(117c) Stripper Configurations for CO2 Capture by Aqueous Amines

A
monumental task for improving the feasibility of carbon capture from the flue
gas of coal-fired power plants is reducing the total energy cost.  A generic
absorption/stripping process using amines would reduce the net power production
of a coal-fired power plant by roughly 30%.  The reboiler duty, solvent
pumping, and CO₂ compression account for most of this power hit. 
Though alternative solvents are being studied which can improve the capture
performance, the Aspen Plus® thermodynamic model for 9 m MEA (35wt%) developed
by Hilliard is a sturdy model for a more concentrated form of the industry
standard for carbon capture, 7 m MEA.  Using this model, a fair analysis of
various stripper configurations can be performed.  It is proposed that
increasing complexity improves performance, but with diminishing return.  Since
increasing the complexity raises capital and operation costs, an optimum
configuration complexity is expected.

In prior
work by other authors, more complex configurations have been modeled to assess
potential improvement in performance of the stripper.  Some of these
configurations included Simple Stripping with Vapor Recompression,
Multi-Pressure, Double Matrix, Split Product, Internal Exchange, and Flashing
Feed.  In addition to these configurations that use packed columns, additional
flowsheets using arrangements of equilibrium flashes are considered in this
work.  This choice arose from previous results which demonstrated that the lean
CO₂ loading which minimizes the energy requirement in the stripper
is higher than typically considered.  For example, for 9 m MEA a typical
lean loading used in operation is 0.2 with a rich loading of 0.5.  However, the
optimum lean loading in the stripper has generally been in the range of
0.35-0.40 for this solvent.  This optimized lean loading reduces the solvent
capacity as well as the CO₂ flux per unit solvent.  Consequently,
the required height of packing drops to practically a negligible amount, making
equilibrium flashes a more financially conscious option.

The
double matrix configuration revealed the benefit of stripping CO₂ at
different pressure levels.  High pressure stripping reduces energy consumption
by reducing compression work and increasing selectivity for CO₂ over
water in the vapor.  Also using a lower pressure level for stripping is
beneficial to achieve the desired lean loading without using unreasonably high
temperatures.  For example, MEA has a ceiling temperature of 120°C due to
elevated thermal degradation rates above that temperature.  In this work,
analyses have been performed for configurations with 1 pressure stage, 2
pressure stages, and 3 pressure stages.  In all cases the reboilers were run
isothermally to have the closest approach to a reversible process.  This
isothermal operation would also have the added benefit of only requiring one
steam temperature, thereby reducing the impact on the steam turbines of the
coal plant.

Multi-stage
flash configurations from 1- to 3-stage flash were modeled.  In addition to the
isothermal cases, the 2- and 3-stage flash configurations were modeled
non-isothermally, where all stages were not necessarily heated.  The 1-stage
flash exhibited poorer performance than the base case simple stripper, which
can be attributed to all of the separation occurring in one irreversible step. 
The performance improved with the 2- and 3-stage flash configurations as the
separation became more reversible, as more CO₂ was collected at
elevated pressure, and as less heat was wasted from water vapor generation in
the flashes.  The multi-stage flash configurations performed well at their
optimum lean loadings, but alternative configurations with packing performed
better with low lean loadings.  These low lean loadings reflect cases in which
the absorber does not achieve adequate performance with the lean loading
optimized for the stripper.

As a
consequence of increased complexity for some configurations, variables arose
which could either be specified arbitrarily or optimized.  An example of one
such variable was the rich solvent split for the double matrix configuration. 
Another example was the distribution of vapor production for configurations
with at least two stages.  The distribution of vapor production is also
directly related to the pressure ratio between stages.  In previous work these
values were specified using good judgment or rule of thumb.  However, in order
to better understand the effect of increasing the configuration complexity,
these values were varied or optimized.

Some of
the final results for selected configurations are shown in Table 1 below.  The
optimized lean loading and the accompanying equivalent work.  Additionally, the
number of net process units for the stripper in each configuration is
tabulated.  This includes separation vessels, packing sections, and
heaters/reboilers.  The process unit count was reduced if a configuration
eliminates standard units.  For example, the stripper with adiabatic lean flash
re-compresses vapor without an intercooler, reducing the unit count by 1.

Table 1: Performance Results for
Selected Stripper Configurations

The
results of this work have elucidated several factors which improve the
performance of the stripper, pumps, and CO₂ compressor.  First,
operating with multiple pressure levels reduces the energy requirement.  This
operation provides the benefit of stripping at high pressure, but also improves
the reversibility when returning to atmospheric conditions for the absorber.  Another
factor that this work confirmed was the benefit of using low-pressure vapor
with high water content for additional stripping.  This was done in two
different ways: contacting the vapor with cool, rich solvent, or recompressing
the vapor for use in a stage with elevated pressure.  These methods condensed
some of the water in the low-pressure vapor stream to reduce the total waste
heat.  In the case of vapor recompression, the intercooling for the first
compression stage was removed entirely, and the increase of the vapor
temperature from compression replace a portion of the reboiler duty.  Lastly,
the configurations which investigated adiabatic flashes or packing indicated
that heating first always yielded preferable operating conditions.  As an example,
only heating the second stage of the 2-stage flash produced cases where the
second stage pressure was 20 times that of the first stage pressure.  Even at
the optimum lean loading, the first stage pressure was only 78% of the second
stage pressure.  Configurations with heat added first always resulted in the
highest pressure stage first, followed by lower pressures in subsequent stages.