Process intensification in biodiesel production with energy reduction by pinch analysis

Energy xxx (2014) 1e15

Contents lists available at ScienceDirect

Energy journal homepage: www.elsevier.com/locate/energy

Process intensification in biodiesel production with energy reduction by pinch analysis Valentin Ples¸u b, Joan Subirana Puigcasas a, Guillem Benet Surroca a, Jordi Bonet a, b, Alexandra E. Bonet Ruiz a, b, *, Alexandru Tuluc b, Joan Llorens a a b

University of Barcelona, Department of Chemical Engineering, 1, Martí i Franqu es Street, E-08028 Barcelona, Spain University POLITEHNICA of Bucharest, Centre for Technology Transfer in Process Industries (CTTPI), 1, Gh. Polizu Street, RO-011061 Bucharest, Romania

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 January 2014 Received in revised form 30 October 2014 Accepted 5 November 2014 Available online xxx

The overall process of biodiesel synthesis from vegetable oil and methanol is spontaneous according to Gibbs energy values. Therefore, a classical process scheme consisting of reactor followed by distillation columns train is grouped in a single hybrid reactive extraction column. Minimum energy consumption is calculated using Pinch Analysis, taking into account the minimum energy thermodynamically required by process units, e.g. distillation. Process Integration decreases dramatically the minimum energy requirements. Using Pinch Analysis, a useful tool is provided to calculate the minimum energy requirements of alternative processes, the effect of inclusion of the distillation column is to be underlined. The intensified process provides biodiesel and glycerol valorisation with very low energy consumption. A conceptual design of hybrid reactive extraction column useful for several input oils and fats is proposed, considering first pure triglycerides as raw materials and then complex mixtures of triglycerides as in real oil compositions. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Energy efficiency Energy conservation Heat engine Pinch analysis Short chain fatty acids Biodiesel additives

1. Introduction Biodiesel is one of the main products of the European biorefineries. In European Union, Directive 2009/28/EC require that every year the production of biofuels increases with the aim that by 2020, 10% of transportation fuels to be biofuels. Most of the UE car engines are diesel and therefore biodiesel plays a crucial role. The classical scheme to produce biodiesel including a reactor followed by a train of distillation columns to purify the products and to recover non-reacted methanol, requires a high-energy demand [1]. Great improvements on biodiesel process are required to compete with the mature petro-diesel process. Combining in the same unit reaction and separation as a technique of process intensification proved to be an effective way to reduce the energy requirements for many processes. Gibbs energy provides a valuable guide to identify when process intensification is advantageous. In case of biodiesel synthesis, the overall intensified process does not require energy

* Corresponding author. University POLITEHNICA of Bucharest, Centre for Technology Transfer in Process Industries (CTTPI), 1, Gh. Polizu Street, RO-011061 Bucharest, Romania. Tel.: þ40 21 4023916; fax: þ40 21 3185900. E-mail address: [email protected] (A.E. Bonet Ruiz).

consumption [2]. The transesterification of vegetable oil with methanol is thermodynamically spontaneous unlike mixing. The temperature and excess of methanol are more useful to favour phase mixing than to displace the reaction equilibrium towards biodiesel formation [3]. This statement is in agreement with experimental results, when mixing is produced by ultrasounds [4], or in supercritical conditions [5], or when using reactors that provide an intensive mixing [6]. All these techniques require an excess of methanol. Our research team has recently proposed a novel overall intensified process for triolein transesterification without the need of methanol excess, using a hybrid reactive extraction column that produces pure biodiesel and glycerol [7]. Both reactants flow in counter-current, which is more efficient than other flow configurations such as cross-flow [8]. Methanol is partially soluble in FAME (fatty acid methyl ester) [9], consequently can be recovered in the non-reactive section, using glycerol as extractive agent. Without the use of the non-reactive section, an excess of methanol is required [10]. The capital costs and risks become lower as the number of process units decreases due to the intensification [11]. Several oils can be used as raw material for biodiesel: waste vegetable cooking oils [12], algae oil [13], fish oil [14], Sapindus mukorossi kernel oil [15], Canola or Camelina oil [16], Karanja [17],

http://dx.doi.org/10.1016/j.energy.2014.11.013 0360-5442/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

2

V. Ples¸u et al. / Energy xxx (2014) 1e15

Table 1 Some biodiesel compositions limited by legislation. % wt

EN 14214

Ester Water Impurities Linolenic methyl ester Methanol Monoglycerides Diglycerides Triglycerides Free Glycerine Total Glycerine

>96.5 <500 mg/kg <24 mg/kg <12 <0.2 <0.8 <0.2 <0.2 <0.02 <0.25

Table 3 Mass percent of fatty acid for several kinds of oils. ASTM D6751-08 <0.05% vol

<0.2

<0.02 <0.24

rubber seeds oil [18], castor oil [19], etc. Algae oil is considered a good oil source for biodiesel. To produce 50% of the fuel required in USA, an agriculture area equivalent to 24% of the USA fields should be dedicated to palm oil production meanwhile only 3% would be required using microalgae [20]. Algae oil can be collected with low energy requirements by CO2 media acidification, low-power pulsed electromagnetic field and static mixer turbulences to break the membrane and posterior decantation to separate oil, water and cellular membranes [21]. However, palm oil is used as base case in Aspen Plus® documentation providing the composition of palm oil. The catalyst considered in the Aspen Plus® example is homogeneous. However, important advances are reported in literature concerning application of heterogeneous catalysts, e.g. MgeZn mixed metal oxide catalysts [22], acidic ionic liquid immobilized on poly divinylbenzene [23], hydrotalcite as basic catalyst [24] or sugar catalyst [8]. Furthermore, it is evaluated the suitability of the hybrid reactive extraction column for several pure triglycerides, fat and oils. The number of required reactive and non-reactive stages required for a hybrid extraction column able to treat several kinds of oils and fats is also determined. The UE and USA legislation related to biodiesel composition is taken into account (Table 1). The effect of presence of free fatty acids, as well as the potential for glycerol revalorization is not the scope of the present study. This allows providing more details about the biodiesel synthesis. The study focuses on aspects such as comparison between classical and intensified process for biodiesel synthesis from minimum energy point of view. When alternative processes are available, the process with lower energy requirements, typically, is more efficient and preferable because it provides lower operation costs, higher energy conservation and lower environmental impact. However, the energy requirements comparison between processes is not an easy task because it depends on how well it is optimized and how process integration is made. The minimum energy consumption

