Reduced Order Modeling of tubular reactors for synthetic methane production

Synthetic methane can be produced by a heterogeneous reaction system of hydrogen (H2) and carbon dioxide (CO2) on the surface of metal-based catalysts. The reaction system is mostly exothermic; therefore, effective heat management is required. Three-dimensional conjugated heat transfer simulations enable a detailed evaluation of the heat distribution in the reactor tubes and the effect of convective cooling. Since the reactor tubes are generally patterned, an optimization exercise with chemical composition and physical parameter space exploration can already be applied to a single-tube reactor in the first step. Still, this can be computationally intensive, which is why a reduced order modeling (ROM) approach in modeFRONTIER is investigated to minimize the computational cost, and maximize the throughput of design variations for the optimization study. A single tube reactor is first considered through which the mixture of CO2 and H2 flows. The exothermic reaction system is solved using a one-dimensional adiabatic reactor model coupled with detailed chemistry. Computed heat release and species source terms are mapped to the single-tube reactor along its axis. The tube reactor is parametrized with pressure, temperature, velocity, and composition. Likewise, the air cross-flow stream is parametrized with velocity and ambient temperature to model the effect of forced convection for cooling of the tubular reactor. The results of an efficient parametric space exploration of the reactor model are then created using a CFD solver. Proper Orthogonal Decomposition (POD) in modeFRONTIER is applied to the respective scalar and vector field variations of the results. From the POD modes the fields can be predicted to obtain results in real-time. The coefficients of the POD are mapped to the respective input parameters through regression, enabling the use of this ROM model in optimization exercises for the single-tube reactor design.