Power to Fuel: Optimisation and characterisation of a small scale methanol synthesis reactor

Abstract

In the effort against climate change the company ZEF has as
its target to make methanol (CH3OH) out of atmospheric CO2; a process which can
be seen as power-to-fuel and/or power-to-x. In light of the developmental cycle
within ZEF a new reactor with an increased size and enhanced mass manufacturability
is required. The reactor contains three main components: reactor bed, condenser
and heat exchanger. The heat exchanger design has been shown to be highly
efficient as well as the underlying design philosophy [1]. However, an exact in
depth evaluation of the condenser performance has not been performed yet.
Furthermore, in light of reactor optimisation, an investigation on the reactor
bed design for the use case of ZEF is seen as necessary. The condenser plays a
dual role in the reactor: it acts as the separation mechanism of products from the
recycle stream and as vital factor in generating mass flow. Due to the current
modelling architecture/philosophy of ZEF computationally simple descriptions of
condensation phenomena are desired. A method based on linear relations for the
latent heat release is developed. In combination with a PT-flash plug-in for
MatLab this allows for modelling of heat effects and liquid-gas separation due
to condensation. According to literature local heat transfer of condensation is
significantly hampered com- pared to ideal Nusselt Film Condensation theory.
The reason for this degradation is the presence of Non-Condensable Gasses which
will limit the heat flux of the condensing species to the condenser surface. A
sub model is used to evaluate whether this effect is significant. A combined
model and experimental approach is used for evaluation of these models. The
reactor bed is the generation site of the methanol and in previous work it has
been found there might be limiting effects in this bed [1]. Evaluation of
literature on the causes indicates that mass transfer and temperature
limitations are likely the cause. Furthermore literature suggests that the
reactor bed might be able to attain a higher Space Time Yield. A set of models
are made to describe the mass and heat transport of the reactor bed. These are
based on 1-D heat transfer correlations, a linearized Thiele modulus, and the
Bussche & Froment kinetic model. Furthermore, the reactor bed and condensation
models are integrated into an existing overall model. This enables the
simulation of synergy between these processes. A new reactor bed design is made
based on these models which should increase the Space Time Yield, and is
subject to experimental validation. For the experimental validation a new
reactor was designed and build based upon the new models developed.
Characterisation experiments indicate satisfactory qualitative behaviour of the
condensation modelling. Quantitatively deviations are observed which are
expected to be due the over prediction of methanol formation by the reactor bed
models. The reactor bed model deviations are mainly attributed to a lower than
predicted mass flow rate, and adverse flow fields in the reactor bed. A new
reactor bed design is proposed which should significantly reduce the adverse flow
field effects while increasing thermal performance. The reactor bed design used
allows for a decrease in catalyst size without causing a significant decrease
in mass flow rate. An increase of a factor 1.25 for the Space Time Yield compared
to the previous design has been observed during experiments. Insulation
performance is satisfactory with the insulation performing within 20 W as
modelled. Thermal efficiency has decreased by a factor 1.8 and is attributed to
the under performance of the reactor bed. Furthermore, the control of the
reactor has been evaluated in terms of mass flow rate measurements, the prevention
of the stalling of flow, and control. A new mass flow rate measurements device
based on differential pressure, a new feed injection design, and further
development of control have been experimentally validated. Designs for each of
these subjects have been found to be satisfactory. Furthermore, it was found
that reactor bed geometry also has an effect on the control of the
reactor.