FFA

C 7:0 C 8:0 C 14:0 C 16:0 C 16:1 C 16:2 C 18:0 C 18:1 C 18:2 C18:3 Others

Jatropha curcas oil

14.2

7.0 44.7 32.8 1.3

Algae oil Tolypoghrix

Spirogyra

Spirulina

5.8 31.8 4.7 2.4 2.7 23.4 8.6 8.4 12.2

6.4 25.2 5.4 3.8 4.5 33.3 10.8 0.7 9.9

0.23 46.07 1.26 3.38 1.41 5.23 17.43 8.87 16.12

calculated with the aid of Pinch Analysis is a great stimulus to achieve efficient process conceptual designs. Although the methodology of taking into account the distillation column energy requirements in the Pinch Analysis was well established [25], the calculation of the minimum energy requirements of a distillation column, in a simple way, based on thermodynamic principles, was not described until recently [26]. The minimum energy requirement of a system is calculated taking into account the maximum heat exchangeable, fulfilling the second thermodynamic law and considering the distillation columns as heat engines that provide separation instead of work. At our knowledge, this approach to calculate the minimum energy requirements of a process has not been yet reported before in literature. This paper proposes to verify if classical process integration could compete with the proposed intensified process from energy point of view. In some cases, the energy savings for process intensification and the process integration are reported to be similar, e.g. the energy requirements for methanol and glycerol recovery can be decreased with 27% by using a divided wall column [27] or 23% with an appropriate heat exchange network [28]. Therefore, a comparison between the hybrid reactive extraction column and the classical scheme is required to verify the potential of the novel intensified process. The novel process is evaluated using several triglycerides and oils, i.e. to be close to real situations from this point of view. 2. Material and method Biodiesel synthesis from palm oil using a classical process scheme available as example in AspenPlus® version 8.2 is used as base case, providing palm oil composition (Table 2). Thermodynamic data and kinetics are ready implemented in AspenPlus® v8.2

Table 2 Palm oil feed composition. Compound

AspenPlus® identifier

x (Mass)

Trioleine Trimirystine Tripalmitine Dipalmitine stearine Dipalmitine oleine Palmitine oleine Stearine Dimiristate palmitine Dipalmitine Linoleine Palmitine Dioleine Palmitine Linoleine Oleine Dioleine Stearine Dioleine Linoleine Mirystine Palmitine Linoleine 1,3-Dipalmitine

OOO MMM PPP PPS PPO POS MMP PPLI POO PLIO OOS OOLI MPLI PP

0.0440 0.0042 0.0551 0.0106 0.2962 0.0490 0.0170 0.0923 0.2326 0.0968 0.0224 0.0058 0.0220 0.0520

Fig. 1. Methanol and triglycerides reacting in co-current in a PFR.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

V. Ples¸u et al. / Energy xxx (2014) 1e15

example [29], the thermodynamic model used is UNIFACDORTMUND. Oils used as raw material provide mixtures of 33 compounds. The triglycerides of following fatty acids are taken into account: myristic (C14:0 M), palmitic (C16:0 P), stearic (C18:0 S), oleic (18:1 O) and linoleic (18:2 Li). The nomenclature used in this paper is the same as the one used in AspenPlus® where “O” means Oleic acid, “M” e Myristic acid, “P” e Palmitic acid, “Li” e Linoleic acid and “S” e Stearic acid. The components denominated with three letters are triglycerides and each letter represents the base fatty acid (for example OOO is a triglyceride obtained from three oleic acid molecules, i.e. triolein). In a similar manner, the molecules denominated with two letters are diglycerides made of two fatty acids according to the initials specified previously. The molecules like 1O are monoglycerides of the fatty acid indicated. Finally, components like Methyl-O represent fatty acids methyl esters (FAMEs) that, in this case, derive from oleic acid. The extraction column operates at 60  C. Jatropha curcas oil [30] and several algae oils compositions [31] are listed (Table 3). The composition of the different algae oils is quite similar. As a reactive extraction column is not implemented as an independent unit operation in AspenPlus®, each equilibrium reactive stage is simulated combining a CSTR (continuous stirred tank reactor) and a decanter. The thermodynamic model UNIFAC-DORTMUND is chosen, as it is proved to be suitable for biodiesel-related LLE (liquideliquid equilibrium) () mixtures [32]. The reactive stages are added one by one until all the input oil is converted to biodiesel. The novel process is compared to the classical scheme in terms of minimum energy requirement, which is calculated using the Pinch Analysis, as implemented in Sprint® software [33]. This software tool is provided by Centre for Process Integration, The University of Manchester, as CTTIP is member of the Process Integration Research Consortium. Sprint® software is a complete tool for Pinch analysis, compared with other similar tools, available as commercial or free license. Distillation columns are considered as heat engines [26]. 3. Results Pure triolein was previously considered as raw material [7] for the proposed novel process. Process simulations for several pure triglycerides and mixtures of these triglycerides according to typical compositions of several oils are provided in this paper. First,

3

Fig. 3. The increase of number of chain length decrease the biodiesel purity.

the influence of pure triglycerides on hybrid extraction column stages requirements is evaluated. Secondly, the classical process scheme consisting of reactor section and separation section, already implemented in Aspen Plus® for palm oil, is compared to proposed process model. Finally, the behaviour of several types of oils is evaluated, to propose column configurations and operation conditions able to treat any type of oil. 3.1. Pure triglycerides transesterification to produce biodiesel A stoichiometric mixture of methanol and pure triglycerides of 4 kmol/h with 0.25 kmol/h NaOH is feed to a reactor modelled as a PFR (plug flow reactor), operated isothermally at 60  C. Although the conversion is somewhat high, no total conversion is achieved when methanol and triglycerides flows in co-current (Fig. 1). The results show that some compounds react faster than others do, e.g. trimyristin. In general, it is observed that the triglycerides reaction rate is faster for low number of double bounds and short chain length. Some other triglycerides, such as the tripalmitin, react slower, but reach higher conversions compared to trimyristin. The lowest conversions are provided for triglycerides containing long chain saturated fatty acids, e.g. O (oleine). However, some exceptions do not fulfil this general tendency. For instance, OOLi (diolein-

Fig. 2. The increase of number of double bounds decrease the biodiesel purity.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

4

V. Ples¸u et al. / Energy xxx (2014) 1e15

Fig. 4. Conversion along the column stages depending on the chain length.

linolein) has more double bounds than OOO (triolein), but conversion slightly higher, according to the kinetic model implemented in Aspen Plus®. The reaction mixture splits in two liquid phases, and therefore a counter-current flow of reactants is feasible. It is commonly accepted that countercurrent flow is usually more efficient than cocurrent flow of reactants. The high amount of the reactant at the entrance displaces the other reactant to high conversions at the exit. Nevertheless, the countercurrent flow pattern means that it cannot be treated any more as a PFR but as equilibrium stages. Its model is closer to the reactive distillation column than to a PFR and liquid phase equilibrium becomes as important as the reaction kinetics. The feed is stoichiometric and the 1 kmol/h of methanol is diluted with a glycerol stream of 1 kmol/h because glycerol is required as extractive agent in the non-reactive section of the proposed process. The influence of chain length and number of double bonds is determined when the reactants flow in counter-current in an

extractive-reactive column of five equilibrium stages. The mass fraction of methyl esters decreases when the number of double bounds in the triglyceride increases (Fig. 2). The decrease of this mass fraction is much influenced when the number of double bounds is greater than three. When there are more than 16 carbon atoms in the chain, the ester mass fraction decreases more abruptly (Fig. 3). With 5 reactive stages, a practical total conversion is obtained for any triglyceride and the ester mass fraction is above 90%. The main impurity in the obtained biodiesel is due to the solubility of methanol in biodiesel. However, the presence of methanol in biodiesel is not a problem as it is extracted in the non-reactive stages by glycerol. In the case of POO (palmitin-diolein) and PPP (tripalmitin), with just one stage, conversions higher than 90% are already attained. The triglycerides conversion and mass fraction variation along the number of stages, depending on carbon chain length, are presented in Figs. 4 and 5, respectively. According to these results, five reactive stages provide enough biodiesel purity, although a

