Student Contest Problem 2018 | EURECHA

Preliminary design of the green diesel production

process by hydrotreatment of vegetable oils
Javier Moreno Fernández-Villamil∗,1 and Aurelio Hurtado de Mendoza Paniagua∗
∗ Polytechnic University of Madrid (ETSII)

ABSTRACT
In this work, a conceptual design is presented of a HVO/green diesel production unit with a processing capacity of 74
ton/h (500 000 ton/year) of vegetable oils and a production rate of 59 ton/h of diesel. Firstly, an extensive literature review
has been conducted regarding the state-of-the-art techniques as well as process equipment, mechanisms of reaction and
thermodynamical properties. A market analysis is also presented which estimates feedstock availability and target production
rate. With this information, a preliminary Process Flow Diagram is proposed, along with explanations on the type of equipment
used and its operating conditions. Process design and simulation has been performed using Aspen Plus®, while Aspen Custom
Modeler® has been used to develop more accurate models where necessary. The present study concludes with an analysis of
process flexibility, considerations for heat integration and an economic assessment.
KEYWORDS Green diesel; Hydrotreatment of Vegetable Oils; Computer-aided Process Engineering

Introduction tional diesel engines, mainly due to its stability and cold flow
properties [2]. On the other hand, the sustainability of these
n the last decade, alternative fuels have gained increasing
I attention from the scientific community, given the growing
energy demand and the global warming effect derived from
processes has been questioned, since the food-based crops that
are commonly used as feedstock tamper with food prices and
availability in the regions of growth, a problem that has been
CO2 emissions. Even though fossil fuels are still the main source addressed carefully by Articles 17 and 18 of the RED [3].
of energy across the globe, the European Union has made great In light of the facts, it only seems logical to propose an alter-
effort to change the current situation through the Renewable native renewable fuel that is fully compatible with current fossil
Energy Directive (RED), which enforces a 20% target for renew- fuels, as not avoid any changes in the current transportation
able energy sources and a 10% in the transportation sector by infrastructure and conventional engines. Green diesel obtained
the year 2020. from hydrotreatment of vegetable oils is one, well-established
With no more than a year to reach the due date, the EU has example. Even though it uses the same feedstock as biodiesel,
been working on a second version of the RED directive, the RED green diesel yields a product with better qualities, which is also
II, which raises the target of renewable energy to 32% by the indistinguishable from fossil fuel-derived diesel. Table 1 com-
year 2030 [1]. Given the current paradigm shift, it is clear that pares the properties of conventional diesel, biodiesel and green
changes must be made in the appropriate direction. diesel.
The two main biofuels used in transportation are bioethanol Hydrotreated vegetable oils (HVOs) seem like a suitable al-
used in gasoline and biodiesel or FAME (Fatty Acid Methyl Es- ternative to biodiesel due to their improved cold flow properties,
ter) in diesel. As promising as the word biofuel sounds, there low content in sulfur, aromatics and oxygen, and engine com-
are two main concerns affecting these products which shall be patibility.
carefully considered. On one hand, engine compatibility makes The objective of the present work is to develop a conceptual
it undesirable to use blends of more than 7% of FAME in conven- design for the processing of vegetable oils into green diesel.
Technical, economical and environmental aspects will be taken
Manuscript compiled: Monday 31st December, 2018 into account as indicators of the quality of the project.
1
[email protected] First, a literature review is presented in order to gather the

A proposal for the HVO process 1

 Diesel HVO (NeXBTL) FAME
CEN standard EN 590 n/a EN-14214
Density at 15◦ C (kg/m3 ) 820-845 770-790 860-900
Cetane number >51 75-99 >51
Higher heating value (MJ/kg) 43.1 44.1 37.2
Blending <7 0-100 <7
Flash point (◦ C) >55 >61 >101
Cloud point (◦ C) >-20 >-40 -5

Table 1 Properties of summer diesel, FAME and Neste’s green diesel NeXBTL [4].

necessary information for the design of the process and selection The two main technologies for hydrotreatment of vegetable
between alternatives, followed by a market review to support oils are NExBTL® by Neste Oil and EcofiningTM by UOP and
the financial viability of the venture. The selected process and ENI. The basic scheme for both is quite alike, as can be seen in
its variables are then put to test through simulation in Aspen Figures 1 and 2. Other companies, such as Axens IFP or Haldor
Plus ®, allowing for the optimization of process variables and Topsoe, have also developed similar technologies by the name
consumption of utilities through heat integration. of VeganTM and HydroflexTM .
Upon completion of the simulation and technical design
stage, an economic assessment will help decide on the economi- Degumming
Feedstock and metal Solids
cal feasibility of the project and identify main cost contributors. removal
Another, no less important key aspects such as carbon consump- CO, CO2, light Hydrogen
Water hydrocarbons
tion and sustainability of the process are also studied.
Naphtha
Hydrogen Separation Product
Hydrotreating Isomerization
Literature review stabilization

Diesel
Literature research for this project attains the technology used
for stand-alone processing of vegetable oils into green diesel. Figure 2 Block diagram for the NExBTL® process [6].
Many refiners prefer instead to follow the coprocessing scheme,
where vegetable oils are injected into the fuel mixture going to Both processes consist of a primary phase called deoxygena-
the hydrotreaters in a refinery. Minor investment and revamping tion, where hydrogen is used to break down the triglycerides
is needed with the latter, but the possibility of a 100% renewable and eliminate oxygen molecules at high temperatures and pres-
fuel is left out. Therefore, this work will focus only in stand- sures. The result is a linear hydrocarbon mixture in the diesel
alone processing of HVOs. boiling range which is mainly paraffinic, as well as CO, CO2 ,
Hydrotreatment of vegetable oils could, as a first approach, water and light hydrocarbons. After separation in a three phase
be compared to the conventional hydrotreatment units used flash tank, the mixture of paraffins is partially isomerized in
to remove sulfur from refinery products. Annex 1 contains a order to improve its cold flow properties. Isomerization is essen-
schematic representation of this process which, when compared tial in order to obtain a high quality product, but it comes at the
with the novel green diesel technologies developed in the last expense of partial loss of the diesel product due to cracking. As
decade, provides substantial proof of the similarity between a result, a final stabilizer column is needed in order to separate
both hydrotreatment processes. lighter hydrocarbons from the heavier ones, resulting in four
possible products: LPG and naphtha as head products, jet fuel
and green diesel as bottom product.
Deoxygenation and isomerization take place in presence of
catalysts, mainly CoMo and NiMo catalysts for the hydrotreat-
ment section and Pt-based catalysts for the isomerization unit [7].
The conditions in which these reactions take place are what deter-
mine the level of cracking, and thus the amount of by-products
that are obtained.
Nevertheless, given the need for a flexible and a robust pro-
cess, the design will include a stabilizing unit which allows to
vary the amounts of green naphtha and jet fuel that are obtained,
depending on market demand and suitability. Regarding the
separation of gaseous products, an amine unit can be used to
Figure 1 Schematic representation of UOP’s Ecofining technol- separate hydrogen, light hydrocarbons, and contaminants such
ogy [5]. as CO and CO2 .

2 Javier Moreno and Aurelio Hurtado de Mendoza

Feedstock and feedstock pretreament
Before entering the hydrotreating reactor, feedstock must go
through a purification unit as to remove any impurities and
metals contained in it. Pretreatment becomes indispensable, as it
ensures metals (Ca, Mg) and phosphorus will not interfere with
the catalysts in the reactor [8].
The pretreatment unit is based on a bleaching unit, and its
operating conditions vary depending on the type of feedstock.
Bleaching takes place by degumming with addition of an acid
such as H3 PO4 , neutralization with NaOH, and washing with
demineralized water. In order to further reduce the level of im-
purities, the feedstock can be fed through silica and/or bleaching
earth [9].
There are three feedstock alternatives for this project: palm, Table 2 Iodine value of different vegetable oils [10].
rapeseed, and sunflower oil. None of these would allow the
final product to enter the second generation biofuels category, was retrieved from the Food and Agriculture Organization of
but on the other hand they may imply reduced expenses in the the United Nations [12].
pretreatment unit, since they are cleaner than waste oils. The final - and probably most important - aspect regarding
Vegetable oils are made up mostly of triglycerides, which are feedstock selection is price. The surge of interest in biofuels has
hydrocarbon structures containing three fatty acid chains and a translated into uncertainty and price volatility in the recent years.
glycerol backbone in an ester. Processing of these triglycerides Sunflower, rapeseed, and palm oil are all within the category of
will include (I) Saturation of double bonds in the hydrocarbon middle-priced vegetable oils, although palm oil is considerably
chains, (II) Separation from the glycerol backbone in the form cheaper than the rest [13]. A better insight of this matter is
of propane, and (III) Elimination of oxygen molecules through presented in the Market Analysis section.
reduction with hydrogen.
Annex 2 contains information on the composition of all three
types of feedstock, according to the literature [14-16], as well
as practical information about the estimation of their physical
properties, according to the fragment approach model proposed
by Zong et al. [17].

