Solution and Parameter Identification of a Fixed-Bed Reactor Model for Catalytic CO2 Methanation Using Physics-Informed Neural Networks

In this study, we develop physics-informed neural networks (PINNs) to solve an isothermal fixed-bed (IFB) model for catalytic CO2 methanation. The PINN includes a feed-forward artificial network (FF-ANN) and constraints, such as governing equations, boundary conditions, reaction kinetics. most effective structure consists of 5–7 hidden layers, 256 neurons per layer, hyperbolic tangent (tanh) activation function. forward solves the plug-flow reactor IFB, whereas inverse reveals unknown effectiveness factor involved in shows excellent extrapolation performance with accuracy 88.1% when concentrations outside training domain are predicted using only one-sixth entire domain. identifies error 0.3%, even small number observation datasets (e.g., 20 sets). These results suggest that PINNs can be used solution system identification models chemical