Fig. 5. Methyl esters mass fraction along the column stages depending on the chain length.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

V. Ples¸u et al. / Energy xxx (2014) 1e15

5

Fig. 8. Influence of the extractive agent flow rate on methanol content in final biodiesel.

Fig. 6. Non-reactive section stages required to decrease the amount of methanol below the specification.

non-reactive extraction section is required to remove methanol from biodiesel. According to the EN 14214, the methanol content in biodiesel must be under 0.2% (Table 1). Fig. 6 shows that with 9 non-reactive extraction stages the methanol content in biodiesel decreases to values under the legal limit allowed, independently of the triglyceride used. However, there is not a big difference related to the number of stages required for each triglyceride, e.g.trimyristin requires the lowest number of stages e 8 stages. Hence, with a hybrid reactive extraction column with 5 reactive stages and 9 non-reactive stages, a total conversion of methanol and any triglyceride stoichiometric mixture to glycerol and biodiesel is attained. An excess of methanol does not influence much the conversion. This fact is observed for MMM (trimystirine) in Fig. 7. However, when methanol quantity increases slightly above

Fig. 7. Influence of methanol excess on biodiesel purity.

the stoichiometric mixture (3 kmol/h methanol: 1 kmol/h MMM (trimyristin)), the conversion presents a sudden small increase that is not observed any more at higher ratios of methanol. This fact is not a problem and a stoichiometric feed mixture is recommended to be used industrially. However, the process simulation considering a high number of compounds and reactions requires a small excess of methanol to assure the mathematical convergence. The convergence of PPS (dipalmitin-stearin) has not been attained with stoichiometric methanol. When the ratio is 3.1 kmol/kmol PPS, mass fraction of methyl esters 99.68% is obtained. Another important parameter is the glycerol flow rate for methanol extraction from the final biodiesel in the extraction section. The minimum mass flow rate percent of glycerol/triglyceride for methanol extraction depends on the kind of triglyceride used (Fig. 8). MMM (trimystirine) requires the highest amount: a flow rate of 7% of glycerol with respect to biodiesel. The flow rate of extraction glycerol also influences the amount of monoglycerides, diglycerides and triglycerides present in biodiesel. There is a glycerol flow rate that minimizes the amount of monoglycerides in biodiesel but there is a wide interval where the legislation is fulfilled (Fig. 9). Therefore, the mass flow rate of glycerol should be between 9 and 13% related to the biodiesel flow rate. In the process simulations, the amount of glycerol is fixed to 0.8 kmol glycerol/ kmol feed oil and therefore the mass percent of glycerol/biodiesel is not constant but within the indicated range. Finally, the mass fraction of methyl esthers is calculated. Most of the triglycerides reach a mass fraction of methyl esthers higher than required by the legislation (Fig. 10). Nevertheless, the palmito-dioleo-glycerol

Fig. 9. Influence of the extractive agent flow rate on monoglycerides content in final biodiesel.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

6

V. Ples¸u et al. / Energy xxx (2014) 1e15

Fig. 12. Influence of number of double bonds on the energy requirements.

Fig. 10. Influence of the extractive agent flow rate on methyl ester fraction in final biodiesel.

(POO) has not reached enough conversion, although it can be solved adding reactive stages. The simulations are performed following the minimum equilibrium stages required but, as the rule of thumb for the distillation columns, at least 40% more equilibrium stages would be advisable for an industrial implementation. An excess of equilibrium stages increases the investment cost, while a deficit of equilibrium stages, once the column is constructed, produces biodiesel out of specifications that could not be improved only changing the operation variables. The above results show that a reasonable number of equilibrium stages allows the production of biodiesel, fulfilling the legislation. The kinetic model implemented in AspenPlus® used is based on the homogeneous NaOH basic catalyst. Although there are some other models available in the literature, e.g. heterogeneous, the same catalyst has been used for an easiest comparison between the classical process and the intensified process. Fig. 11 shows that the minimum amount of catalyst required for OOO (triolein) is 0.07 kmol/h for feed flow rate 4 kmol/h. In the process simulations, a flowrate of 0.25 kmol/h has been used, following a ratio similar to the one provided in the classical scheme of AspenPlus. Assuming the feed at 60  C, energy is required to ensure reaction and mixing enthalpies. It is obvious that it is very low. Energy requirement increases for compounds with large number double bounds (>4) (Fig. 12) and passes through a minimum around 16 carbon chain length (Fig. 13).

Fig. 11. Influence of the catalyst amount.

Results previously presented show that the counter-current flow is able to attain a total reactant conversion to pure biodiesel and glycerol when an enough number of equilibrium stages are used, without an excess of methanol. Therefore, the next section compares the classical scheme for biodiesel production with the novel proposed scheme. 3.2. Palm oil transesterification, comparison between the classical and intensified process The classical process scheme for biodiesel synthesis is composed of a reactor followed by a separation section with several distillation columns (Fig. 14). The flowrate of palm oil entering the process is of 1.28 kmol/h. The classical scheme has a first distillation column to recover part of the non-reacted methanol in high excess. A LLExt (liquideliquid extraction) column with water separates glycerol from non-polar phase, the non-polar phase in a second distillation column is split in non-reacted oil to be recycled and pure biodiesel. In a third column, glycerol is purified. The distillation columns operate at low pressure, e.g. 0.2 bar. There are also several heat exchangers, e.g. HX1 for reactor feed, HX2 for extraction column feed and B2 for the oil recycled. A single hybrid reactive extraction column can replace the classical scheme (Fig. 15a). As the reactive extraction column is not implemented in AspenPlus®, the flowsheet shown in Fig. 15b is considered to be equivalent. A CSRT (continuous stirred tank) model is considered for reaction section and a decanter simulates the LiquideLiquid Equilibrium for each reactive equilibrium stage. The stream GLI (glycerol) þ MeOH (methanol) is the internal column stream of the polar phase from the extractive to the reactive region; the main components are glycerol that dragged methanol from the non-polar phase. The temperature of the inlet stream GLYCERIN (extractive agent) is 60  C because is a recirculation from the glycerol produced (S2).