Deoxygenation of vegetable oils

Most of the literature regarding the processing of vegetable oils
focuses mainly on its deoxygenation. This is due to the recent
interest hydrotreatment processes have sparked in the recent
years, but also because companies carefully keep confidential
Figure 3 Example of triglyceride structure typically found in information safe and away from the eyes of the competition.
palm oil. As a result, there are many questions regarding the mech-
anism of this process: reaction kinetics, catalyst use and
One important difference between feedstock types is their io- deactivation, and reactor design and conditions. Through this
dine value, which refers to the number of double bonds present work, the authors have the objective of laying down the basic
in the hydrocarbon chains belonging to the fatty acids. A higher information regarding this complex process and its simulation.
iodine number implies a higher number of double bonds and
thus an increased consumption of hydrogen when being pro- Basics of deoxygenation of vegetable oils
cessed [10]. Table 2 proves that, among the three options, palm Many authors have contributed greatly to the understanding
oil is the one with a smaller iodine value. of the deoxygenation of VOs [2,18-25]. Out of all of them, the
Another important aspect of feedstock is its availability. work of Kubicka et al.and Tirado et al.[26-27] deserve a special
World production of vegetable oils has increased dramatically mention, due to their outstanding and practical approach on
in the past decade. As of 2017, world production of vegetable both reaction mechanisms and modeling approaches. While
oils and fats is 226 million tons [11]. Top world producers of Kubicka has been a prominent researcher in the field, the work
rapeseed oil include China, Germany, and Canada. Germany of Tirado elaborates a synthesis of previous works regarding
produces up to 3.5 million tons of rapeseed oil per year, mainly reaction kinetics and reactor modeling.
to feed livestock. Sunflower oil is produced in less quantities, The mechanism of hydrotreatment of triglycerides takes place
Russia being it biggest producer with an annual average produc- in two phases. First, double bonds in the hydrocarbon chains
tion between 1993 and 2014 of 1.9 million tons. are saturated by hydrogen. Then, reduction takes place by the
Palm oil takes the lead in terms of production: 65 million tons elimination of oxygen atoms through three possible parallel
per year. This means that palm oil accounts for more than half of mechanisms: hydrodeoxygenation, decarboxylation and decar-
the world market share together with soy bean oil. In terms of bonilation. A schematic representation of this mechanism is
location, palm oil is mainly grown in tropical countries, Malaysia shown in Figure 4.
and Indonesia being its main producers. All production data As seen in Figure 4, decarbonylation and decarboxylation

A proposal for the HVO process 3

 +3 H₂ +3 H₂ of sulfur-containing agents in order to prevent deactivation,
Gly-R₃ Gly-R’₃ C₃H₈ + 3C₁₇H₃₅COOH worsening the emission performance of the final product [31].
+3 H₂ Nevertheless, given the sulfur-free quality of vegetable oils,
Decarboxylation Hydrodeoxygenation
+ H₂ green diesel contains less sulfur than conventional diesel [4]. As
Decarbonylation
for the addition of these agents, sulfur concentration in any typi-
C₁₇H₃₆ + CO₂ C₁₇H₃₆ + CO + H₂O C₁₈H₃₈ + 2H₂O cal hydrogen feed coming from a conventional hydrotreatment
unit in a refinery will suffice.
Figure 4 Schematic representation of deoxygenation of a typi- As alternatives, several studies have focused mainly on the
cal triglyceride [28]. use of sulfur free, noble metals such as Pt and Pd. Whereas
NiMo and CoMo sulphided catalysts yield products through
reactions (DCO) yield paraffins with and odd number of carbons, both HDO and DCO mechanisms, noble metals show high selec-
as fatty acids are usually conformed of an even number of them, tivity for DCO routes, resulting in higher CO2 and CO produc-
whereas hydrodeoxygenation reactions (HDO) yield paraffins tion [20, 23].
with an even number of carbon atoms. Other disadvantages of noble metals such as Pt or Pd include
It is unknown whether decarbonylation is part of the pro- their high price, lower yields, and rapid deactivation in the
cess, as the reaction byproduct CO could also be involved in presence of sulfur and water. Xin Le et al.[29] made a thorough
the complex reaction system between CO2 , CO, CH4 and water review of these types of catalysts and concluded that, in spite of
that is depicted in Figure 5. As a result, there is no unequivo- their cleaner performance, there is still progress to be made in
cal evidence to support whether decarbonylation takes place, these aspects.
according to gas chromatograms obtained from different experi- Furthermore, Sotelo-Boyás et al.[7] studied the effect
ments [2, 18-19, 30]. of zeolite-supported Pt catalysts and compared them to
NiMo/Al2 O3 . He concluded that the stronger acidic sites of
the first induced a higher yield of isomerized products at the
expense of a higher degree of cracking, with the subsequent loss
of valuable diesel product.
As a result, it seems more proper to invest in two separate
reactor systems: one for deoxygenation and another one for
isomerization. This way, NiMo/Al2 O3 catalysts can be used to
achieve total conversions in the first reactor and, after proper
separation of products, cold flow properties can be improved
over Pt-zeolites by control isomerization to avoid cracking.

Reactor conditions
Figure 5 Possible additional reactions taking place alongside When controlling deoxygenation, main reactor variables and
deoxygenation [29]. conditions are: temperature, pressure, H2 /oil ratio and liquid
space hourly velocity (LSHV). As shown in Table 3, typical op-
erating conditions do not vary considerably from one study to
Regarding hydrogen consumption, one could infer from
another.
the reaction mechanism depicted in Figure 4 that HDO routes
should be avoided, as it consumes three times the amount of In one particular study, Bezergianni et al.[32] concluded that
hydrogen. However, this would only be true if CO2 and CO increasing reactor temperature above 360 C translated into an
obtained in these reactions didn’t take part in the hydrogen- undesirable degree of cracking.
consuming reactions depicted in Figure 5. In fact, if the opposite In another study, Srifa et al.[21] experimented with a wide
was to happen, studies confirm as much as 19% more hydrogen range of hydrotreating conditions for palm oil over a sul-
would be consumed in the DCO route [2, 26]. phided NiMo/Al2 O3 catalyst. Results supported the hypothe-
Furthermore, the HDO route yields a cleaner green diesel sis that, under 300◦ C, partial conversion of triglycerides took
product, as only 3 atoms of carbon per triglyceride are not place, whereas cracking and isomerization broke down diesel
recovered in the liquid fuel, as opposed to the 6 carbon atoms molecules above 360◦ C.
lost in the DCO route, three more carbon atoms lost in the form Regarding H2 /oil ratios, 3 to 5 times the theoretical consump-
of CO and CO2 which will have to be separated downstream. tion is recommended in order to ensure gas-to-liquid mass trans-
fer resistance of hydrogen does not play a role of interference in
Catalyst systems in deoxygenation reactions the process [2, 21].
Sulphided NiMo and CoMo catalysts supported on alumina LSHV values, on the other hand, may vary between 1 and
are the most common catalysts used in the industrial scale pro- 5 h−1 [33], although higher values could affect the level of
duction of green diesel [7]. These are also the same type of conversion.
catalysts used in conventional hydrotreatment units, which, to-
gether with their high activity, explains their widespread use. Estimation of properties
The high activity of NiMo /Al2 O3 catalyst systems allow to For the estimation of thermophysical properties in the hy-
reach the 100% conversion rates widely reported in the literature drocarbon mixtures, it is recommended in the literature that the
[2, 26, 29]. However, these catalysts need a continuous addition Redlich-Kwong or Peng-Robinson equations of state be used [19,

4 Javier Moreno and Aurelio Hurtado de Mendoza

 Study number Operating conditions Reference
1 390 C, 50 bar, 1.5 h−1 [34]
2 300 C, 30-50 bar, 1.2 h−1 , 500-1500 Nm3 /m3 [21]
3 300 C, 50 barg, 10-100 h−1 (WSHV), 1250 Nm3 /m3 [2]
4 350 C, 40 bar, 7.6 h−1 , 800 Nm3 /m3 [35]
5 340 C, 40 bar, 1 h−1 , 500-1000 Nm3 /m3 [36]
6 335-360 C, 30-60 bar, 20 mol H2 /mol oil [18]

Table 3 Typical reactor operating conditions for hydrotreatment of vegetable oils over sulphided catalysts, as found in the literature.

27, 37]. If the presence of water is important, then the NRTL converges easily and is exempt of intermediate products. Fur-
property package needs to be included [37]. All of these property thermore, the model is in good agreement with what is known
packages are included in Aspen Plus®. about the mechanism of the hydrotreatment of VOs, as discussed
Regarding the properties of triglycerides, two methods were earlier.
found in the literature for their estimation with a reasonable Nevertheless, Vélez did not take into account the effect of
level of precision within less than 10% of error: Zhongs’s hydrogen pressure and mass transfer resistance, and the predic-
fragment approach method, which allows to calculate both tions are not accurate enough (R2 = 0.8). Table 4 represents the
thermodynamic and transport properties and is illustrated in proposed reaction mechanism.
Annex 2, and Smejkal’s group contribution method, which Once a model that is robust enough is developed, it could be
is based in Joback’s group contribution method and shows adapted towards another and more sophisticated one, such as
excellent accuracy [17, 38]. that of Kubicka et al. As for Landberg’s study, it was discarded
as it does not explain the formation of CO2 since it does not
Kinetics take into account the decarboxylation reaction in its proposed
Finding proper kinetic data proved to be harder than ex- reaction pathway.
pected, given the presence of complex triglyceride molecules
in the reactions. Furthermore, since the reactions take place in
the surface of catalysts, kinetic parameters depend not only in
reactor conditions (temperature, mainly) but also in the catalyst
system being used.
The team decided that, in order to select appropriate kinetic
data among the few options found in the literature, some restric-
tions had to be put in place:
1. Kinetic data must be calculated in a sulphided NiMo, CoMo
catalyst, in order to match catalysts typically used in the
industry, as discussed earlier.

