Biodiesel Plant Optimisation Study by using

Aspen-HYSYS® Process Simulator

ALEXANDRU TULUC1, PETRICA IANCU1*, VALENTIN PLESU1, JORDI BONET-RUIZ2, GRIGORE BOZGA1, GHEORGHE BUMBAC1
1
University Politehnica of Bucharest, Centre for Technology Transfer in Process Industries, Department of Chemical and
Biochemical Engineering, 1, Gh. Polizu Str., 011061, Bucharest, Romania
2
University of Barcelona, Department of Chemical Engineering, Martí I Franquès, núm. 1, Planta 6, E-08028 Barcelona, Spain

The work presents a simulation and optimisation study for biodiesel synthesis process, based on rapeseed
oil transesterification with methanol in homogeneous catalysis (KOH catalyst). All necessary calculations
are performed with Aspen HYSYS® 8.3 process simulator. For the transesterification reaction, after a thorough
investigation, two perfectly mixed reactors in series are considered, insuring a triglyceride conversion
corresponding to biodiesel standards as EN14214 conditions for biodiesel purity. The optimisation study is
performed based on economic type objective function representing main costs associated to the
transesterification step of the process, for a fixed production. Decision variables are the volume of each
transesterification reactor, the methanol to triglycerides ratio, the temperature of each reactor and the three
phase separator temperature. The restrictions consider limit for total reactors volume, limit for reactors
temperature, three phase separator temperature interval, minimum methyl oleate mass fraction, and
maximum methanol content of biodiesel. The results reveal that the minimum production cost, in the
optimisation problem so formulated, is 14% lower than the reference case. This is achieved considering
higher methanol to oil ratio (e.g. limited however by methanol recycling cost), each reactor temperature
close to maximum limit, and three separation temperature close to the lower limit. The results are dependent
on the quality of kinetic model. Aspen HYSYS® is appropriate instrument to solve flowsheet optimisation.

Keywords: biodiesel plant, kinetic models, transesterification, process simulation,

cost objective function, flowsheet optimisation

Biodiesel is an alternative renewable fuel with rapid materials. In addition, the catalyst removal from the
worldwide development especially in last two decades. reaction mixture is not a simple issue.
Chemically, biodiesel is a mixture of fatty acids alkyl esters, Industrial scale biodiesel processes are typically setup
obtained by the transesterification of triglycerides with light in continuous stirred tank reactors, achieving triglycerides
alcohols, usually methanol or ethanol. The reactions below conversion higher than 99.3%. An important process
represent transesterification with methanol. parameter is the molar ratio between the two reactants
(methanol: tryglicerides). Conventional processes use
molar ratios between 6:1 and 10:1. As the reactions are
(1) slightly exotermic, reaction temperature is relatively low,
usually in the range of 60°C to 80°C. The transesterification
(2) reaction is performed in at least two reactors in series, to
achieve high conversion.
Four different transesterification processes using virgin
(3) canola oil and waste cooking oil as raw materials are
compared in [15] and [16]: alkali-catalyzed process, using
Those reactions highly dominated by chemical virgin canola oil as raw material; alkali-catalysed process
equilibrium. As source of triglycerides usually there are using waste cooking oil as raw material (including a
vegetable oils and animal fats as most important [3]. The pretreatment unit for free fatty acids removal); acid
reaction takes place in the presence of basic or acid catalysed process using waste cooking oil as raw material;
catalysts, in homogenous or heterogeneous medium [4, the fourth process has same conditions as third process,
7]. Usually, industrial scale processes are applied with the modification consists in hexane or petroleum ether as
basic homogenous catalysis, as high conversion in a solvent in FAME separation step. The authors compared
relatively short time is achieved [5, 7, 9, 12]. Typical basic the techno-economic performance of four processes, using
catalysts are alkaline hydroxides (NaOH, KOH) or Aspen HYSYS® process simulator, and concluded [17] that
methoxides (CH 3ONa) in the case of homogenous the alkali catalysed process using virgin oil (first process)
catalysis. Calcium oxide, hydrotalcites or sodium silicate gives the lowest capital cost.
are typical heterogeneous catalysts [4], [10]. Acid catalysts A techno-economic analysis for a biodiesel synthesis
give relatively poor performance, especially regarding the process was performed also in [1], with the aim to
amount of time needed to achieve reasonable triglyceride investigate the influence of critical profitability indicators
conversion [15, 16]. on production capacity. The authors concluded that the
The main disadvantages of the homogenous catalysis raw materials cost accounts for 75 % of the production
are: catalyst recovery is not possible and the equipment cost for small plants, and it is possible to increase up to 90
should be corrosion resistant, which involve expensive % for large plants.