Fig. 13. Influence of chain length on the energy requirements.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

V. Ples¸u et al. / Energy xxx (2014) 1e15

7

Fig. 14. Example of biodiesel synthesis provided by AspenPlus® 8.2.

The reactants, oil and methanol, flow in counter-current in the reactive extraction column section. Methanol is the lowest density compound, glycerol is the highest density compound, and therefore the polar phase methanol-glycerol flows downwards, as the reaction proceeds and becomes richer in glycerol. Methanol is fed at the top of the reactive section and glycerol is collected at the bottom of the reactive section. On the other hand, the vegetable oil is fed at the bottom of the reactive section and impurified FAME is collected at the top of the non-reactive section. Methanol is partially soluble in FAME and therefore a non-reactive extraction section at the upper part of the column is required to extract it from the obtained biodiesel, using glycerol as extraction agent. This is obtained by recycling part of the glycerol product stream collected in the bottom of the column. The numbering of column stages is performed from bottom to top. Fig. 16 shows the evolution of the composition versus the number of reactive stages in a reactive extraction section, keeping constant the number of non-reactive stages. For simplicity purposes, the compounds are grouped in biodiesel, triglycerides, diglycerides and monoglycerides without representing in the figure the type of fatty acid present. Biodiesel is the main component. Although an almost total conversion is achieved using only few stages, the biodiesel stream purity is limited due to the methanol feed in counter-current. Nevertheless, the main aim of reactive stages is to convert the vegetable oil to biodiesel (99% in this system). A conversion of 99.2% of palm oil to biodiesel can be achieved in 5 reactive stages. Notice that a column with 5 reactive stages provides biodiesel containing 16% methanol. This quantity can be decreased to the allowed quantity of 0.2% in mass of methanol, according to the European standards EN14214, by increasing the number of extraction stages or flow rate of extractive agent. The relation between minimum glycerol flowrate and the number of non-reactive stages allowing to obtain limiting MeOH concentration is shown in Fig. 17, obtained in AspenPlus®. An acceptable result is obtained for 9 non-reactive stages (99.6% of biodiesel purity as established in EU) for glycerol flow rate of 0.89 kmol/h for an imposed palm oil flow rate of 1.28 kmol/h. The number of reactive and non-reactive stages is in agreement with the previously calculated values for pure triglycerides. A convenient

extractive agent flow rate could be 0.077 kg glycerol/kg palm oil which is in the range defined previously for pure triglycerides with 14 theoretical stages (5 reactive þ 9 non-reactive). The 5 reactive stages at the lower part of the column are used to achieve the complete palm oil conversion, and the 9 non-reactive stages at the upper part is used to separate the biodiesel from methanol by extraction with glycerol. Fig. 18 shows the composition profile for the polar phase. The main components of this phase are methanol and glycerol, the latter flowing from the top to the bottom, also having the higher density. MeOH increases its composition from top to bottom of non-reactive section (from stage 14 to stage 5). In agreement with reaction scheme, MeOH reacts with glycerides forming FAME and glycerol in the reactive section (from stage 5 to stage 1), as illustrated in Fig. 18. Fig. 19 shows the composition profile for the non-polar phase. In the first stage, most of the triglycerides are converted to monoglycerides, which react to biodiesel in the second stage. Although in the first stage the composition of biodiesel is lower, in the second stage the mass fraction is higher than 0.9 and remains constant through the reactive phase. Between stages 3e6, methanol concentration passes through a maximum, enabling high conversion. Between stages 6e14, which are non-reactive, methanol concentration decreases until the desired value, under 0.2%. In the extractive phase, the biodiesel composition rises to achieve the required value (99.6%). As mentioned previously, for simplicity, the compounds used in the simulation are presented in Figs. 16e19 lumped as triglycerides, diglycerides, monoglycerides, biodiesel, glycerol and methanol. However, in Table 4 detailed composition for the input and output streams of the reactive extraction column are presented. In Table 4, raw materials (oil and MeOH) are fed at 25  C. Reactive extraction column operates at 60  C. Appropriate utilities are used to provide this condition. The purity of biodiesel collected is 99.6% and the purity of glycerol is 88.2%. The purity of glycerol is limited by the presence of a slight excess of 6.7% methanol in the feed, otherwise mathematical convergence is not achieved due to the great quantity of compounds present in palm oil. Running the economic analysis in AspenPlus® with the default settings for the classical scheme, for a biodiesel flow rate of 1051 kg/

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

8

V. Ples¸u et al. / Energy xxx (2014) 1e15

Fig. 15. Hybrid reactive-extraction column: (a) proposed process scheme and (b) simulation flowsheet.

h, the capital costs are of about 9,111,180 USD. For the intensified process, the capital costs calculated in AspenPlus® are of 1,951,360 USD Therefore, the capital cost for the classical process is 4.6 times higher than for the intensified process. Nevertheless, these values suggest that the intensified process is more advantageous from economic point of view. 3.3. Comparison of the minimum energy requirements for the intensified process with the classical scheme The classical scheme proposed in AspenPlus® example requires 3 distillation columns, two for the product purification (ESTCOL to

purify biodiesel and to recirculate the non-reacted oil and GLYCOL to purify the glycerol) and one for methanol recovery (MEOHCOL) which is recirculated to the reaction stage. In addition, an extraction column (WASHCOL) to separate the polar and non-polar phases using water is considered and a second reactor is needed to neutralize with H3PO4 the NaOH catalyst present in the polar phase using. The product Na3PO4 is commonly used as fertilizer. In Table 5, distillation columns specifications are given. Distillation columns operate at high temperature and low pressure, involving important energy consumption, due to low volatility of glycerol and FAME, as revealed in Table 6. However, the extraction column WASHCOL operates adiabatically at 1 bar.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

V. Ples¸u et al. / Energy xxx (2014) 1e15

9

Fig. 18. Composition profile for the polar phase. Fig. 16. Composition results of the reactive phase simulation.