2. Reaction pathways for the proposed mechanism must ac-

count for the formation of gases typically formed in the
HDO and DCO mechanisms that take place in the industrial
systems mentioned above (CO, CO2 , propane and H2 O).

3. Kinetic data must include Arrhenius’s constant, so that

temperature dependence can be included in the model.

4. If possible, kinetic data should also include the effect of hy-

drogen pressure in their equations, in order to understand
the effect of hydrogen presence in the reactor.
Table 4 Proposed mechanism and kinetic data for hydrotreat-
In light of these requirements, the studies of Vélez [18], Land-
ment of palm oil [18].
berg [19], Kubicka [26] and Zhang et al.[39] all fit most of them.
One of the common and advantageous aspects of all these mech-
anisms is that they are based on simple power kinetic models, Reactor type
given the incompatibility between more complicated models Reactors used in this process are most likely the same type of
such as Langmuir-Hinshelwood’s and uncertainties linked to reactors as the ones used in hydrotreating desulphuration units,
operation parameters in trickle-bed reactor systems [40]. this is, trickle-bed reactors.
As a starting point, Vélez’s kinetic studies provide a model Trickle-bed reactors are excellent choices when it comes to
which is simple enough to develop a simulation model that heterogeneous reactions where liquid and gas phases (mainly

A proposal for the HVO process 5

hydrogen) come into contact over a catalyst bed at high temper- On the other hand, taking a look at the long term outlook
atures and pressures [41]. of the oil market, supply is ensured due to new deposits be-
When compared to other heterogeneous reactors, trickle-beds ing continuously discovered. Regarding the demand for oil
offer poor heat transfer rates, thus leading to radial temperature and its future prices, experts do not expect the market share to
gradients and the formation of hotspots. In order to prevent increase, as environmental restrictions in developed countries
this situation, the reactor may count with interstage quenching grow tougher [53].
systems, cooling jackets or, in more extreme cases, be designed One such example of these restrictions is the cap for sulfur
as a multi-tubular reactor. Jeczmionek et al.[42] presents a useful content in marine fuel, which is being lowered from 3.5 to 0.5%
study regarding the heat of reactions involved in hydrotreatment by the International Maritime Organization (IMO) [54]. The
of vegetable oils. measure, which will be effective on the 1st of January of 2020,
Another important aspect of trickle-bed reactor systems are has brought many discussions to the table as to how to achieve
its hydrodynamics. Ranade and Chaudhari [43] present in their it in the most cost-efficient manner.
book a complete overview of the main parameters affecting the As a possible solution, marine vessels could move towards
hydrodynamic effects in trickle-bed reactors. cleaner versions of marine fuel such as low-sulfur diesel blends.
For the estimation of the liquid hold-up in high pressure In this aspect, green diesel seems like an excellent choice, as its
systems, Larachi et al.[44] provides simple correlations based content in sulfur is exceptionally low [4].
in dimensionless numbers, while the wetting efficiency of the The measure, which affects the demand of 3.5 million barrels
catalyst particles can be estimated with the Reynolds number, per day of high-sulfur fuel oil, will likely cause an increase in
according to Giannetto and Specchia [45]. diesel prices in the future. As of 2018, the mean price of diesel
Trickle-bed models were thoroughly reviewed in order to in Spain is 1.1 e/kg [55].
develop a reactor model robust and complex enough to predict
with fair accuracy the operation in a hydrotreatment plant for Process capacity and feedstock availability
vegetable oils [46-48]. A detailed explanation of the equations After studying feedstock prices for vegetable oils (Figure
used in the model can be found in Annex 4. 7), one may notice there is a short gap between retail price for
biodiesel and feedstock price. This is also typical for fossil fuel
Isomerization refining products, since the fuels market is close to perfectly
Isomerization takes place once gaseous products and water competitive.
formed during deoxygenation are separated. As discussed ear-
lier, noble metals such as Pt over SAPO zeolite supports are
typically used in this operation [29].
Temperatures and pressures are similar to those in deoxy-
genation, and even if hydrogen is consumed in low quantities,
the reaction can only take place in a high pressure hydrogen
atmosphere [49].
For simulation, typical conversions found in the literature
[50] can be used together with Calemma’s proposed pathway
and kinetic information for isomerization of long hydrocarbon
chains [51]. Figure 7 Biodiesel (FAME) differentials in $/ton [57].

Market analysis The Finnish company Neste oil, which provides valuable
information on the fuel market distribution, proposes a method
to fix the price of biodiesel based on margins and production
costs which will be used and explained more thoroughly in the
Economic Analysis section.
As a result, in order to ensure profitability of the venture,
process capacity has to be carefully measured, following the
economy of scale principle.
Main players in the HVO game for stand-alone units are,
as mentioned earlier, Neste Oil and UOP/ENI. Neste owns 4
production facilities, two in Finland with a capacity of nearly
200,000 tons per year and another two with a capacity of 1000,000
tons per year in Rotterdam and Singapore. As for ENI, they are
Figure 6 Oil barrel prices ($/bbl) in the last 4 years [57]. currently constructing a plant with a capacity of 500,000 tons in
Italy [56].
Crude oil prices are the most determining factor for fuel cost. In view of this data, production capacity for the plant is set
As seen in Figure 6, the price of the barrel has been volatile in to be 500,000 tons of green diesel per year, which implies a
the last few months, mainly due to the uncertainty in global consumption of around 600,000 tons of vegetable oil per year
economy created by the trade war between China and the U.S. (roughly 20% is converted into gas and cracking products, ac-
and political problems in Europe [52]. cording to mass balances).

6 Javier Moreno and Aurelio Hurtado de Mendoza

 Off gas Green LPG
16
22
Pre- C-101 26 Hydrogen 24
Palm oil 1 treatment 23 E-104
unit P-101
Green V-103
2 Naphtha
25 Amine unit T-101
P-101 P-102
Recycled hydrogen
F-101 3
4 27 17

21 Green
Naphtha
20
28
R-101 T-102

8 R-102 E-105
13
Green
29
15 Kerosene
10 V-102
18

5
V-101

6
14
19
E-101 7
12

E-106 Green
Water Diesel
11 E-103
9 E-102

Figure 8 Process flow diagram (PFD) of the proposal for the hydrotreatment of vegetable oils. See corresponding annex for more
details on stream characteristics.

As discussed in the Literature Review, only the annual pro- Process Design
duction of palm oil is high enough to ensure feedstock availabil-
The information found and portrayed in the Literature Review
ity.
section allowed the team to perform a conceptual design of
Furthermore, as seen in Figure 9, palm oil prices are consid-
the process, the result of which is described in the following
erably lower than rapeseed and sunflower oil prices, the latter
paragraphs and depicted in Figure 8. A summary table for the
of which is not depicted but data proves its price lands between
input and output stream conditions is shown in Table 5. A more
the two [58].
comprehensive table for all process streams can be found in
Annex 6.
The process can be divided into three main sections: hy-
drotreatment, isomerization, and final product separation. First,
palm oil is purified in a pretreatment plant, where it undergoes
degumming and bleaching in order to remove metals, phos-
phorous and any solid particles that might affect the catalyst
negatively.
After purification, palm oil enters reactor R-101, which is
a trickle bed reactor with catalyst beds disposed in series for
intermediate quenching with hydrogen. The catalyst is made of
Figure 9 Vegetable oil prices ($/ton) for rapeseed (above) and NiMo supported on alumina and conversion is nearly total.
palm oil (below) in the last 4 years [57]. Liquid products obtained in the hydrotreatment reactor are
mainly n-alkanes in the diesel boiling range (270-320◦ C), result-
The main drivers for choosing palm oil as feedstock in this ing in a product with high cetane number but poor cold-flow
project are price and availability. properties. In order to improve the quality of the product, it is
However, the Renewable Energy Directive from the EU de- essential for the mixture to enter a second reactor.
mands that GHG emissions be reduced by 60% or more in order Reactor R-102 operates similarly to the hydrotreatment reac-
to be considered as a biofuel [3]. tor, this is, at high temperatures and pressures and with consid-
Oil companies such as Neste and Repsol count with special erable supply of hydrogen gas. In this case, catalyst bed is made
certifications known as the International Sustainability and Car- of Pt supported in a SAPO zeolite and alumina. Any triglyc-
bon Certifications (ISCC) that ensure biofuel lifecycle from cradle erides that have not reacted in R-101 will thus finish reacting
to cradle is sustainable. One representative example depicting here.
control over GHG emissions in the transport of palm oil to the Improving the cold-flow properties of the product comes at
refinery is shown in Annex 3. the expense of inevitable cracking of the diesel, which yields