* email: p_iancu@chim.upb.ro

REV. CHIM. (Bucharest) ♦ 66 ♦ No. 4 ♦ 2015 http://www.revistadechimie.ro 565

 Table 1
SIMULATION AND ECONOMIC
STUDIES REGARDING BIODIESEL
SYNTHESIS PROCESSES

An economic analysis of a biodiesel production plant Brassica Carinata oil transesterification kinetics was
was reported in [6]. The raw material for the trans- studied in [11], for following conditions : temperature 25°C
esterification process was soybean oil. The predicted to 65°C, molar ratio methanol to oil 6:1, basic catalyst
biodiesel cost was 0.4 euro/L and the major contributor to concentration 0.5-1.5% KOH (based on oil mass).
this cost was also the cost of the oil feedstock, which Soybean oil transesterification in the presence of BuONa
accounted for 88% of the total estimated production cost. or sulfuric acid as catalysts were reported in [5]. They
The transesterification of soybean oil in alkali based concluded that the basic catalysis was the most indicated
catalysis was reported in [14]. The reaction temperature for this kind of reaction. The study was performed in
was 60°C and the molar ratio between methanol and oil following conditions : temperature between 20 and 60°C,
was 6:1. Following the economic analysis performed by reactants molar ratio methanol to oil 6:1, and catalysts
the authors, break-even cost for biodiesel was 0.53 euro/L. concentration was 1% (based on oil mass).
Another economic analysis for an alkali based Sunflower oil transesterification of with methanol kinetic
transesterification plant was reported in [10]. The reaction analysis, in the presence of NaOH as catalyst was reported
temperature was considered 60°C, and the molar ratio in [2]. The kinetic model was developed for following
methanol to oil was 6:1. The methanol and water were conditions: temperature 20 and 60°C, methanol to oil molar
removed by vacuum distillation from the biodiesel product. ratio 6:1, and catalyst concentration in the interval 0.25-
The estimated biodiesel cost was 0.47 euro/L. 1% NaOH (based on oil mass).
The main studies investigating the economics of Based on the reaction mechanism, the majority of
biodiesel synthesis processes, based on computer published kinetic models consider second order reaction
simulations are presented in table 1. rates in molar concentrations (the form 4 to 6), neglecting
In spite of the significant number of simulation studies, the partial solubility of reactants and the non-ideal
overall biodiesel plant optimisation is less considered in behaviour of the reaction medium. The kinetic parameters
the open literature. The aim of this work is to evaluate the published in various studies are presented in table 2.
main process parameters value, minimizing the cost of
biodiesel production, and to analyze the potential of Aspen-
HYSYS® process simulator for flowsheet optimisation. (4)

Reaction kinetics
Because the transesterification reaction is reversible and (5)
slightly exothermal the equilibrium conversion of
triglycerides is favoured by low temperatures and high
methanol excess.
In order to simulate the transesterification reactor, an (6)
indispensable information is regarding the process kinetics.
One of the first studies regarding the kinetics of the It the above kinetic models, reaction rate constants are
transesterification reaction was carried out by [8]. The arrhenius form:
reaction conditions were: temperature 30-70°C, molar
reaction 6:1, catalyst concentration 0.2% (based on oil
(7)
mass) and the raw material was soybean oil.