The use of heat exchanger HX2 is irrelevant because only increases the temperature with 1  C and its energy consumption is insignificant. One unclear aspect of this system is the recirculation of non-reacted oil that leaves the distillation column ESTCOL at 338  C, later is cooled to 25  C in the heat exchanger B2 and finally is heated again to the reaction temperature of 60  C. Therefore, when comparing different processes, it is important to assess rational use of energy. In this respect, Pinch Analysis is very useful methodology. The streams are heated to 60  C before entering the reactor and then the effluents are feed to distillation columns at the corresponding boiling point. Process scheme output streams are cooled to 35  C. In this respect Pinch Analysis is used. In Table 7 and Fig. 20, data extraction results are presented, for hot and cold process streams. Fig. 21 shows the Composite Curves for the classic process considering only the streams, without taking into account the distillation columns. A DTmin ¼ 40  C is necessary to obtain a process integration problem. The heat utility target is small, 8.35 kW due to the output distillation column streams at high temperature. Cold utility is as well relatively small, 47.1 kW. But, energy consumption of distillation columns is very important, as presented in Table 6. The energy of process streams considered above is just a small part of flowsheet heat requirements, therefore the columns must be also considered in process integration study. Following recent results [8], it is possible to calculate the minimum energy consumption of a distillation column, for consideration into the Pinch Analysis. Fulfilling the second thermodynamic law, the distillation column acts as a heat engine, which exchanges energy between a cold and a hot focus (condenser and reboiler) to ensure

Fig. 17. Minimum glycerol flowrate to obtain MeOH legal composition in biodiesel versus number of non-reactive stages.

stream composition modification (separation task). The energy required for this purpose is calculated with equations given below (enthalpy of mixing is assumed negligible): 0

Qr ¼

Tc $DSsep þ l$D h

(1)

P  X  3 ðD þ BÞ$  D$ xdistil  $ln xfeed $ln xdistil xfeed i i i i 5 P  DSsep ¼ R$4 B$ $ln xresidue xresidue i i 2

(2) h¼

Tr  Tc Tr

(3)

The minimum energy requirements are shown in Table 8. These values are introduced in the heat cascade of Pinch Analysis. The reboiler duty has the same sign as a cold stream because it consumes energy. The same value is used for the condenser duty, but with the opposite sign. The column duties are introduced in the cascade at the reboiler and condenser temperatures, respectively. At our knowledge, it is the first time that the minimum energy requirements of a process are determined based on the second law of thermodynamics, taking into account the operation units requirements. The minimum energy consumption is lower than a simple sum of the units' energy requirements shown in Table 6, as the distillation columns can be energy integrated. The energy

Fig. 19. Composition profile for the non-polar phase.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

10

V. Ples¸u et al. / Energy xxx (2014) 1e15

Table 4 Detailed composition for the input and output streams. Stream

OIL

x (mass) METHANOL OOO MMM METHYL-O GLYCEROL NaOH PPP METHYL-P PPS PPO POS MMP PPLI POO PLIO OOS OOLI MPLI METHYL-M METHYL-S METHY-LI 1-M 1-P 1-S 1-O 1-LI MM PP OO PO PLI MP PS OS LIO MLI Total Flow kg/h SUMMARY Triglycerides Diglycerides Monoglycerides FAME Methanol Glycerol NaOH

MEOH

FAME

GLYPR

GLY þ MEOH

82

0.00052 0.00000 0.00000 0.38932 0.00019 0.00000 0.00001 0.48103 0.00000 0.00001 0.00000 0.00000 0.00000 0.00001 0.00000 0.00000 0.00000 0.00000 0.01993 0.02934 0.07655 0.00005 0.00123 0.00009 0.00123 0.00025 0.00000 0.00005 0.00004 0.00009 0.00002 0.00000 0.00001 0.00001 0.00001 0.00000 1062

0.06613 0.00000 0.00000 0.00002 0.88242 0.04749 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00001 0.00000 0.00001 0.00009 0.00355 0.00001 0.00025 0.00003 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 225

0.50941 0.00000 0.00000 0.00033 0.48095 0.00786 0.00000 0.00075 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00007 0.00002 0.00009 0.00003 0.00030 0.00001 0.00016 0.00002 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 171

1.0000

0.00004 0.00022 0.00285 0.99617 0.00052 0.00019

0.00000 0.00000 0.00393 0.00004 0.06613 0.88242 0.04748

0.00000 0.00000 0.00052 0.00126 0.50941 0.48095 0.00786

GLY

0.92707 0.04699 0.00366 1.0000 0.07293 0.05366 0.01068 0.29774 0.05091 0.01540 0.09256 0.24112 0.10011 0.02398 0.00618 0.02132

0.03568

1059 x (mass) 0.96432 0.03568

146

0.92707 0.07293

requirement to operate at low pressure is not considered. Applying the Pinch Analysis and taking into account the minimum energy requirement for the distillation columns, the hot minimum energy requirement of the system is of 28.0 kW. The minimum duties (condenser and reboiler duties) calculated for the columns in Table 8 are considering a very small temperature difference between the reboiler and condenser and the respective heating and cooling fluids according to minimize the irreversibility. For better real values, the temperature of the units should be shifted as it was made for the process streams, to take into account the irreversibilities in condenser and in reboiler. The temperature shift is performed automatically when the data is introduced in Sprint®. A very small corresponding difference of temperature (0.1  C) is used to consider the cold stream in Table 5 Distillation columns specifications. Name

Stages

Reflux ratio

Pressure (bar)

Treboiler ( C)

Tcondenser ( C)

MEOHCOL GLYCOL ESTCOL

7 6 6

2 2 0.95

0.2 0.4 0.1

59.2 262.2 338

28.8 70.9 148.3

reboilers and hot streams in condensers. The very small increase of temperature must be positive for the reboilers, which are considered as a cold stream, and negative for the condensers. As the reaction is exothermic, the reactor is considered as a hot stream. Stream data introduced in Sprint® is shown in Fig. 22. The Composite Curves taking into account the irreversibilities in Table 6 Energy consumption of process equipment. Equipment

Type

Heat (kW)

Heat requirement

MEOHCOL GLYCOL ESTCOL HX1 REACTOR

Heat to eliminate

MEOHCOL GLYCOL ESTCOL HX2 B2 NEUTR

Reboiler Reboiler Reboiler Heat exchanger Reactor TOTAL Condenser Condenser Condenser Heat Exchanger Heat Exchanger Reactor TOTAL

118.0 283.8 374.4 22.0 15.3 813.5 122.0 261.2 304.7 0.1 9.8 1.5 ¡699.3

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

V. Ples¸u et al. / Energy xxx (2014) 1e15

11

Table 7 Data extraction: hot and cold process streams for classical process scheme.