A proposal for the HVO process 7

Stream Number 2 26 7 17 18 19 22 23 24
From (feed) (feed) V-101 P-102 E-105 E-106 Amine unit Amine unit Amine unit
To F-101 C-101 (out) (out) (out) (out) (out) (out) (out)
Temperature (ºC) 60 50 30 26 26 26 30 30 30
Pressure (bar) 1,0 20,0 40,0 1,0 1,0 1,0 1,0 1,0 1,0
Enthalpy flow (MW) -48,92 0,27 -38,70 -0,51 -1,13 -32,71 -3,23 -0,25 -1,76
Mole flow (kmol/h) 87 1296 488 8 11 243 110 5 128
Mass flow (kg/h) 73666 2613 8795 767 1993 58522 4943 377 883
Volume flow (m3/h) 83,50 1760,64 8,82 1,13 2,71 82,58 2741,99 0,60 3201,39
Mass Fractions:
Tripalmitin 41,97%
Tristearin 3,75%
Triolein 46,28%
Trilinolein 8,00%
Propane 2,47% 93,47%
C4H10 2,48% 6,53%
C5H12 13,91% 84,20%
C6H14 1,56% 2,28%
C7H16 26,16% 11,19%
C8H18 2,25% 0,29%
C9H20 50,06% 2,05%
C12H26 4,93% 0,00%
C13H28 71,67% 0,00%
C14H30 4,06% 0,04%
C15H32 1,75% 2,52%
C16H34 0,27% 9,15%
C17H36 0,00% 0,65%
C18H38 0,00% 13,00%
I-C15 9,73% 2,42%
I-C16 7,51% 28,89%
I-C17 0,01% 2,05%
I-C18 0,06% 41,29%
Hydrogen 100,00% 25,07%
Carbon dioxyde 64,77%
Water 100,00% 1,10% 10,14%

Table 5 Input (in green) and output (in blue) stream conditions for the proposed Process Flow Diagram.

variable quantities of lower boiling point mixtures such as jet Process simulation and analysis
fuel and naphtha, depending on reactor conditions.
The proposed solution for the process was simulated using As-
The resulting mixtures are separated in a distillation column pen Plus® V10. Considering that reactor simulation options
where naphtha and gaseous products are obtained as top prod- offered in this software are fairly limited, it was decided that
ucts, diesel as bottom product, and jet fuel is obtained through the hydrotreatment reactor would be simulated using its pro-
a side column. This way, the process is flexible enough to yield gramming modeling environment, known as Aspen Custom
variable quantities of products different from green diesel, de- Modeler® (see support files).
pending on the market demand. Explanations for the developed model of the hydrotreatment
reactor, which include mass transfer resistance calculations, are
Furthermore, since hydrogen is present in great quantities explained in detail in Annex 4. From the simulations, it was
in both reactors but is not totally consumed, an amine unit was found that reactor dimensions are 25 meters in length and 5 me-
considered as the best option to recover and recycle most of it, as ters in diameter. Conversion attains 99.9%, at nominal operation
well as the light hydrocarbons and contaminants (CO and CO2 , pressure (40 bar), slightly varying upon changing the operation
mainly) present in the gaseous flows which are obtained in the conditions.
flash tanks situated after each reactor.
Regarding the isomerization reactor, only reaction kinetics
The team did not go deeper into the design of the amine was modeled using the RPlug block and the proposed reaction
unit, as it was assumed the process would be integrated into pathway and kinetic constants by Calemma et al.[51]. A brief
an existing refinery. However, one interesting option which explanation of this model is shown in Annex 5.
deserves further investigation would be to study the feasibility Basic calculations were made using typical LSHV values
of transforming the light hydrocarbons into hydrogen in order found in the literature for this process [50] and process data.
to reduce external dependence. Results provide an estimated reactor size of 18 meters in length

8 Javier Moreno and Aurelio Hurtado de Mendoza

 Unit Process conditions Comments
P-101 60◦ C, 1 atm Below 60◦ C palm oil is semi-solid
C-101 360◦ C, 40 bar Interstage cooling
R-101 300-360◦ C, 40-60 bar, 10-100 h−1 (WSHV), H2 /oil ratio 500 NiMo/Al2 O3 catalyst
Nm3 /m3
R-102 360-380◦ C, 40 bar, 1 h−1 , H2 /oil ratio 500 Nm3 /m3 Pt/SAPO-11/Al2 O3 catalyst
V-101 30◦ C, 40 bar Flash tank
V-102 30◦ C, 15 bar Flash tank
T-101 26-300◦ C, 1 bar Main column, 28 plates
T-102 215-240◦ C, 1 bar Side column, 10 plates

Table 6 Operating conditions for main process equipment.

and 4 meters in diameter, with which a high quality diesel prod- 1 0,007
uct with excellent cold-flow properties is obtained.

Concentration of H2 in bulk liqud (kmol/m3)

0,95 0,006
Simulation was useful when testing the fitness of the con-
ditions and data found in the literature, as well as to system- 0,9
0,005

Tripalmitin conversion
atically find and exploit possibilities for process optimization
0,85
and flexibility (i.e. most important contributors to cost and heat 0,004
integration options). Such possibilities are analyzed further in 0,8
this chapter. 0,003
0,75
Final selected conditions for the main process equipment are
0,002
shown in Table 6, while the mass and energy balance for all 0,7
process flows is shown in Annex 6. 0,001
0,65

Process flexibility 0,6 0

36 38 40 42 44 46 48 50
One of the main advantages of simulation is the ability to test
Operating pressure (bar)
the flexibility of the process. This allows to perform a sensitivity
analysis and study the effect key design parameters have on the
quality and quantity of the product obtained. Figure 10 Tripalmitin degree of conversion as a function of op-
Furthermore, evaluating process performance at different erating pressure. In this case, the pressure threshold is placed
operating conditions proves the validity of the models used for at 44 bar.
simulation. In this case, flexibility of the process was studied for
its main operations: hydrotreatment and isomerization. in Figure 10). If pressure becomes lower than this thresh-
The use of a custom model for the hydrotreating reactor has old, Tripalmitin conversion will steeply descend, although
allowed for interesting analysis of the influence of operation and Triolein conversion is not as significantly affected.
feed conditions on overall reactant conversion. Triolein and Tri- • The temperature profile is also important for reactants’ con-
palmitin being the main reactants in terms of feed composition, version. The fast Triolein consumption in the first half of
the first one is consumed in the first half of the reactor length, the reactor causes a great increase in temperature, which
while Tripalmitin and Tristearin (generated from hydrogenated must be kept under 360◦ C with refrigeration. Interestingly,
Triolein) reactions suffer from competitive inhibition and occur Tripalmitin and Tristearin consumption rates are enhanced
only in the second half of the reactor. by temperatures close to 300◦ C instead of higher ones. This
Two variables were found to be critical for reactant conver- is due to the higher volatility of hydrogen at higher tem-
sion: concentration of H2 in the bulk liquid and temperature peratures, which decreases the amount of H2 in the liquid,
profile along the reactor. and to a worsening of the liquid-solid mass transfer coef-
ficients also at higher temperatures. Temperature profiles
• Concentration of H2 in the bulk liquid is rather sensitive to
should be kept as flat as possible, with higher values for
pressure variations. A plot of Tripalmitin degree of conver-
Triolein reactions and lower values for Tripalmitin reaction.
sion - the others being higher - against operating pressure
Consequently, hydrogen quenching arises as an interesting
can be seen in Figure 10. Note that other operating condi-
possibility for refrigeration, something that has nevertheless
tions, such as inlet temperature or cooling power, remain
not been explored in this work.
constant for this plot, but can be varied to operate at 40 bar
with 99.9% degrees of conversion. A pressure threshold The reactor performance is not significantly affected by
has been observed which varies between 38 and 48 bar, de- changes in feed composition or feed temperature. In general, it
pending on other operating conditions (it is placed at 44 bar can be said that operation of the hydrogenation reactor is flexible

A proposal for the HVO process 9

and that total conversion (> 99.9%) can be achieved in a wide
range of situations.
Another interesting study was made regarding isomerization
reactor R-102. Operating conditions that affect the degree of
isomerization and cracking are: temperature, pressure, LSHV
values and H2 /oil ratios.
While the effect of these parameters is studied elsewhere in
the literature [50-51, 59-60], the team found most interesting
to study the effect of temperature of the reactor. Reactor tem-
perature was varied between 360 and 450◦ C and the yield of
products was evaluated. Of course, this analysis also allowed to
understand more deeply which variables had to be manipulated
in the distillation columns to adapt to this variation.
As shown in Figure 11, increasing the temperature in R-102
increases the ratio of light to heavy products to the point where
at 450◦ Cconsidered. The composite
no diesel is obtained curve
as a product. depicting
Another heat
interesting the HEN proposal depicted in Figure 11.
flows
finding was from
that the hot andofcold
the proportion currents
jet fuel is reduced
is in fact depictedfrom
in
◦ C , due to the kinetics of the model.
Figure
420 to 450 10. To keep
Figure a balance,
12 Composite one for
curves possibility would
hot and cold beconsid-
flows to
remove
ered in the hydrogen in aAshigh
pinch analysis. it can temperature
be seen, demandflash
for cold
vessel isafter
utilities high each
given reactor, reducing
the misbalance the power
between to
the currents.
temperature ratio and thus allowing the pinch
temperature to decrease before heat exchange.
The pinch temperature is the highest in the process: 360◦ C,
anThe
interesting
gaseousresult which
currents can be explained
containing mainlybyhydrogen
the high power
to temperature (kW/◦ C) ratio of the hot currents coming from
could then be cooled using air coolers since amine
the reactors compared to the cold currents. This is a result of the
unitsamounts
large operateofathydrogen
30-40 ºC.these
Another option would
hot currents carry. be
toAsdesign different unit for hydrogen purification
a result, the calculated heat exchanger design networks
(HEN) do not offeratmany
that can operate higher temperatures,
possibilities, givensuch
the as a
misbalance
pressure swing absorption unit (PSA).
between hot and cold currents and the excess heat being wasted
in the process. Table 7 shows the potential savings from the
Tableproposal
HEN 6: potential savings
depicted from proposed
in Figure 13. integrated heat
exchanger networks over a plant life of 15 years.