Table 2
KINETIC PARAMETERS FOR
VARIOUS BIODIESEL SYNTHESIS
STUDIES

566 http://www.revistadechimie.ro REV. CHIM. (Bucharest) ♦ 66 ♦ No. 4 ♦ 2015

Process description - mixing in the two autoclave reactors is considered to
One of the most important issues regarding trans- be ideal (Continuous-Stirred Tank Reactor, CSTR).
esterification of vegetable oils with an alcohol is the The thermodynamic model selected to describe the
immiscibility of the reactants. Thus, to achieve high behaviour of different mixtures involved in the simulation
triglyceride conversions, reactants advanced contacting is NRTL (liquid phase) and ideal (vapour phase) [17].
(intensive mixing) is required. Two types of reactors are Previous studies revealed that this model seems to be most
most accessible and commonly used in practice, the suitable for glycerides transesterification process [15]. The
mechanically stirred autoclave, and the tubular reactor. The missing binary coefficients are estimated using UNIFAC
mechanical stirred autoclave has the advantage of a very group contribution method. Among the three kinetic
good reactants mixing, but the required reaction volume models presented above (table 2), the study published in
necessary to achieve high conversions is relatively large. [2] is considered as the most appropriate for simulation of
The tubular reactor presents the advantage of a lower selected process.
volume necessary to achieve high levels of conversion (due One of the most important restrictions of biodiesel
to low intensity of axial mixing), but intimate reactants synthesis process is product purity. In agreement with
contacting is difficult to be achieved. European Standard EN14214, glycerides concentration in
To achieve good reactants contacting and high biodiesel has to be smaller than 0.2 wt % for triglycerides,
conversions, in reasonable reaction volumes, a series of 0.2 wt % for diglycerides, and 0.7% for monoglycerides.
two or more stirred reactors appears to be a suitable This purity can be achieved for very high triglyceride
solution. In this work, a flowsheet for the biodiesel conversion in the reaction section (at least 99.3%). Lower
synthesis process, having the structure presented in figure triglyceride conversion involves expensive separation of
3 is considered. The raw material streams are mixed in a fatty acid methyl esters and unreacted glycerides [16].
static blender, heated at reaction temperature and pumped A study is carried out to decide the number of reactors
in the first reactor autoclave type of the series, at pressure required to achieve high triglycerides conversion. Four plant
2 bar, with pump P-100. The effluent of the last reactor of configurations are studied, with the transesterification step
the series is heated, flashed and fed to three phase performed, alternatively, either in one CSTR reactor or in a
separator D-100. Three streams are obtained : ester phase, series of two, three or four reactors. Simulation results are
glycerol phase and methanol vapours. As methanol is presented in table 3 and figure 1. It is to notice a strong
favouring the reciprocal miscibility of glycerol and fatty decrease of total reaction volume with the number of
acid esters, the glycerol separation is promoted by reactor units to achieve same conversion. This is the result
diminishing methanol concentration in the liquid mixture. of the average transesterification rate increase with the
So, for good separation of glycerol product, methanol and number of reaction stages. Nevertheless, two reactors are
glycerol have to be simultaneously separated from the selected to perform the transesterification step, due to
mixture. Preliminary calculations shown that an adequate increase of operation and investment expenses associated
separation of methanol from liquid phases is insured at with increased number of reactors. All the reaction
temperature 90°C and a pressure of 0.5 bar in three phase conditions are kept constant (methanol flowrate 126 kmol/
separator D-100. h, reaction temperature 65°C, three phase separator
Methanol vapour stream from D-100 is condensed in temperature 90°C, for a triglycerides conversion 99.3%).
heat exchanger E-102 and recycled to feed point with pump Following this study two reactors configuration is chosen,
P-101. The glycerol stream resulted from three phase because it has the highest influence on the overall reaction
separator D-100 contains the base catalyst which should volume, and there is also relatively simple to implement.
be neutralized with a mineral acid (HCl aqueous solution) The total reaction volume and its distribution amongst
in the reactor CSTR-102, and further fed to the rectifying the two CSTR units considered in the simulation are given
column T-101 for methanol advanced separation. The raw in table 4.
glycerol stream resulted from the bottoms of this column The effluent from the second reactor is heated up to
is further purified and valorised in different ways (this step 90°C, then flashed to 0.5 bar pressure, then separated in
is not considered in this work). Similarly, the biodiesel three phase separator D-100. The glycerol stream is fed in
product stream resulted from the separator D-100 is fed to CSTR-102, where the catalyst is neutralized with
rectification column T-100 for advanced methanol hydrochloric acid (20% aqueous solution). The separation
separation. To avoid thermal degradation of heavy of methanol-glycerol mixture is achieved in rectification
components (glycerol and fatty acid esters respectively), column T-101 (4 theoretical trays - NTT). The methanol
it is necessary to operate rectification columns at low contained in biodiesel stream is also separated in T-100
pressure (0.5 bar), so that the bottoms temperature of these rectification column (with 4 theoretical trays). The bottom
rectification columns to be lower than 170°C. temperature for both rectification columns is maintained
Process simulation Table 3
A first step of the study consists in a process simulation REACTOR VOLUME (METHANOL FLOWRATE 126 KMOL/H ,
at conventional values of operating and constructive REACTION TEMPERATURE 65°C, DECANTER OPERATING
parameters, considering a rapeseed oil transesterification TEMPERATURE 90°C, TRIGLYCERIDES CONVERSION 99.3%)
plant with the capacity of 100,000 t/year. The process is
simulated using Aspen HYSYSÒ v8.3 process simulator. To
simplify the process calculations, the following
assumptions are considered:
- triglyceride content in oil raw material is considered as
pure triolein;
Table 4
REFERENCE WORKING CONDITIONS (FOR
TRIGLYCERIDES CONVERSION 99.3 %)