1 2 3 4 5 6 7 8 9 10 11

Stream

TS ( C)

TT ( C)

Type

CP (W/ C)

DH

MIX 1 MEOH recirculated OIL REACT-MEOHCOL AQU3 EST4-ESTCOL WATMEOH MEOHWAT FAME GLYCEROL OILREC

25 28.8 25 60 50 54 71 148.3 148.3 262.2 338.01

60 60 60 40 100 136 35 35 35 35 60

COLD COLD COLD HOT COLD COLD HOT HOT HOT HOT HOT

153 98 622 883 221 645 152 2 694 93 35

5357 3043 21,764 17,653 11,052 52,857 5465 272 78,650 21,108 9697

(W)

reboiler and condenser of distillation columns and the reactor are presented in Fig. 23. Then, as usual for such processes DTmin ¼ 10  C, energy required by the units, i.e. distillation and reaction, are the horizontal regions of the curves. The pinch point temperature is 66.1  C, but as it is shown in Fig. 23. From Sprint® energy report hot utility target is 33 kW, cold utility target is 79.3 kW, and process heat recovery is 121 kW. These values are very low compared to the requirements of AspenPlus® base example without heat integration, i.e. 813.15 kW. These results underline the importance of process integration, taking into account all process units. Grand Composite Curve (Fig. 25a) indicates the potential heat recovery inside the process (a big pocket in the hot zone, which lowers the demand and temperature of hot utilities. Therefore, it is possible for example to use about 25 kW hot utility with temperatures less than 150  C. Just a small amount of hot utility ~10 kW is needed around 300  C and a very small amount of hot utility ~3 kW is needed for more than 350  C. For our proposed process, the extractive stages operate at 60  C, therefore energy should be considered to maintain constant temperature (Table 9). This can be achieved using a jacketed column. Table 9 shows also that the energy is consumed mostly by the first and fifth stages of the reactive zone, because oil stream and methanol stream are fed at 25  C. Therefore, the non-integrated alternative process has a relatively low energy requirement 31.8 kW, compared to the basic AspenPlus® example process of 813.5 kW. It is also bit lower when compared to the heat integrated AspenPlus® basic process (33 kW). The composite curves of the isothermal extraction column with heat integration of the streams (Table 10) are shown in Fig. 24. For same reasons as above, DTmin ¼ 10  C can be considered. As shown in Fig. 24, in this case Composite Curves give minimum hot utility requirement is 10.5 kW, minimum cold utility is 13.6 kW, and process heat recovery is 26.2 kW. The pinch point temperature for the intensified processes is 55  C. Therefore our

Fig. 21. Composite Curves for the classic process without considering the units, where DTmin ¼ 40  C.

Table 8 Minimum energy requirements of the distillation columns.

MEOHCOL GLY COL ESTCOL REACTOR

Tc ( C)

Tr ( C)

h

DS

(W/ C)

Dl (kW)

Minimum energy consumption (kW)

28.8 71 148.3 60

60 262.2 338.01

0.091 0.357 0.310

15.3 7.8 1.9

38.5 2.6 0.1

46.4 10.7 2.5 7.4

proposed flowsheet is a better option, considering process intensification and rational use of energy (Table 11). Furthermore, the intensified process requires a heat source of lower temperature than the AspenPlus® base process (Fig. 25) as is shown by Grand Composite Curve. Minimum hot utility temperature is 70  C and maximum cold utility temperature is 25  C. 3.4. Evaluation of using the hybrid extraction column for several oils The presence of FFA (free fatty acids) is not the scope of the present study, although it is also an important point to be considered as for instance J. curcas oil can contain up to 15% FFA [34]. Same composition has as well as cooking oils. Excess FFAs lead to increased catalyst requirement as well as soap formation. FFA should be esterified with an acid catalyst before being fed to the column. The oil compositions are approximate to triglyceride composition (Table 12) with an equivalent amount of fatty acids indicated in Table 3. The flow rate of oil fed to the process is of 1 kmol/h, requiring a stoichiometric amount of methanol and 0.8 kmol/h glycerol.

Fig. 20. Classical process scheme with streams according to Table 7.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

12

V. Ples¸u et al. / Energy xxx (2014) 1e15

Fig. 22. Data input at Sprint®. From 1 to 11 are streams, from 12 to 17 are column reboilers (R) and condensers (C) and 18 is the reactor.

Biodiesel composition obtained from J. curcas oil is presented in Table 13. The biodiesel obtained using stoichiometric methanol ratio fulfils EN 14214-legislation provision with exception of diglycerids and glycerol content. The excessive diglycerides quantity is unexpected result. The diglycerides with a high mass faction are PO and PLi. These are not directly obtained by transesterification of the initial triglycerides (PPS and OOLi), but by recombination. This fact can be justified by the stability of the compounds, e.g. PO is the more stable compound, according to the enthalpy of formation. Glycerol removal is easy by water extraction. J. curcas oil has been assumed as a mixture of 77% OOLi and 23% PPS and the percent of methyl esters is 97.77% instead of the 99.47% obtained using OOLi alone. Therefore, the presence of PPS leads to a lower conversion and a higher amount of diglycerides. A slight increase of the methanol amount, higher than the stoichiometric one, produces an increase of methyl esters quantity, in the same time, the mass fraction of diglycerides decreases. However, a higher number of equilibrium stages is preferred rather than the use of a methanol excess to fulfil the biodiesel specifications.

Spirulina oil composition is assimilated to be 90% PPLi (dipalmitin-linolein) and 10% OOLi. The linolenic acid content has been assimilated to linoleic in the process simulation because it is not in AspenPlus® database. An interesting result is that the main triglyceride, which appears in biodiesel, is PPO (dipalmitoyl-oleiylglycerol), although this triglyceride is not initially present in algae oil. This behaviour can be explained considering a transesterification process. In this way, the main triglycerides present in biodiesel are always PPO or PLiO, independently if they are present or not, in the initial oil. Contrary to the result obtained for J. curcas oil, biodiesel obtained from spirulina oil do not contain significant amounts of diglycerides PO and PLi. The reason is that as the OOLi is only a 10% of oil composition, a higher conversion is obtained with fewer stages. Spirogyra oil is assimilated to a composition of 37.86% PLiO, 26.28% OOO, 15.75% PPS, 12.63% PPP and 7.48% MMM. An excess of methanol to 3.5 kmol/h has been required to achieve process

Fig. 23. Composite curves of the classical process.

Fig. 24. Composite curve of the intensified process.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

V. Ples¸u et al. / Energy xxx (2014) 1e15

13

Table 10 Data extraction: hot and cold process streams for the intensified process.

1 2 3 4

Stream

T1 ( C)

T2 ( C)

Type

CP (W/ C)

DH (W)

OIL MEOH FAME GLYPR

25 25 60 60

60 60 35 35

COLD COLD HOT COLD

584 465 1444 147

20,438 16,267 36,106 3667

Table 11 Energy requirements for classic and intensified process alternatives (DTmin ¼ 10  C).