Scenario Base Integrated Integrated

Figure 10: composite curves for hot and cold flows considered case process process
in the pinch analysis. As it can be seen, demand for cold (unbalanced) (balanced)
utilities is high given the misbalance between the currents. Heating (MW) 35,9 1,8 3,7

The pinch temperature is the highest in the Cooling (MW) 55,7 145,3 64,1
Figure 11 Revenues from LPG, naphtha, jet fuel and diesel Operating
process: 360 ºC, an interesting result which can be cost (M€/yr)
3,35 1,12 0,92
obtained at different temperatures in the isomerization (R-102)
reactor. explained
Total revenuebyis shown
the high power
as a line to temperature
above. Capital cost
1,84 2,89 3,39
(kW/ºC) ratio of the hot currents coming from the (M€)
Total cost
3,6 1,49 1,82
Fromreactors compared
this sensitivity to the
analysis cold
it can be currents.
concludedThis
that, is a
with (M€/yr)
result of the large amounts of hydrogen these
current market prices, the process becomes less profitable with hot Savings
- 2,11 1,78
(M€/yr)
currents
increasing carry. of the reactor R-102.
temperature
Steam reforming of propane
Table 7 Potential saving from proposed integrated heat ex-
HeatAs a result, the calculated heat exchanger design
integration
From the
changer process
networks overflows
a plantitlife
canof 15
be years.
seen propane
networks (HEN)
Considerations for heat do not offer
integration weremany
made possibilities,
according to
could be utilized for hydrogen generation through
the pinchgiven theresults.
analysis misbalance between an
Aspen Plus®offers hotEnergy
and Analy-
cold
sis package that is suitable for this operation. steam reforming:
To keep a balance, one possibility would be to remove hy-
currents and the excess heat being wasted in the
In order to perform the pinch analysis, only the main process drogen in a high temperature flash vessel after each reactor,
process. Table 6 shows the potential savings from !"power
#$ + 6#
streams were included in the study, all of which flow through a reducing the ' ( ⟶ 3!('ratio
to temperature + 10 #
and 'thus allowing the
heat exchanger in Figure 8. Thermal fluids for reactor cooling pinch temperature to decrease before heat exchange.
were not considered as options for heat integration. The gaseous currents containing mainly hydrogen could then
As for the utilities, cooling water, medium and high pressure be cooled using air coolers since amine units operate at 30-40◦ C.
steam and fired heat were considered. The composite curve Another option would be to design a different unit for hydrogen
depicting heat flows from the hot and cold currents is depicted purification that can operate at higher temperatures, such as a
in Figure 12. pressure swing absorption unit (PSA).

10 Javier Moreno and Aurelio Hurtado de Mendoza

 Since 105 kmol per hour of propane are obtained the to
and the required input for hydrogen is 1296 kmol hydrotr
per hour, propane could cover 80 % of the perspec
hydrogen consumption, while the rest could be
obtained from methane reforming, thus reducing Investm
its negative greenhouse impact. €550 m
investe
ECONOMIC ANALYSIS 60 % m

The economic evaluation of the project starts by Produc

calculating the capital cost of the plant and the
variable production costs using Towler and Variab
Figure 13 Proposed HEN network for the process. From top to bottom, process Sinnot’s
flowsmethod [61].water,
are: cooling Product
streamprice is then
11, stream 5,
consum
stream 19, stream 18, stream 9, stream 2, boiler from column 1, boiler from calculated according
column 2, stream to heat,
14, fired the high
margin
pressuremethod
steam stream
and medium pressure steam. Minimum temperature difference is 10◦ C (9,7proposed
◦ C in yellow exchanger)
by Neste oil, where margins are applied waste
in order to obtain a reasonable benefit. Plus® a
which
Propane reforming network was already obtained in the heat integration section.
Payback period and net present value are also values
From stream 22 in the Process Flow Diagram (8) it can be seen Costs must be updated using CEPCI index for 2016, which is
propane could be utilized for hydrogen generation through 541,7. calculated, and sensitivity analysis is performed in
Coolin
steam reforming: ISBLorder to calculated
costs are check the as thedependence
sum of equipment of project
costs mul-
profitability on the main contributors to cost. Aspen
tiplied by Lang installation factors, while OSBL, engineering
and contingency costs are calculated as percentages of the inside is assu
C3 H8 + 6 H2 O −→ 3 CO2 + 10 H2 (1)
batteryCapital investment
limits cost (30, 30 and 10%, respectively). hydrog
Since 105 kmol per hour of propane are obtained and the estimat
Finally, working capital is estimated as 15% of the fixed capi-
required input for hydrogen is 1296 kmol per hour, propane
tal investment in order to obtain the total required investment. also ne
could cover 80% of the hydrogen consumption, while the rest
Tables Table
8 and7:9 main
sumprocess
up these results.costs (installed).
equipment
could be obtained from methane reforming, thus reducing its The pr
negative greenhouse impact.
Equipment Cost (M€) unit w
Reactor R-101 91,3 times
Economic analysis Vessel V-101 0,2
handlin
Reactor R-102 39,8
The economic evaluation of the project starts by calculating
Vessel V-102 0,3
the capital cost of the plant and the variable production costs
Tower T-101 0,2 Catalys
using Towler and Sinnot’s method [61]. Product price is then Tower T-102 0,04 Mo an
calculated according to the margin method proposed by Neste Compressor C-101 9,4 their p
oil, where margins are applied in order to obtain a reasonable Heat exchangers 10,8 assume
benefit. Pretreatment unit 10
hydrod
Payback period and net present value are also calculated, Total 162,0
and sensitivity analysis is performed in order to check the de-
Regard
pendence of project profitability on the main contributors to Table 8 Main process equipment costs (installed).
Table 8: total capital investment of the venture. 0,7 €/k
cost.
As expected, biggest contributors to capital investment are values
Capital investment Investment Cost (M€) price,
reactor costs. As much as 29% of the total investment is coming
Sizes and capacities for the main process equipment are obtained ISBL
from the hydrotreatment reactor, a value which162,0
gives perspective assumi
through simulation. Costs are then calculated using curves for OSBL
on the importance of its design. 48,6
Table 8:
each process equipment from Turton [62] and Rules of Thumb in InvestmentEngineering line with the e550
of the plant seems to be in 48,6
Engineering Practice [63], while the cost for the heat exchanger million which Neste Oil is estimated to have 16,2
Contingency invested in its plant
Raw material
Fixed cap. Inv. 275,4
Working cap. 41,3 process
A proposal for the HVO 11 Hydrogen
Total investment 316,7 Palm oil

