(706c) Conceptual Design of a Novel Efficient Hydrogen Production Process from Natural Gas Using the Systematic "G-H" Methodology

Authors: 
Anantharaman, R., SINTEF Energy Research
Gundersen, T., Norwegian University of Science and Technology

Hydrogen is a clean fuel and is thus expected to play an
important role as an energy carrier in a future decarbonized energy scenario
that has a high share of renewable energy. However, global hydrogen production
is currently dominated by fossil fuel-based processes with significant
inefficiencies. Quantifying the source of these inefficiencies and designing
novel efficient processes is essential to realizing a hydrogen-based energy
economy. Further, these novel processes have to include carbon dioxide capture
and storage (CCS). 

Natural gas (containing predominantly methane) is a common
feedstock for hydrogen production processes that use the steam-methane
reforming (SMR) reaction followed by a water-gas shift (WGS) reaction. Methane
is combusted in a furnace to provide heat for the endothermic reforming
reaction. We simulate a state-of-the-art hydrogen production process in Aspen
HYSYS, and then perform an exergy analysis of the model in order to pinpoint
the unit operations responsible for the largest sources of inefficiency. The
exergy analysis includes physical as well as mixing and chemical exergy
parameters that are obtained using user-defined subroutines programmed into
Aspen HYSYS. The results of the analysis show that the source of largest exergy
destruction (and hence inefficiency) is the furnace (43.8%), followed by the
steam generation heat exchanger (20.88%) and the reformer unit (13.73%). Using
insight from the exergy analysis, we deduce that the cause of inefficiencies is
incorrect operating conditions (temperature and pressure) of different unit
operations such as the furnace, SMR and WGS reactors. However, it is not
practical to design a more efficient process by merely changing the operating
conditions of one unit operation since all interconnected unit operations would
be affected by this change. Thus, it is necessary to employ a systematic
methodology for conceptual design of an efficient hydrogen production process.
This methodology should use a “systems-level” approach, i.e. it should
concurrently consider all the unit operations in the process as well as their
interconnections.

The objective of this paper is to use the systematic “G-H”
methodology [1] for conceptual design of a novel hydrogen production process.
The G-H methodology uses information about the change in Gibbs free energy (G)
and enthalpy (H) at standard conditions for chemical reactions to derive the
heat and work balances of the corresponding reactor. For instance, in our
project, changes in G and H are established for SMR, WGS and methane combustion
reactions at standard conditions in order to set up heat and work balances for
each reactor. The heat and work balances of the overall process are obtained by
summing up the contributions of each reactor unit operation weighted by the
extent of the reaction taking place in each operation. A desirable operating
point is chosen for the overall process and this fixes the operating conditions
for the unit operations as well. The operating temperature for each unit is
given by its “Carnot temperature” or “reversible temperature”, which is defined
as the temperature at which the reaction work requirements are provided exactly
by the heat supplied such that both requirements are met concurrently. Units
operating at their Carnot temperatures are reversible and hence are efficient
since they do not destroy any exergy. Herein lies the value of the G-H
methodology: It uses the concept of reversibility to provide a target for the
operating conditions that give the highest efficiency from the enormous choice
of reactor unit operating conditions. In addition, the G-H methodology is
applied at the systems-level ensuring that the entire process is highly
efficient rather than focusing on the efficiency of individual unit operations.

The Carnot temperatures of the SMR, WGS and methane
combustion reactions in the furnace are calculated to be 960.83 K, 974.67 K and
145,866.25 K respectively, though the furnace is designed to operate at a more
realistic temperature of 1000.00 K. The flow compositions of the material
streams are determined from the extent of reaction of the corresponding unit
operations, and a block diagram of the flowsheet is shown in Figure 1 below.

Figure 1: Block diagram of a process for hydrogen production
from natural gas    

While the G-H methodology ensures a high efficiency, it does
not take conversion of the reactions into consideration. In this paper, we
extend the G-H methodology by investigating alternative ways of achieving
maximum conversion. Specifically, we study two options: separation and
recycling of the unreacted feed and integrated separation and reaction in a
hydrogen membrane reactor. For the first case, we model the process with
separation and recycle of unreacted products for both the SMR and WGS reaction
in Aspen HYSYS. We examine how the conversion of each reaction varies with the
separation split ratio, and design a flowsheet that gives complete conversion.

For the second case, we formulate a novel procedure to
determine the conversion of a hydrogen membrane reactor given the reactor
operating conditions and membrane parameters. The problem can be stated as
follows: Assume that the SMR and WGS membrane reactors run at the operating
conditions obtained from the G-H methodology; what is the expected range of
achievable conversion for a particular set of hydrogen membrane characteristics
- permeability, thickness of the membrane layer, hydrogen partial pressure
profiles at both the feed and retentate side, hydrogen flux and the hydrogen
permeation rate-determining step. The procedure involves dividing the membrane
reactor into a series of reaction steps followed by membrane separation steps.
In each reaction step, the change in conversion is calculated analytically for
the membrane reactor operating temperature, and the products of this reaction
step are the feed of the next separation step. We assumed a palladium membrane
for hydrogen permeation and modelled the separation step according to Sievert's
law [2]. Each separation step is further divided into a number of sub-steps in
which the feed partial pressure of hydrogen is updated. The final retentate
concentration forms the feed of the next reaction step. The results of this
procedure give an attainable region of conversion for the SMR and WGS reactions.
This attainable region is used to develop a framework for estimating required
membrane areas by plotting area contours for different reactor operating
pressures. Finally, we show how this procedure fits into the G-H methodology.

Thus, we present two alternative flowsheets showing
conceptual design of novel hydrogen production processes from natural gas.
These processes are more efficient than the state-of-the-art showing the value
of the systematic G-H methodology.

References:

[1] James Alistair Fox, Diane Hildebrandt, David Glasser,
and Bilal Patel. "A Graphical Approach to Process Synthesis and its
Application to Steam Reforming." AIChE Journal 59, no. 10
(2013): 3714-3729.

[2] Samhun Yun, and S. Ted Oyama. "Correlations in
palladium membranes for hydrogen separation: a review." Journal of
membrane science
 375, no. 1 (2011): 28-45.