(454f) Power-to-Gas : Dynamic Modeling of a Catalytic Methanation Reactor

Electrical power from intermittent renewable sources can be used, during extra production periods, to generate hydrogen through water electrolysis.

Although hydrogen can eventually be stored and used as such, methane provides higher possibilities in terms of storable quantities. Methane transportation requires less energy and lighter equipment than hydrogen transportation does (compression, canalizations sizes…) Furthermore, existing and well implemented methane grids already benefit from a strong feedback.

 Conversion of hydrogen into methane, using captured CO₂, is made possible by Sabatier reaction :

CO₂ + 4 H₂ = CH₄ + 2 H₂O

This strategy gives a new value to CO₂, turning it from a waste to a means of facilitating renewable energy development.

In this context, we build a mathematical model of a reactor performing CO₂ methanation on a nickel-based catalyst.

H₂ supply undergoes strong temporal variations, as a consequence of the power-excess variability. Using a temporary H₂ storage (in order to feed the reactor with constant reactants flow) would lead to a significant increase in process-cost. Therefore, a power-to-gas reactor has to enable dynamic operation.

Our model was thus developed in order to simulate the transient behavior of a reactor.

This abstract sums-up the main hypotheses we used for our modeling purposes, as well as the goal we pursue. Examples of simulation result curves are also available on simple request.

Models :

We simulate a multitubular, fixed-bed reactor. A cooling fluid flows parallel to the tubes, these latter being filled with catalytic pellets where methanation occurs.

1-D model is used in each reactor component: cooling fluid, tubes walls, and reacting mixture for which plug-flow is assumed. Here are the main hypotheses of the model :

• The catalytic section is modeled as a pseudo-homogeneous reactive medium (there is no mass or energy diffusion inside the catalyst pellet).
• Experimental data from Xu & Froment (1989) have been used to estimate the kinetics of the different reactions involved in the model.
• Heat transfer between the tube wall and the reaction medium is calculated according to Tsotsas (2006) recommendation.
• Heat transfer between the tube and the coolant is calculated thanks to Gnielinski correlation, based on hydraulic diameter.
• Thermal losses through the whole reactor walls are computed, considering that the insulating layer provides the only thermal resistance between the reactor and the outside. This leads to overestimation in thermal loss, which we consider as a secure choice.
• Lastly, in momentum conservation, fluid inertia is neglected with respect to pressure drop due to friction. Ergun equation, assuming catalyst pellets are spherical, is used.

Assuming all tubes behave the same way (i.e. neglecting side effects), calculations are made in one single tube concerning :

• mass, momentum and energy balances in the reacting medium
• energy balance in the metal constituting the tube and in the cooling fluid

The total amount of an extensive quantity in the reactor is simply calculated as the corresponding value at tube-scale, multiplied by the number of tubes.

However, the calculation is based on the real multitubular reactor-dimensions, for all the correlations where scale-effect exists. In particular, thermal inertia of the baffle-metal, heat loss from the reactor to the external environment, and hydraulic diameter for coolant flow, are computed on the basis of the whole reactor size.

 Purpose of our work : 

Our computation code is used to simulate a reactor operating in an on-off mode, in order to determine a range for conception parameters values, with regards to a couple of complementary (and possibly conflicting) issues :

• steady-state operation requires to meet various constraints : local temperature should not exceed a given value (catalyst preservation), catalyst amount and reactor volume should be as low as possible while enabling a high conversion yield…
• after a standby period, we want our reactor to get back to steady-state operation as fast as possible, in order to avoid using CO₂/H₂ at loss.

Running and standbys typically last many hours. Gas re-feeding/stopping take 15 to 30 seconds, as required by some European power networks regulation.

 Conclusion and prospects :

Our simulation tool enables us to calculate many quantities of interest (temperatures, mixture composition along the axial position…) in a multitubular fixed-bed methanation reactor, from starting to steady-state and vice versa. We now intend to build a pilot unit, in order to adjust our model thanks to experimental data. We will then be able to design and to size an industrial reactor.