Abstract: The equilibrium composition of biogas-to-hydrogen conversion was simulated using Aspen Plus software, and the influence of process parameters on the thermodynamic characteristics and energy consumption distribution of the reaction system was systematically investigated. It was further explored that the enhancement mechanism of adding adsorbents to the hydrogen production process. Subsequently, response surface methodology was employed to construct binary interaction function models for the effects of temperature, pressure, and steam-to-carbon ratio on hydrogen yield and energy consumption, achieving multi-parameter collaborative optimization. Finally, to address the nonlinear relationship between process parameters and target responses, an innovative neural network prediction model was developed and combined with genetic algorithm optimization to determine the optimal process conditions for biogas-to-hydrogen conversion. The results indicate that the optimal conditions for adsorption-enhanced biogas-to-hydrogen conversion are: T = 540 °C, p = 0.1 MPa, n(steam)/n(C) = 3, and n(CaO)/n(C) = 1. Under these conditions, the hydrogen yield reaches 2.38 m³/kg, energy consumption is 3.76 MJ/m³, and the energy conversion efficiency is 91.96%. The neural network model achieved a fitting accuracy of 0.999 for hydrogen yield prediction. Compared to conventional biogas-to-hydrogen processes, the adsorption- enhanced biogas reforming process demonstrates significant advantages in terms of hydrogen yield and energy consumption control.

Key words: biogas-to-hydrogen, adsorption enhancement, response surface methodology, Gibbs free energy minimization, neutral network, genetic algorithm