REV. CHIM. (Bucharest) ♦ 66 ♦ No. 4 ♦ 2015 http://www.revistadechimie.ro 567

 Table 5
METHANOL SEPARATION CONDITIONS AND PERFORMANCE IN
COLUMNS T-100 AND T-101
Fig. 1. Total reaction
volume versus the
number of reactors

Fobj = Cmethanol + Cutilities + Creactors (8)

Following restrictions are considered for the optimisation

problem:
-total biodiesel reactors volume (CSTR-100 and
CSTR101) – less than 50 m3;
below 170°C, as long as the pressure is smaller than 0.5 -reactors temperature – less than 65°C;
bar. Table 5 summarises the performance of both -three phase separator temperature from 90°C to 140°C;
rectification columns. -methyl oleate mass fraction in biodiesel stream >97%;
The simulation study shows that a triglyceride -methanol mass fraction in biodiesel stream <0.2%;
conversion of 99.3 % can be achieved considering two CSTR Biodiesel reactors (CSTR-100 and CSTR-101) total
reactors in series, having each the volume of 9 m3, operation volume is restricted to a maximum of 50 m3, because
temperature 65°C, pressure 2 bar, catalyst concentration higher volume can introduce mixing problems, thus
of 1% KOH, based on oil flow rate, and methanol to oil degrading oil overall conversion. The methanol flow rate
molar ratio of 9:1 in the feedstock. has the highest impact on the objective function, so the
limits are less restrictive. Biodiesel synthesis temperature
Process flowsheet optimisation is limited to 65oC, considering literature reported data (Table
Based on the simulation results previously presented, 1), to avoid secondary saponification reactions [9]. Three
subsequently an optimisation study is carried out for same phase separator temperature is chosen with regard to the
biodiesel plant, with capacity of 100,000 t biodiesel/year. degree of methanol separation required. Higher
As the plant capacity and the triglyceride conversion (min temperatures have the disadvantage of higher overall costs,
99.3 %) are fixed, the triglyceride feed flowrate is also fixed but at lower temperatures (below 90°C) the separation of
and corresponds to approximately 14 kmol/h. All the methanol is week, and high content of methanol
separation equipment is considered of fixed capacity, given determines that biodiesel - glycerol liquid mixture becomes
by the simulation study previously presented. For homogenous, and two liquid phases cannot be separated.
optimisation the decision variables are : methanol to oil Biodiesel stream concentration restrictions observe the
ratio, the volume of each of biodiesel reactors (CSTR-100 provisions of European Standard for biodiesel purity.
and CSTR-101), and the three phase separator D100 Prices for raw materials and utilities are presented in
temperature. table 6. Vegetable oil is the most expensive raw material
The objective function to be minimised is annualised used in the transesterification process. Although methanol
production cost (eq. 8). This represents the summation of has a higher unit price than oil, its contribution for the overall
the amortization costs for both CSTR reactors (considering process cost is small (the methanol net flow rate is much
a depreciation period of 10 years), the cost of reacted smaller as compared to oil flowrate). Vegetable oil is the
methanol, and the energy costs for different utilities most expensive raw material used in the transesterification
consumed in flowsheet heat exchangers (E-100, E-101, E- process. AspenHYSYS® process simulator has a number of
102, E-103, E-104, etc.), [13]. different methods implemented in its optimisation toolbox.
In this work, the Box method appeared the most suitable.