Utilities
 hydrodesulfuration catalyst lifecycles [67].
Total 162,0
Regarding feedstock, palm oil price is selected as
Table 8: total capital investment of the venture. 0,7 €/kg a base case scenario, based on the recent
values shown in the Market Analysis. Hydrogen
Investment Cost (M€) price, on the other hand, is fixed at 1 €/kg,
ISBL 162,0 assuming it can be generated within the plant.
OSBL 48,6
Table 8: summary review of the variable cost of production
Engineering 48,6
Contingency 16,2 Catalyst prices were estimated based on their Ni, study [68]
Cost
Raw materials Consumption
Mo (t/yr) asCost
and Pt content (€/kg)in the literature and
found accounted f
Fixed cap. Inv. 275,4 (M€/yr)
Working cap. 41,3 Hydrogen
their 20901,5
price [21, 50]. Lifecycle
1
of the catalyst
20,9
was
assumed as 1,5 years, based on typical Given the v
Total investment 316,7 Palm oil 589331 0,7 412,5 weight on
hydrodesulfuration catalyst lifecycles [67].
Consumption Cost recommend
Utilities Cost (€/MWh)
(MW) (M€/yr)
Table 9 Total capital investment of the venture. Regarding feedstock, palm oil price is selected as studies in
As expected, biggest contributors to capital Cooling 145,3 0,08 0,089
0,7 €/kg a base case scenario, based on the recent adapting th
investment are reactor costs. As much as 29 % of values shown in the Market Analysis. Hydrogen waste oils
in Singapore, which produces 60% more product [64].
price, on the other hand, is fixed at 1 €/kg, and animal
assuming it can be generated within the plant.
Production costs
Figure 14 Operating costs of the plant.
Variable costs come from raw materials, utilities consumption, Table 8: summary review of the variable cost of production
fungibles (catalysts) as well as waste stream treatment. Refer-
Consumption Cost
ence costs for utilities and waste water treatment were taken Raw materials Cost (€/kg)
(t/yr) (M€/yr)
from Aspen Plus® and Turton except for the price of electric- Hydrogen 20901,5 1 20,9
ity, which was fixed at 83 e/MWh, according to the values for Palm oil 589331 0,7 412,5
industrial consumption in Spain [65]. Consumption Cost
Utilities Cost (€/MWh)
Cooling duty for the reactors was calculated using Aspen (MW) (M€/yr)
Plus®, according to the heat of reactions. It is assumed that Cooling 145,3 0,08 0,089
part of this heat is absorbed by the hydrogen gas (quenching), Heating 1,8 1,5 0,023
although half of it is estimated to be removed by a thermal fluid, Electricity 31,4 83 20,85
which also needs to be considered. Thermal fluid 11,9 2 0,19
The price of hydrogen purification in an amine unit was
Gas, water and Cost
estimated as 0,1e per treated m3 , three times its typical value in Flow (Mm3/yr) Cost (€/m3)
palm oil treatment (M€/yr)
order to account for handling expenses [66].
Catalyst prices were estimated based on their Ni, Mo and Water 0,7 0,043 0,03
Hydrogen Product pr
Pt content as found in the literature and their price [21, 50]. 44,9 0,1 4,49
purification
Lifecycle of the catalyst was assumed as 1,5 years, based on Palm oil Product pr
0,7 0,5 0,37
typical hydrodesulfuration catalyst lifecycles [67]. purification
Consumption Cost adding mar
Regarding feedstock, palm oil price is selected as 0,7 e/kg Fungibles
(t/yr)
Cost (€/kg)
(M€/yr) the result w
a base case scenario, based on the recent values shown in the NiMo/Al2O3 81430,3 4,2 0,65 to ensure co
Market Analysis. Hydrogen price, on the other hand, is fixed at
Pt/SAPO-11 72382,464 139,5 10,10
1 e/kg, assuming it can be generated within the plant. Plant produ
Total variable costs 470,2
Fixed costs are calculated assuming the plant is two oper- year of gre
ators are present at all times and a base salary of e50,000 per fuel (1590
year. The rest of the costs can be calculated as percentages of Table 10 Summary review of the variable cost of production.
Fixed costs are calculated assuming the plant is green LPG
labor and investment (ISBL and OSBL) costs, including: supervi-
two operators are present at all times and a base
sion, maintenance, property taxes, plant overhead, royalty fees market values for€50,000
biodiesel Table 9 sh
salary of pertoyear.
ensure
Thecompetitiveness
rest of the costsof the
and depreciation. For the latter, 5% of fixed capital cost was product.can be calculated as percentages of labor and diesel pro
considered, assuming a plant life of 20 years. Plantinvestment
production(ISBL is roughly market val
and 468.000
OSBL) metric
costs, tons per year of
including:
Calculations for all costs as well as estimation details can green diesel. prices are:
be found in Annex 7. The proportion of production costs is supervision, maintenance, property taxes, plant t/yr),
Byproducts include: green jet fuel (15.900
green naphtha (9.100 t/yr) fees
and green LPG (39.500 For t/yr). naphtha, an
depicted in Figure 14. The results confirm what was found in overhead, royalty and depreciation. the
Tablelatter,
11 shows the calculated margin for green diesel pro- Spanish ma
another similar study [68], where the cost of the raw material 5 % of fixed capital cost was considered,
duced in the process, considering a market value of 1,1 e/kg.
accounted for 75% of the production costs. assuming a plant life of 20 years. The values
The rest of the product prices are: 1e/kg for kerosene, 1,15 e/kg
Given the volatility of the price for palm oil and its weight on for naphtha, and 0,54for e/kg shown by
Calculations all for LPG,
costs as according to the Spanish
well as estimation
the production costs, it is highly recommended in view of these market [55]. The values obtained in Table 9 are similar to margin est
details can be found in Annex 7. The proportion of those
results that future studies in the field focus on the possibility of shown by for Neste O
adapting the process to less clean (but cheaper) waste oils such
production costs is depicted in Figure 12. The esti-
Neste Oil for their Renewable product margin
mation, although final margin values forinNeste Oilsimilar
are twice as to tax exem
as cooking oil or oil from the fish and animal industry. results confirm what was found another
high [69], probably due to tax exemptions and the high quality which allow
of the diesel which allows it to be sold at higher prices.
Product price and profitability To calculate profitability of the plant, annual cash flow is de-
Product price and revenue can be calculated by adding mar- termined considering the EBIT value obtained after subtracting
gins to production costs and comparing the result with current production costs to revenues and a taxation of 15% over profits,

12 Javier Moreno and Aurelio Hurtado de Mendoza

 800
600
600 over the plant’s lifespan of 20 years (5 %), as
400 shown in the fixed costs of production. A realistic

Cash flow (M€)

400 approach was used where the plant undergoes
200 technical shutdowns once every two years,
reducing its productivity to 90 %. Furthermore,
NPV (M€)

200
0 productivity on the first year was assumed to be 75
0,06 0,07 0,080 0,09 0,1 0,11 0,12 0,13 0,14 % due to technical difficulties in start-up
-200 operations.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
-200
-400 Annual cash flow is shown in Figure 10, where it
canYears
be noticed that the slope of the curve varies
-600 -400
every two years due to the above-mentioned
schedules technical shutdowns. Also, construction
Interest rate
of the plant
Figure 15 Cash flow over the lifespan of the plant and its construction, in is
Me.2 years.
Table 9: product margin for a market value of green diesel of
1,1 €/kg.
As seen in Figure 9 (for which the values are
Green Diesel shown in Annex 7), payback period is 7 years.
600
Production (t/yr) 468175
Cost of palm oil (€/t) 700
In order to determine the economic interest of the
400 value has been used. The net
project, net present
Variable cost of production (€/t) 121,8
Fixed cost of prod (€/t) 70,1 present value or NPV can be calculated from the
cash flow in each 200year and the discount rate, which
Margin (€/t) 208,1

NPV (M€)
Final product cost (€/t) 1100
is around 10 %. However, since the interest rate
0
depends on several factors and
Considering palmis hard to predict
oil and diesel atprices are of kerosene, nap
this stage, the net present
decisive value
0,07has0,08
importance
0,06 been calculated
for 0,09
the profitability
0,1 0,11 0,12 of the
0,13 0,14 simulated, and
Table 11 Product margin for a market value of green diesel of -200
To calculate profitability of the plant, annual cash venture,
at different interest ratesavarying
sensitivity
fromanalysis
6 to 14 %. was performed to
1,1 e/kg. The team foun
flow is determined considering the EBIT value study the effect of their fluctuation on the
As seen in Figure-40010, negative NPV values are not design of the h
obtained after subtracting production costs to feasibility of this project. Palm oil prices are varied
positive except for low values, a consequence of
between 900 and 500 €/ton, while diesel prices are ensure total
due to revenues from greenvalue
the environmental dieselofand
thea project.
taxation of 15 % -600 cost the project requires.
the high investment importance is
over profits,
Depreciation wasdueconsidered
to the environmental
as a linearvalue of theover the
function varied between 0,9 and 1,3 €/kg.
through simula
plant’sproject.
lifespan of 20 years (5%), as shown in the fixed costs Interest rate
of production. A realistic approach was used where the plant Table 10 represents the payback period and NPV
Moreover, th
Depreciation was considered as a linear function
undergoes technical shutdowns once every two years, reducing Figure for bothPresent
16 Net analyses.Values
As seen from therates
at interest results,
from these
6% to 14%.shifting the p
its productivity to 90%. Furthermore, productivity on the first variations have great power over the results, thus
allowing to ad
year was assumed to be 75% due to technical difficulties in start- increasing the risk of the investment.
800
up operations. analyses. As seen from the results, these variations have great Regarding he
Table 10: sensitivity analysis results on profitability for diesel
Annual cash flow is shown in Figure 15, where it can be effect over the results, both in a positive and in a negative way. imbalance bet
and palm oil prices.
noticed that the600 slope of the curve varies every two years due process, mak
to the above-mentioned scheduled technical shutdowns. Also, Diesel price (€/kg) Payback period (yr) NPV (M€) resulting in a
construction of the plant is 2 years. 0,9 Negative cash flow - investigations
400
As seen in Figure 15 (for which the values are shown in Annex 1,3 3 4149
Cash flow (M€)

