IM et al. Significant thermal energy reduction in lactic acid production process. Appl Energy (2010), doi:10.1016/j.apenergy.2010.11.031