Classic Intensified

Energy consumption (kW)

Minimum thermodynamic energy (kW)

Maximum temperature (oC)

813.5 31.8

33 10.5

338 60

4. Conclusions A hybrid reactive extraction column to produce biodiesel (five reactive stages and 9 non-reactive stages) is rigorously simulated for several oils considering pure triglycerides or combinations of heterogeneous triglycerides. The minimum number of equilibrium stages and operating parameters are studied for pure triglycerides. A higher number of double bonds decreases the conversion and requires more process energy. Unsaturated fats Fig. 25. Grand composite curves for: a) the classic process scheme; b) intensified process.

simulation convergence. In a similar way as J. curcas, a too high amount of diglycerids and glycerol is present in the collected biodiesel. For 3.3 kmol MeOH/h, 1.50% diglycerids are collected; for 3.5 kmol MeOH/h, 1.47% diglycerids and for 4.0 kmol MeOH/h, 1.52% diglycerids are collected, therefore a minimum of diglycerids exists between 3.3 and 4 kmol MeOH/h. The Tolypothis algae oil is assimilated to 7.88% MMM, 34.40% PPP, 18.74% PPLi and 38.98% OOLi and it is simulated with a stoichiometric reactants mixture. As discussed previously, more stages should be used to assure total conversion and stoichiometric feed than the calculated one for pure triglycerides. There are 14 (5 þ 9) necessary equilibrium stages when considering pure triglycerides, but when using oils, the simulation results suggest considering a larger number of equilibrium stages to assure reasonable conversion. The energy requirements are very low for all kind of oils. Most of the consumption is due to heating the reactant streams and to the first stages of the reactive section where most of the conversion takes place (Fig. 26). The lowest energy requirement is for Spirogyra algae oil.

Table 9 Energy requirement of the alternative process. Extractive-reactive section

Extractive section TOTAL system

Heat duty (kW) Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Total All stages TOTAL

20.4 7.1 2.6 0.4 4.4 34.9 3.1 31.8

Table 12 Oils and corresponding biodiesel compositions. Stream: Mass flow [kg/h] Jatropha curcas Spirogyra Spirulina

METHANOL MMM METHYL-O GLYCEROL NAOH METHYL-P OOO PPP PPS PPO POO POS MMP PPLI POO PLIO OOS OOLI METHYL-M METHYL-S METHY-LI 1-M 1-P 1-S 1-O 1-LI MM PP OO PO PLI MP PS OS LIO

Tolypothryx

FAME

OIL

FAME OIL FAME OIL

0.4270 0 449 0.2510 0.0007 116 0 0 0.0226 0 0 0 0 0 0 0.3780 0 0.0307 0 68 225 0 1.80 0.2650 3.12 0.8720 0 1.43 0.3980 8.40 2.03 0 0.0154 0 0.1060

0 0 0 0 0 0 0 0 192 0 0 0 0 0 0 0 0 680 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

2.12 0 0.3130 0 0.2630 0 0 63 0 0 0 56.90 157 0 43 0 188 0 0.4480 0 0.2400 0 0.2350 0 0.0006 0 0.0007 0 0.0007 0 290 0 466 0 348 0 0 0 0.7170 0 25.40 0 0 0 0.0042 0 0.0032 278 0.0048 133 0 0 0 0 0.0222 328 34 0 47.90 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0090 753 0.0043 156 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3.98 319 0.0001 83.60 0.0011 344 63 0 0 0 57.10 0.0000 47 0 0 0 0 0 59 0 294 0 169 0 0.1610 0 0 0 0.1620 0 9.94 0 0.9230 0 1.22 0 0.1700 0 0 0 0 0 6.16 0 0.1440 0 0.6460 0 0.1810 0 1.00 0 0.5970 0 0.0010 0 0 0 0.0011 0 0.3180 0 0.0390 0 0.0393 0 125 0 0.0005 0 0.0097 0 0.2370 0 0.0089 0 0.0389 0 0 0 0.0614 0 0.0355 0 0 0 0 0 0 0 0.0372 0 0 0 0 0 0.0383 0 0 0 0 0 96 0 0.0067 0 0.0177 0

FAME OIL

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

14

V. Ples¸u et al. / Energy xxx (2014) 1e15

Table 13 Mass percent composition of biodiesel from several oils (3 kmol/h methanol is the stoichiometric amount). Oil

Jatropha curcas oil

Spirulina algae oil

Compound

3

3.3

3

Methanol Methylic esters Monoglycerids Diglycerids Triglycerids Glycerol

0.049 97.775

0.048 98.335

0.037 95.505

0.689 1.409 0.049 0.029

0.525 1.030 0.034 0.028

0.246 0.014 4.169 0.029

Spyrogira algae oil

Tolypothrix algae oil

3.2

3.5

3.0

0.044 99.607

0.045 98.005

0.03 90.88

0.299 0.019 0.003 0.028

0.424 1.470 0.027 0.029

0.31 0.02 8.73 0.03

Fig. 26. Energy requirements for the column operated isothermally at 60 1 kmol/h of oil.

C

for

are most benefic from health point of view but, for biodiesel synthesis, they decrease the conversion and determine more energy consumption. Therefore, it would be recommended to separate the unsaturated fats for human consumption and use the saturated fats for biodiesel synthesis. The oils composed of several triglycerides require larger reactive extraction column. In some cases, a too short column can be compensated using an excess of methanol. Therefore, a column with more stages would be recommended for industrial application. However, it is expected to attain a total conversion of stoichiometric reactants with an acceptable number of stages. The proposed intensified process requires just one unit and low energy consumption. For instance, hot utility target for the intensified process using palm oil is significantly lower than the classical process, i.e. 31.8 kW vs 813.5 kW respectively. The main energy consumption is due to distillation columns needed to purify products and to recover methanol. The Pinch Analysis, to determine the target for hot and cold utilities required, considering the distillation columns as part of the whole process, is used. It is shown that the hot utility target required by the classical scheme presented in AspenPlus® documentation can decrease significantly to our intensified scheme, from 33 kW to 10.5 kW.

Acknowledgements The authors would like to thank the financial support of POSCCE project ID 652 (Structural Funds for Development and Cohesion) and the project CTQ2009-11465 (Ministry of Science and Innovation e Spanish Government) who provided the opportunity to complete this research.