Another poss
7), payback period is 7 years. In order to determine the economic Palm oil price (€/t) Payback period (yr) NPV (M€)
200 net present value has been used. The net propane steam
interest of the project, 500 3 5330
calculations sh
present value or NPV can be calculated from the cash flow in 900 Negative cash flow -
consumption r
each year and the discount rate, which is around 10%. However,
0
since the interest rate depends on several factors and is hard to Sensitivity results on profitability for dieselFinally, the ec
0 1 2 3 4 5 6 7 8 9 10 11 12 13 Table 14 151216 17 18 19analysis
20 21 22
predict at this stage, the net present value has been calculated at and palmCONCLUSIONS
oil prices. Figure process after t
different interest-200
rates varying from 6 to 14%. 10:
oil fluctuation
As seen in Figure 16, negative NPV values are not positive The EU’s Renewable Energy net Directive RED is
Years presen to improve
except for low values, a consequence of the high investment cost imposing ambitious improvements in the energy
t governmental
-400
the project requires. Conclusions
market in order to shift Europe’s energy mix
Considering palm oil and diesel prices are of decisive impor- The EU’stowards a more sustainable
Renewable selection of
Energy Directive REDfuels.is imposing am-
tance for the profitability of the venture, a sensitivity analysis bitiousInimprovements in thefrom
energy marketoils
in are
order
this regard, biofuels vegetable a to shift
was performed to study the effect
Figure of their
11: Cash fluctuation
flow over the lifespan on the
of the Europe’s
plant and its energy
construction in mixoftowards
millions Euros a more sustainable selection of
valid option to allow the transition in the energy
feasibility of this project. Palm oil prices are varied between 900 fuels.
market, although regulations and process design ACKNOWLE
and 500 e/ton, while diesel prices are varied between 0,9 and In this regard, biofuels from vegetable oilsfeedstock
are a valid option
should consider the sustainability of the
1,3 e/kg. to allow the transition in the energy market, although regula- The team wo
very closely for both plantation and transport.
Table 12 represents the payback period and NPV for both tions and process design should consider the sustainability Galán of for his
Furthermore, the high effect feedstock price has project as we
over process profitability should invite designers Also, we than
A proposal for the HVO process 13
to consider alternative sources of oils such as valuable in
waste products from the fish and animal industries regulations for
or used cooking oil.
the feedstock very closely for both plantation and transport. relevant refining palm oil. Fuel Processing Technology 116 (2013), pp.
Furthermore, the high effect feedstock price has over pro- 16-26.
[9] Neste Oil, 2012. NExBTL® Renewable Diesel Singapore Plant.
cess profitability should invite designers to consider alternative
Tallow pathway description.
sources of oils such as waste products from the fish and animal [10] María Cecilia Vásquez, Electo Eduardo Silva, Edgar Fernando
industries or used cooking oil. Castillo. Hydrotreatment of vegetable oils: A review of the technologies
When hydrotreated and isomerized, vegetable oils not only and its developments for jet biofuel production. Biomass and Bioenergy
yield a cleaner, high quality product, but also present improved 105 (2017) pp. 197-206.
[11] Food and Agriculture Organization of the United Nations. Food
cold flow properties, making it an attractive product in northern
Outlook, July 2018.
European countries. [12] FAO, www.fao.org, faostat/crops processed. Information re-
In this design, a plant with a capacity of roughly 500,000 tons trieved in December 2018.
per year of green fuel (diesel, kerosene, naphtha and LPG) has [13] R.E.A. Holdings PLC. www.rea.co.uk/markets. Information
been proposed, simulated, and its feasibility evaluated. retrieved in December 2018.
[14] Graboski MS, McCormick RL. Prog Energ Combust 24 (1998),
The team found that, main aspects considered, the design of
pp. 125-64.
the hydrotreatment reactor is essential to ensure total conver- [15] Pryde EH. Northern Regional Research Center, US Department
sions and profitability. This importance is corresponded with of Agriculture, Peoria, Illinois, The American Oil Chemists’ Society,
proper design through simulation in Aspen Custom Modeler®. Champaign, IL (1979).
Moreover, the isomerization reactor allows shifting the pro- [16] M. Snare, I. Kubickova, P. Maki-Arvela, D. Chichova, K. Eranen,
D.Yu. Murzin. Catalytic deoxygenation of unsaturated renewable feed-
duction in the desired direction, allowing to adapt the process stocks for production of diesel fuel hydrocarbons. Fuel 87 (2008), pp.
to market demand. 933-945.
Regarding heat integration, the team found an imbalance be- [17] Li Zong, Sundaram Ramanathan, and Chau-Chyun Chen. Pre-
tween cold and hot streams in the process, making heat integra- dicting Thermophysical Properties of Mono- and Diglycerides with the
tion difficult and resulting in an excess of heat for which future Chemical Constituent Fragment Approach. Ind. Eng. Chem. Res. 49
(2010), pp. 5479-5484.
investigations are needed. Another possible process improve- [18] Juan Felipe Vélez Manco. Conceptual Design of a Palm Oil
ment includes propane steam reforming to obtain hydrogen, as Hydrotreatment Reactor for Commercial Diesel Production. Master’s
calculations show that it could cover 80% of its consumption thesis within the chemical engineering program at Universidad Nacional
rate. de Colombia, 2014.
[19] Katarina Landberg. Experimental and Kinetic Modelling Study
Finally, the economic analysis shows a profitable process after
of Hydrodeoxygenation of Tall Oil to Renewable Fuel. Master’s thesis
the 7th year, although diesel and palm oil fluctuations make it within the Innovative and Sustainable Engineering program. Chalmer’s
a risky investment. In order to improve this situation, funding University of Technology. Gothenburg, Sweden, 2017.
and governmental incentives would be needed. [20] M. Snare, I. Kubickova, P. Maki-Arvela, K. Eranen, J. Warna,
D.Yu. Murzin. Production of diesel fuel from renewable feeds: Kinetics
of ethyl stearate decarboxylation. Chemical Engineering Journal 134
Acknowledgements (2007), pp. 29-34.
[21] Atthapon Srifa , Kajornsak Faungnawakij, Vorranutch Itthiben-
The team would like to thank Professor Santos Galán for his chapong, Nawin Viriya-empikul, Tawatchai Charinpanitkul, Suttichai
support and guidance through the project as well as our path Assabumrungrat. Production of bio-hydrogenated diesel by catalytic
through university. Also, we thank Belén García from Repsol hydrotreating of palm oil over NiMoS2 /γ-Al2 O3 catalyst. Bioresource
for the valuable information received regarding regulations for Technology 158 (2014), pp. 81-90.
[22] Worapon Kiatkittipong, Songphon Phimsen, Kunlanan Kiatkit-
biofuels in the European Union.
tipong, Suwimol Wongsakulphasatch, Navadol Laosiripojana, Suttichai
Assabumrungrat. Diesel-like hydrocarbon production from hydropro-
References cessing of relevant refining palm oil. Fuel Processing Technology 116
(2013), pp. 16-26.
[1] Proposal for a Directive of the European Parliament and of the [23] R. W. Gosselink, S. A. W. Hollak, S.-W. Chang, J. van Haveren, K.
Council on the promotion of the use of energy from renewable sources. P. de Jong, J. H. Bitter, D. S. van Es. Reaction Pathways for the Deoxygena-
Corrigendum from the COM(2016) 767 document, made at Brussels in tion of Vegetable Oils and Related Model Compounds. ChemSusChem
23/02/2017. Reviews, pp. 1-20.
[2] Rasmus G. Egeberg, Niels H. Michaelsen, Lars Skyum. Novel [24] Pavel Simácek, David Kubicka, Gustav Sebor , Milan Pospísil.
hydrotreating technology for production of green diesel. Haldor Topsoe. Hydroprocessed rapeseed oil as a source of hydrocarbon-based biodiesel.
[3] Official Journal of the European Union. Directive 2015/1513 of the Fuel 88 (2009), pp. 456-460.
European Parliament and of the Council of 9 September 2015 amending [25] Pavel Simácek, David Kubicka, Iva Kubicková, Frantisek Ho-
Directive 98/70/EC relating to the quality of petrol and diesel fuels and mola, Milan Pospísil, Josef Chudoba. Premium quality renewable diesel
amending Directive 2009/28/EC on the promotion of the use of energy fuel by hydroprocessing of sunflower oil. Fuel 90 (2011), pp. 2473-2479.
from renewable sources. Art. 17 and 18. [26] David Kubicka, Vratislav Tukac. Chapter Three - Hydrotreating
[4] Neste Renewable Diesel Handbook. Neste Corporation, May of Triglyceride-Based Feedstocks in Refineries. Advances in Chemical
2016. Engineering 42 (2013) , pp. 141-194.
[5] UOP, 2012. UOP/ENI EcofiningTM Process. www.uop.com [27] Alexis Tirado, Jorge Ancheyta, and Fernando Trejo. Kinetic and
[6] Nils-Olof Nylund, Kimmo Erkkila, Matti Ahtiainen. Optimized Reactor Modeling of Catalytic Hydrotreatment of Vegetable Oils. Energy
usage of NExBTL renewable diesel fuel. VTT, 2011. Fuels 32 (2018), pp. 7245-7261.
[7] Rogelio Sotelo-Boyás, YangYong Liu, Tamoaki Minowa. Renew- [28] I. Karame, in Hydrogenation, 1st ed. (Eds: R. Sotelo-Boyas, F.
able Diesel Production from the Hydrotreating of Rapeseed Oil with Trejo-Zarraga, F. J. Hernandez-Loyo), InTech, Winchester, UK 2012, Ch.
Pt/Zeolite and NiMo/Al2O3 Catalysts. I&EC Research. Ind. Eng. Chem. 8.
Res. 50 (2011), pp. 2791-2799. [29] Xin Li, Xingyi Luo, Yangbin Jin, Jinyan Li, Hongdan Zhang,
[8] W. Kiatkittipong, S. Phimsen, K. Kiatkittipong, Aiping Zhang, Jun Xie. Heterogeneous sulfur-free hydrodeoxygenation
S.Wongsakulphasatch, N. Laosiripojana, and S. Assabumrun- catalysts for selectively upgrading the renewable bio-oils to second
grat. Diesel like hydrocarbon production from hydroprocessing of generation biofuels. Renewable and Sustainable Energy Reviews (2017).

