Break

Determination of
Kinetics for Methane Dry Reforming Using Plasma

Presenting
author: Cassiane Oliveira
Martins (cassiane.oliveira-martins@univ-pau.fr)

Co-authors: Frédéric Marias
(frederic.marias@univ-pau.fr), Jean-Paul Robert-Arnouil
(jprarnouil@europlasma.com), Stéphanie Moyal (smoyal@europlasma.com)

Keywords: Reactor Design, Kinetics, CFD, Modeling,
Plasma Reactor, Methane Dry Reforming

The steel industry is crucial for
global development. The worldwide steel demand is forecasted to reach 1,735 Mt
in 2019 and all scenarios foresee strong additional growth over the coming
decades, considering the rapid development of the BRICS countries. Although in
Europe the forecasted growth is lower than the developing countries, since most
of its infrastructure is already established, the European steel demand still
increases proportionally with population growth.

Since steel production is mainly
made from iron ore using carbon-based reduction processes, the steel industry
is reasonably confronted with the major challenge of climate change. It is
estimated that nowadays the steel industry is responsible for 4-6% of global
anthropogenic greenhouse gas emissions and for 25-30% of industrial emissions.

In this scenario, the IGAR (Reducing
Gas Injection) project has been devised by European steel industry, in
particular by ArcelorMittal, to reduce the carbon dioxide emission. This
project is based on the proposition of directly injecting reduction gases,
consisting primarily of carbon monoxide and hydrogen, in the blast furnace to
substantially reduce its solid carbon requirement and consequently the steel
industry CO₂ emissions. 

The IGAR project considers that steel
gases, produced in large quantities in the integrated plants, will be treated
by a plasma reactor where the dry reforming of methane ()
will take place. The reducing gas produced by the plasma reactor will be then
injected into the blast furnace, as shown by the figure below.

Hence, the IGAR Project has two
main objectives:

·       
To test on a pre-industrial scale a plasma conversion system for
iron and steel gases to ensure their heating and conversion by the methane dry
reforming.

·       
To validate, under industrial conditions, a tuyère specifically
designed for the injection of gas produced with pure oxygen and coal, under the
operating conditions of the blast furnace process.

The plasma reactor will be
installed in ArcelorMittal plant in Dunkerque-France and might be used during a
long test test period for pre-industrial validation. In order to design the
reactor several studies are being carried out including the computational fluid
dynamic (CFD) model of the reactor and a CFD model of a small-scale pilot that
will be constructed by Europlasma Company. This pilot will be installed in
Europlasma research and development plant in Morcenx-France.

The CFD model of the pilot was
developed by using the ANSYS 19.2 software package and its geometry was constructed
in the ANSYS Space Claim Direct Modeler 19.2, which is presented by the figure
below. Due to confidential issues, the pilot dimensions are not shown in the
figure.  

The mesh of the CFD model was
composed by 7.362.606 tetrahedral elements and a system of 12 homogeneous
reactions was considered to describe the reactional dynamic in the CFD
simulation, as presented below:

One of the challenges of the IGAR
project involves the determination of the reactions kinetics considering the
plasma environment. The CFD model was developed by using the Arrhenius kinetics,
announced in the scientific literature for a non-plasma environment, for each
one of these 12 reactions. However, considering the presence of ions and free
radicals in the plasma, the intensity of these reactions might be stronger; therefore,
this model should be adjusted.

To adjust the kinetic model, experiments
will be carried out by using the small-scale pilot that will be constructed in
the Europlasma R&D plant. As a first step, in order to have a better
comprehension of how to modify the kinetic parameters to adjust de model, a study
was performed to analyze the effect of their variation on the molar fraction of
the main products (CO₂, H₂ and C_(s)) at the
outlet of the pilot.

The kinetic parameter chosen for
this study was the activation energy (E) of Arrhenius equation. The approach
used for the study was to impose a variation on the activation energy of the
most important reactions of the system and then run a series of CFD simulations
of the pilot’s 3D model.

Thus, to follow this approach it
was firstly necessary to determinate the most important reactions, the method
used for this purpose is described as follow:

1.      The
weight of each reaction on the overall conversion of every species was calculated
by using the following expression:

Where:

·     
 :
plasma torch volume,

·       
 :
stoichiometric coefficient of the species i on the reaction j,

·       
 :
 reaction rate. 

This
calculation considered the reaction rate results obtained with the 3D CFD
simulation of the pilot by using the kinetic parameters issued from the
literature for a non-plasma environment for each one of the 12 reactions.

2.      For
each reaction the species average weight was calculated by the following
expression:

In which  
is the number of species of the system.

3.       The
number of species with weight SR_i,j  greater than zero was counted.

The
reactions with largest species average weight  and
highest number of species with weight SR_i,j  greater than zero are
ranked as the most important in the system. Thus, by using this method it was
possible to determine the three reactions that most affect the results in the
pilot outlet.

The
methodology chosen to study the effect of a variation in the activation energy
of each one of these three reactions on the simulation results at the pilot
outlet was the 2^k factorial design, which means that for each reaction
the effect of a variation of the activation energy will be studied in two
levels: high and low.

Taking
into account the presence of ions and free radicals in the plasma, the
intensity of the reactions might be greater, which can be considered as a lower
activation energy in the Arrhenius equation. Following this reasoning, the
activation energy announced in the literature in the non-plasma environment was
considered as the high level. The low level, in turn, was chosen considering a
result obtained for a preliminary study that have shown that the lower the
activation energy value used in the pilot’s CFD simulation, the higher the
necessary number of iterations to achieve convergence. Hence, the first 2³
factorial design was made with a variation of 10% between the low and high
levels ().

Therefore,
the complete factorial design was made with 8 CFD simulations contemplating all
the possible combinations of the activation energy of the three most important
reactions in the low and high levels. The response variable was considered as
the main products’ molar fraction at the reactor outlet and the main effects were
used to construct a regression model such as:

Where:

·     
:
molar concentration of the specie i on the reactor outlet,

·       
 :
regression coefficients,

·       
 :
coded variable calculated by ,

·       
FD : Factorial Design,

·       
RO : Reactor Outlet.

The
interest of this regression model is to create an expression, as simple as
possible, to associate the activation energy of the main reactions directly to
the molar fraction of the main products at the reactor outlet. Thus, with the
results that will be obtained during the experimental campaigns at the pilot,
it will be possible to have a first estimation of the activation energy values
that must be used in the CFD kinetic model to obtain simulation results that
match the experimental ones. The adjusted kinetic model will be then used to
design the industrial reactor that will be installed in the ArcelorMittal plant
in Dunkerque-France.

This research was supported by
the Investissement d’Avenir program funded by the Agence de l'environnement et
de la maîtrise de l'énergie (ADEME).