Fig. 2 Triglycerides
transesterification process
flowsheet

568 http://www.revistadechimie.ro REV. CHIM. (Bucharest) ♦ 66 ♦ No. 4 ♦ 2015

 Table 6 Table 7
METHANOL AND UTILITIES COSTS OPTIMAL CONDITIONS

Results and discussions Acknowledgement: Authors are deeply recognising the financial
As the optimisation problem previously formulated is support of Structural Programme POSCCE, project ID 652.
characterized by four independent variables (methanol to
oil ratio, CSTR-100 volume, CSTR-101 volume, and D-100 Notations and abbreviations
temperature) The process is highly nonlinear, the global ·Ci – molar concentration of component i, kmol/m3;
extreme proved relatively difficult to be located using Aspen ·rj – reaction rate, kmol/(m3·h);
HYSYS® built-in functions. To identify the global minimum ·Ea,i – activation energy for reaction j, kJ/kmol;
of the objective function, an important number of ·T – reaction temperature, K;
successive optimisation trials with Aspen HYSYS® are ·R – ideal gas constant, 8.31 kJ/(K·kmol);
performed with different initial values of the independent ·k0,j – preexponetial factor for reaction j, m3/(kmol·h);
variables. The final results are presented in table 7. Indexes :
To minimize the objective function value and to keep oil -j = 1, 2, 3 the reactions of transesterification, presented in (1), (2),
conversion above 99.3 %, the fresh methanol flowrate and (3)
decreases and, as a consequence, the reactors volume is -i = TG, DG, MG, GL, ME, M chemical species involved in biodiesel
slightly increased. The optimum results of this problem synthesis
represent a trade-off between methanol recycling cost and
equipment costs. Lower methanol recycling cost favours Abbreviations
-TG – triglyceride
the excess of methanol (high methanol: oil ratios) and
-DG – diglyceride
smaller reactor volumes, and vice versa.
-MG – monoglyceride
The optimized operating conditions shown in Table 7 -GL – glycerol
correspond to reactors temperature close to maximum -ME – methyl ester of fatty acids
limit. -M – methanol.
Comparing with the reference case, considered in the
simulation study presented previously, the objective References
function decreases by approximately 14 %, which 1.APOSTOLAKOU A.A., KOOKOS I.K., MARAZIOTI C., ANGELOPOULOS
represents a significant reduction of production expenses. K.C., Fuel Processing Technology, 2009, doi:10.1016/
Naturally, the correctness of these results is directly j.fuproc.2009.04.017.
dependent on the quality of the kinetic model used in 2. BAMBASE M.E.JR., NAKAMURA N., TANAKA J., MATSUMURA M.,
calculations. In this context, it is worth to underline that, in Journal of Chemical Technology and Biotechnology, 82, 2007, p. 273-
spite of its practical importance, only few kinetic models 280.
considering the reaction reversibility and accurate rate 3. DIMIAN A.C., BILDEA C.S., Chemical Process Design, WILEY-VCH
temperature dependencies for transesterification of Verlag GmbH & Co. KGaA, Weinheim, Germany, p. 426, 2008.