14 Javier Moreno and Aurelio Hurtado de Mendoza

 [30] Juan Tapia, Nancy Y. Acelas, Diana López and Andrés Moreno. catalyst. Reactivity and Reaction Pathway for Base Oil Formation. Ind.
NiMo-sulfide supported on activated carbon to produce renewable Eng. Chem. Res. 43 (2004), pp. 934-940.
diesel. Universitas Scientiarum, 22 (2017), pp. 71-85. [52] Oil Price Outlook. Web article from
[31] Hachemi I, Jenistova K, Maki-Arvela P, Kumar N. Comparative https://www.globalpetrolprices.com retrieved on the 26th of
study of sulfur-free nickel and palladium catalysts in hydrodeoxygena- December of 2018.
tion of different fatty acid feedstocks for production of biofuels. Catal. [53] World Energy Council. World Energy Issues Monitor 2018. As-
Sci. Technology 6 (2016), pp. 1476-87. sessing the energy agenda for Euope, pp. 48-81.
[32] Stella Bezergianni, Aggeliki Kalogianni. Hydrocracking of used [54] International Maritime Organization (IMO). Guidance on the
cooking oil for biofuels production. Bioresource Technology 100 (2009), Development of a Ship Implementation Plan for the Consistent Imple-
pp. 3927-3932. mentation of the 0,5% sulfur limit under Marpol Annex VI, November
[33] Graca I, Lopes JM, Cerqueira HS, Ribeiro MF. Bio-oils upgrading 2018.
for second generation biofuels. Ind. Eng. Chem. Res. 52 (2012), pp. [55] Precios de los derivados del petróleo en Espana, information
275-87. retrieved at www.datosmacro.com in December 2018.
[34] Luca Zullo - VerdeNero LLC. Green Jet and Green Diesel [56] Chris Castanien. Renewable Diesel: Sustainable Hydrocarbon
Second Generation Biofuel. Powerpoint presentation available Fuel. Detroit Advisory Forum 17 April 2018.
at https://www.auri.org/assets/2013/05/Auri-April-30-2013-zullo.pdf. [57] Neste Oil (2018). Palm and rapeseed oil prices. Data source:
Last consulted on the 25th of December of 2018. Thomson Reuters. Information retrieved from www.neste.com in De-
[35] Yanyong Liu, Rogelio Sotelo-Boas, Kazuhisa Murata, Tomoaki cember 2018.
Minowa, and Kinya Sakanishi. Hydrotreatment of Vegetable Oils to [58] International Monetary Fund. Market Prices for Non-fuel
Produce Bio-Hydrogenated Diesel and Liquefied Petroleum Gas Fuel and fuel commodities from 2014 to 2017. Information retrieved from
over Catalysts Containing Sulfided Ni-Mo and Solid Acids. Energy www.imf.org in December 2018.
Fuels 25 (2011), pp. 4675-4685. [59] Michael J.Girgis, and Y. Peter Tsao. Impact of Catalyst Metal-
[36] Mikulec J, Cvengros J, Joríková L, Banic M, Kleinová A. Second Acid Balance in n-Hexadecane Hydroisomerization and Hydrocracking.
generation diesel fuel from renewable sources. J Cleaner Prod 18 (2010), Ind. Eng. Chem. Res. 35 (1996), pp. 386-396.
pp. 917-26. [60] Nataliya Belinskaya, Emiliya Ivanchina, Elena Ivashkina, Ev-
[37] Carlo Edgar Torres-Ortega, Jian Gongb, Fengqi Youb, Ben-Guang geniya Frantsina,Galina Silko. Optimal technological parameters of
Ronga. Optimal synthesis of integrated process for coproduction of diesel fuel hydroisomerization unit work investigation by means of
biodiesel and hydrotreated vegetable oil (HVO) diesel from hybrid oil mathematical modeling method. Procedia Chemistry 10 (2014), pp. 258-
feedstocks. Proceedings of the 27th European Symposium on Computer 266.
Aided Process Engineering - ESCAPE 27. October 2017. Barcelona, [61] Towler, G.P. and Sinnott, R.K. Chemical engineering design: Prin-
Spain. ciples, practice and economics of plant and process design. Butterworth-
[38] Quido Smejkala, Lenka Smejkalová, David Kubick. Thermody- Heinemann, 2nd Edition (2012).
namic balance in reaction system of total vegetable oil hydrogenation. [62] R. Turton, R. C. Bailie, W. B. Whiting, and J. A. Shaeiwitz Anal-
Chem. Eng. Journal 146 (2009), pp. 155-160. ysis, synthesis and design of chemical processes. Prentice Hall, 3rd
[39] Haiping Zhang, Hongfei Lin, Weizhi Wang, Ying Zheng, Peijun Edition (2008).
Hu. Hydroprocessing of waste cooking oil over a dispersed nano cata- [63] Donald R. Woods. Rules of Thumb in Engineering Practice.
lyst: Kinetics study and temperature effect. App. Cat. B 150-151 (2014), Wiley-VCH, 1st edition (2007).
pp. 238-248. [64] Chemicals Technology. Neste Oil Biodiesel Plant. Information
[40] Octave Levenspiel. Chemical Reaction Engineering, Ind. Eng. retrieved from https://www.chemicals-technology.com/projects/neste-
Chem. Res. 38 (1999), pp. 4140. oil-plant/ in December 2018.
[41] Charles N. Satterfield. Trickle-bed reactors. AlChe Journal 21 [65] Eurostat Statistics. Electricity Price statistics. In-
(1975), pp. 209-228. formation retrieved from https://ec.europa.eu/eurostat/statistics-
explained/index.php/Electricity_price_statistics in December 2018.
[42] Lukasz Jeczmionek, Krystyna Porzycka-Semczuk. Hydrodeoxy-
[66] Charles R. Perry, Basic Design and Cost Data for MEA Treating
genation, decarboxylation and decarbonylation reactions while co-
Units. Portable Treaters Inc., Odessa, Texas.
processing vegetable oils over a NiMo hydrotreatment catalyst. Part I:
[67] M.A. Ramos Carpio. Refino de petróleo, gas natural y petro-
Thermal effects - Theoretical considerations. Fuel 131 (2014), pp. 1-5.
química. Madrid, 1997.
[43] Vivek V. Ranade, Raghunath V. Chaudhari and Prashant R. Gun-
[68] Tim J. Hilbers, Lisette M. J. Sprakel, Leon B. J. van den Enk, Bart
jal. Trickle Bed Reactors. Reactor Engineering & Applications (2011).
Zaalberg, Henk van den Berg, Louis G. J. van der Ham. Green Diesel
ISBN: 978-0-444-52738-7
from Hydrotreated Vegetable Oil Process Design Study. Chem. Eng.
[44] F. Larachi, A. Laurent, N. Midoux, G. Wild. Experimental Study
Technol 38 (2015), pp. 1-8.
of a Trickle-bed Reactor Operating at High Pressure: Two Phase Pressure
[69] Neste Oil. How to calculate Renewable Products margin. Infor-
Drop and Liquid Saturation. Chem. Eng. Science 46 (1991), pp. 1233-
mation retrieved from https://www.neste.com in December 2018.
1246.
[45] Gianetto, A., and V. Specchia, Trickle-bed Reactors: State of Art
and Perspectives, Chem. Eng. Sci. 47 (1992), pp. 3197.
[46] Hans Korsten, Ulrich Hoffmann. Three Phase Reactor Model for
Hydrotreating in Pilot Trickle-bed Reactors. AlChE Journal 42 (1996),
pp. 1350-1360.
[47] Shigeo Goto, J.M. Smith. Trickle-Bed Reactor Perfomance. AlChE
Journal 21 (1975), pp. 706-713.
[48] A.A. Forghani, M. Jafarian, P. Pendleton, D.M. Lewis. Mathe-
matical modeling of a hydrocracking reactor for triglyceride conversion
to biofuel: model establishment and validation. Int. J. Energy Res. 38
(2014), pp. 1624-1634.
[49] M.Alifanti, C.M.Visinescu. V.I.Parvulescu, P.Grange, G.Poncelet.
Preparation and characterization of WOx-CeO2 catalysts. Studies in
Surface Science and Catalysis 143 (2000), pp. 337-344.
[50] T. Kasza, J. Hancsok. Investigation of Fuel Components Pro-
duced by the Isomerization of Bio-paraffin Mixtures. Hungarian Journal
of Industrial Chemistry 39 (2011), pp. 121-126.
[51] V. Calemma, S. Peratello and C. Perego. Hydroisomerization and
hydrocracking of Long-Chain n-Alkanes on Pt/amorphous SiO2-Al2O3

A proposal for the HVO process 15