Nomenclature

DS DTmin

entropy change in the distillation column pinch temperature (K) Qr energy required by the distillation column (kW) Tc temperature of the distillation column condenser (K) Tr temperature of the distillation column reboiler (K) h Carnot efficiency l enthalpy of vaporization for the distillate stream (kJ/ kmol) R ideal gas constant (kPa m3/kmol K) D distillate flow rate (kmol/s) B bottoms flow rate (kmol/s) xi molar composition of compound i in the corresponding stream. CSTR continuous stirred tank reactor NRTL non random two liquid thermodynamic model EU European Union MeOH methanol NaOH sodium hydroxide FAME fatty acid methyl ester FFA free fatty acids O oleic acid (C18:1) M myristic acid (C14:0) P palmitic acid (C16:0) Li linoleic acid (C18:2) S stearic acid (C18:0) MMM trimyristin OOLi diolein-linolein OOO triolein PLIO palmitin-linolein PPLI dipalmitin-linolein PPP tripalmitin PPS dipalmitin-stearin POO palmitin-diolein POS palmitin-olein-stearin PPO dipalmitin-olein Methyl-M methyl myristate C N:M FFA oil composition, where N is carbon chain length and M is the number of double bonds, e.g. linolenic acid: C18:3. References [1] Plesu V, Bonet-Ruiz J, Bumbac G, Teodorescu F, Tuluc A. Modelling and simulation for sustainable biodiesel production. In: CHISA 2012e20th International Congress of Chemical and Process Engineering and PRES 2012-15th Conference PRES Conference Proceedings, Prague (Czech Republic); 2012. [2] Plesu V, Bonet J, Bonet Ruiz AE, Llorens J. Advantages of process integration evaluated by gibbs energy: biodiesel synthesis case. In: Computer Aided Chemical Engineering eProceedings of the 24th European Symposium on Computer Aided Process Engineering e ESCAPEvol. 24; 2014 [In press]. [3] Bonet J, Plesu V, Bonet Ruiz AE, Iancu P, Llorens J. Thermodynamic study of batch reactor biodiesel synthesis. Rev Chim 2014;65(3):358e61. [4] Badday AS, Abdullah AZ, Lee KT. Ultrasound-assisted transesterification of crude Jatropha oil using cesium doped heteropolyacid catalyst: interactions between process variables. Energy 2013;60:283e91. [5] Ong LK, Effendi C, Kurniawan A, Lin CX, Zhao XS, Ismadji S. Optimization of catalyst-free production of biodiesel from Ceiba pentandra (kapok) oil with high free fatty acid contents. Energy 2013;57:615e23. [6] Alenezi R, Santos RCD, Raymahasay S, Leeke GA. Improved biodiesel manufacture at low temperature and short reaction time. Renew Energy 2013;53:242e8. [7] Jurado MBG, Plesu V, Ruiz JB, Ruiz AEB, Tuluc A, Llacuna JL. Simulation of a hybrid reactive extraction unit. Biodiesel synthesis. Chem Eng Trans 2013;33: 205e10. [8] Cheng J-K, Chao C-C, Ward JD, Chien I-L. Design and control of a biodiesel production process using sugar catalyst for oil feedstock with different free fatty acid concentrations. J Taiwan Inst Chem Eng 2014;45(1):76e84. [9] Oh PP, Chong MF, Lau HLN, Chen J, Choo YM. Liquid-liquid equilibrium (LLE) study for six-component transesterification system. Clean Technol Environ Policy 2013;15(5):817e22.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013

V. Ples¸u et al. / Energy xxx (2014) 1e15 [10] Cadavid JG, Godoy-Silva RD, Narvaez PC, Camargo M, Fonteix C. Biodiesel production in a counter-current reactive extraction column: modelling, parametric identification and optimization. Chem Eng J 2013;228:717e23.  mez GE, Ramos MA, Cadena JE, Go mez JM, Munoz F. Inherently safer design [11] Go applied to the biodiesel production. Chem Eng Trans 2013;31:619e24. [12] Prussi M, Chiaramonti D, Recchia L, Martelli F, Guidotti F, Pari L. Alternative feedstock for the biodiesel and energy production: the OVEST project. Energy 2013;58:2e8. [13] Iancu P, Plesu V, Velea S. Flue gas CO2 capture by microalgae in photobioreactor: a sustainable technology. Chem Eng Trans 2012;29:799e804. [14] Bhaskar K, Nagarajan G, Sampath S. Optimization of FOME (fish oil methyl esters) blend and EGR (exhaust gas recirculation) for simultaneous control of NOx and particulate matter emissions in diesel engines. Energy 2013;62: 224e34. [15] Chakraborty M, Baruah DC. Production and characterization of biodiesel obtained from Sapindus mukorossi kernel oil. Energy 2013;60:159e67. [16] Miller P, Kumar A. Development of emission parameters and net energy ratio for renewable diesel from Canola and Camelina. Energy 2013;58:426e37. [17] Chauhan BS, Kumar N, Cho HM, Lim HC. A study on the performance and emission of a diesel engine fueled with Karanja biodiesel and its blends. Energy 2013;56:1e7. [18] Ng WPQ, Lam HL, Yusup S. Supply network synthesis on rubber seed oil utilisation as potential biofuel feedstock. Energy 2013;55:82e8. [19] Halek F, Delavari A, Kavousi-rahim A. Production of biodiesel as a renewable energy source from castor oil. Clean Technol Environ Policy 2013;15:1063e8. [20] Chisti Y. Biodiesel from microalgae. Biotechnol Adv 2007;25(3):294e306. [21] http://www.originoil.com/pdf/OriginOil_NAA_January_2011.pdf, [accessed 28.06.2011]. [22] Lee HV, Taufiq-Yap YH, Hussein MZ, Yunus R. Transesterification of jatropha oil with methanol over Mg-Zn mixed metal oxide catalysts. Energy 2013;49(1):12e8.

15

[23] Liang X. Synthesis of biodiesel from waste oil under mild conditions using novel acidic ionic liquid immobilization on poly divinylbenzene. Energy 2013;63:103e8. [24] Martins MI, Pires RF, Alves MJ, Hori CE, Reis MHM, Cardoso VL. Transesterification of soybean oil for biodiesel production using hydrotalcite as basic catalyst. Chem Eng Trans 2013;32:817e22. [25] Kemp IC. Pinch analysis and process integration: a user guide on process integration for the efficient use of energy. 2nd ed. Burlington, MA: Elsevier/ Butterworth-Heinemann; 2007. [26] Plesu V, Bonet Ruiz AE, Bonet J, Llorens J, Iancu P. Minimum number of transfer units and reboiler duty for multicomponent distillation columns. Appl Therm Eng 2013;61:67e79. [27] Ignat RM, Kiss AA. Energy efficient recovery of methanol and glycerol in biodiesel production. Chem Eng Trans 2012;29:1141e6. [28] Palmeira JVN, Silva JM, Matos HAS. Industry energy optimization: a case study in a biodiesel production site. Chem Eng Trans 2012;29:919e24. [29] Narvaez PC, Rincon SM, Sanchez FJ. Kinetics of palm oil methanolysis. J Am Oil Chem Soc 2007;84:971e7. [30] Huerga IR, Zanuttini MS, Gross MS, Querini CA. Biodiesel production from Jatropha curcas: integrated process optimization. Energy Convers Manag 2014;80:1e9. [31] Pankaj KM, Suseela R, Toppo K. Physico-chemical characterization of algal oil: a potential biofuel. Asian J Exp Biol Sci 2011;2:493e7. [32] Hakim M, Najafabadi HA, Pazuki G, Vossoughi M. Novel approach for liquidliquid phase equilibrium of biodiesel (canola and sunflower) þ glycerol þ methanol. Ind Eng Chem Res 2014;53:855e64. [33] SPRINT® software http://www.cpi.umist.ac.uk/software/products/sprint.asp [accessed 30.06.2014]. [34] Berchmans HJ, Hirata S. Biodiesel production from Jatropha curcas L. seed oil with a high content of free fatty acids. Bioresour Technol 2008;99:1716e21.

Please cite this article in press as: Ples¸u V, et al., Process intensification in biodiesel production with energy reduction by pinch analysis, Energy (2014), http://dx.doi.org/10.1016/j.energy.2014.11.013