vegetable oils in homogeneous catalysis, are accessible 4. DOSSIN T.F., REYNIERS M.F, MARIN G.B., BERGER R.J., Applied
in open literature. Catalysis B Environmental, 67, 2006, p. 136–148.
5. FREEDMAN B, BUTTERFIELD R.O., PRYDE E.H., Journal of American
Conclusions Oil Chemical Society, 63, 1986, p. 1375-1380.
One of the main steps in the biodiesel synthesis from 6. HAAS M.J., McALOON A.J., YEE W. C., FOGLIA T. A., Bioresource
vegetable oils is triglyceride transesterification. Technology, 97, 2006, p. 671–678.
Based on a published kinetic model, the present 7. NOORDAM, M., WITHERS, R.V., Idaho Agricultural Experiment Station
simulation study is evaluating the reactor volume Bulletin, 785, 1996, p. 12.
necessary for a series of two CSTRs, in conditions of limited 8. NOUREDDINI R., ZHU D., Journal of American Oil Chemical Society,
methanol recycling. Several simulations revealed the 74, 1997, p. 1457.
influence of reactors number on overall reaction volume 9. RASHID U., ANWAR F., Fuel, 87, 2008, p. 265–273.
to achieve a given total conversion. The largest volume 10. SAKAI T., KAWASHIMA A., KOSHIKAWA T., Bioresource Technology,
difference is between one reactor and a series of two 100, 2009, p. 3269-3276.
reactors. Therefore, in the remaining part of the work a 11. VICENTE G., MARTINEZ M., ARACIL J., Energy & Fuels, 20, 2006, p.
series of two reactors configuration is chosen for the 1722-1726.
optimisation study. 12. WEBER J.A., The economic feasibility of community-based
An optimisation study is also performed to find optimal biodiesel plants, MSc. Thesis, University of Missouri, Columbia, USA,
solution for methanol to triglyceride ratio, the volume of p. 108, 1993.
each unit of the reactor series and the temperature of three 13.***http://en.wikipedia.org/wiki/Biodiesel,http://www.indexmundi.
phase separator, minimizing the cost of biodiesel com/commodities/?commodity =rapeseed-oil, http: //en .wiki pedia.
org/wiki/ Electricity_pricing, consulted at 23.08.2014.
production. As stated above, the accuracy of this study is
14. YOU Y.D., SHIE J.L., CHANG C.Y., HUANG S.H., PAI C.Y, YU Y.H.,
essentially dependent on the quality of the available kinetic
CHANG C.H., Energy & Fuels, 22, 2008, p. 182-189.
model. The study showed that a decrease of 14 % 15. ZHANG Y., DUBÉ M.A., MCLEAN D.D., KATES M., Bioresource
compared to reference data is achievable by flowsheet Technology, 90, 2003, p. 1-16;
optimisation. 16. ZHANG Y., DUBÉ M.A., MCLEAN D.D., KATES M., Bioresource
The study underline the capacity of the AspenHYSYS® Technology, 90, 2003, p. 229-240.
process simulator to perform optimisation studies directly 17. *** ASPEN HYSYS V7.3 SIMULATION BASIS, Aspen Technology,
in the simulation environment, on the flowsheet, even for Inc., Burlington, USA, 2011
quite complex flowsheet as that studied in this paper.
Manuscript received: 8.09.2014
REV. CHIM. (Bucharest) ♦ 66 ♦ No. 4 ♦ 2015 http://www.revistadechimie.ro 569