Passive autocatalytic recombiners are widely used as a safety measure for hydrogen risk mitigation in water-cooled nuclear reactor containments. In severe accident scenarios, hydrogen generated mainly by zirconium–steam oxidation may be released into the containment atmosphere, where local accumulation can create flammable or detonable mixtures. System-level containment analyses often apply lumped-parameter models, which are computationally efficient but usually represent recombiner operation by simplified empirical characteristics. This paper presents a concept of enhancing a lumped-parameter model of passive autocatalytic recombiners by introducing selected chemical-kinetic and transport-related effects. The proposed approach is intended to bridge the gap between detailed CFD modelling of recombiner operation and simplified system-scale simulations. The model considers the influence of hydrogen and oxygen concentrations, gas temperature, steam content and catalyst surface conditions on the effective hydrogen recombination rate. The catalytic reaction is represented in a reduced form suitable for implementation in a containment thermal-hydraulic code, without resolving the full surface chemistry. The enhanced model is compared with the previous simplified recombiner formulation implemented in the HEPCAL code. A representative loss-of-coolant accident scenario in an EPR-type containment is used as a demonstration case. The comparison focuses on hydrogen concentration histories in selected containment control volumes, total recombined hydrogen mass and the time intervals during which flammability limits may be exceeded. The proposed modelling strategy allows chemical effects to be included in fast-running lumped-parameter containment simulations while maintaining their practical applicability for multi-variant safety analyses. The results are expected to support improved assessment of passive hydrogen mitigation systems and provide a basis for further coupling of chemical-kinetic models, CFD-derived correlations and system-level accident analysis tools.