(42ag) Three-Dimensional CFD Model of a Multitube Pd/Au Membrane Module for Hydrogen Purification | AIChE

(42ag) Three-Dimensional CFD Model of a Multitube Pd/Au Membrane Module for Hydrogen Purification

Authors 

Ma, R. - Presenter, Worcester Polytechnic Institute
Castro Dominguez, B. - Presenter, University of Limerick
Ma, Y. H. - Presenter, Worcester Polytechnic Institute
Dixon, A. G. - Presenter, Worcester Polytechnic Institute

Three-dimensional CFD Model of a Multitube Pd/Au
Membrane Module for Hydrogen Purification

 

Rui
Ma, Bernardo Castro-Dominguez, Anthony G. Dixon, Yi Hua Ma

Department of Chemical Engineering, Worcester
Polytechnic Institute,
100
Institute Road, 01609 Worcester, MA, USA.

           Hydrogen is an important chemical product
due to its broad use in the generation of ammonia and methanol as well as a clean
energy carrier. The U.S Energy
Information Administration estimated that in the U.S., energy demand grows 0.4%
per year. Coal gasification is a
potential method to solve this problem since it generates electricity and H2
simultaneously. Nevertheless, this method
also produces syngas, which is composed of methane, carbon dioxide, carbon
monoxide and other impurities. Consequently, a separation process is required for
the purification of H2. Compared to traditional separation methods, palladium
(Pd) membrane separation technology has the advantage of high H2 selectivity,
continuous operation and economic manufacturing costs. The objective of this
work is to model accurately the performance of a multitube Pd membrane module and
gain insight on the effects of different operating conditions.

 

           A seven-tube membrane module has
been designed and manufactured as shown in Fig. 1, capable of separating pure H2
from syngas mixtures. In this module, H2 enriched coal-derived syngas
is introduced into the system as the feed; H2 permeates across the
membrane and is collected at the inner tube side, while the unwanted gases are washed-out
at the retentate side. Due to the
broad variety of operating conditions, the membrane performance is hard to
predict; thus, a computational fluid dynamics (CFD) simulation is used to
optimize the operating conditions of this process. Compared to previously
reported simulations, the module in this work takes heat transfer into
consideration, making this framework more realistic since the temperature influences
considerably the permeability and stability of the membranes. The performance of the multitube membrane
module is accurately simulated, by generating a 3D model using COMSOL
Multiphysics 5.0 solving simultaneously the continuity and Navier-Stokes
equations as well as conservation of energy. Sieverts' law was used to define
the permeation rate of H2 across each
membrane, as shown in equation 1.

                                                                                                 
                                      (1)

Where  is H2 flux across the membrane, Q is the permeability
and  is the thickness of membrane. and  are the H2 partial pressures at retentate
side and permeate side, respectively. Equation 1 represents the
solution-diffusion mechanism for H2 transport; the permeability (Q) is
highly dependent on temperature. The overall effect of temperature on permeability
is introduced using the Arrhenius correlation as shown in equation 2.

                                                                                  
                                                                (2)

Where Q0 is permeability constant and Ep
is activation energy. Higher temperatures lead to higher permeabilities and thus
higher H2 diffusion rates through
the membrane.

           The simulation was carried out
under different Reynolds numbers (Re) within the laminar regime and various feed
temperatures (523K-723K); the influences on the Péclet
number and mass transfer resistances are clearly described under these
circumstances. The depletion of H2 in the proximity of the membrane surface
induces the formation of a gas boundary layer often called concentration
polarization. The H2 partial pressure adjacent to the surface of the
membranes is reduced due to the mass transfer resistance generated by concentration
polarization. Furthermore, the simulation shows this phenomenon accurately by
depicting a cross-sectional concentration profile of the shell side, as shown
in Fig. 2. The results show that higher Re reduces concentration polarization by
reducing the mass transfer boundary layer.

 

The performance of the
system is analyzed based on the total H2 recovered on the tube side from the feed under the different operation
conditions. Axial H2 concentration profiles (Fig. 3) show the H2
distribution, implying better H2 recovery at low Re. On the other
hand, operating at high Re number or high Péclet numbers minimizes
the effect of diffusion by markedly enhancing the convective forces. Although low
feed flow rates lead to a high H2 recovery, they diminish the usage
of the membrane. This is caused by early H2 depletion at the
beginning of the membrane and a noteworthy axial H2 partial pressure
gradient.

 

            Furthermore, temperature
distribution profiles show the influence of convective heat transfer of the gases
to the membranes. Permeability distribution profiles along the membranes display
the influence of temperature on H2 permeance. Operating at higher
feed temperatures, close to the membrane temperature, increases the H2 recovery
of the module; nevertheless lower temperatures display lower efficiencies due
to the reduction of temperature at the surface of the membranes.

            Based on the results obtained, the optimum
operating conditions maximize H2 recovery; typically enhanced by
low-moderate feed rates and high feed temperatures. This heat and mass transfer
model is validated by comparing the results obtained in this study with
previously reported experimental results of a single tube membrane module. The evaluation
of the actual industrial model will be performed afterwards at the National
Carbon Capture Center in Alabama.

 

 

                           

 

 

 

                                                            Fig.
1. Membrane module set-up

 

 

 

Fig. 2. Cross-section concentration
profile

 

 

Fig. 3. Hydrogen concentration profile across
the retentate side of the multitube membrane module