Ethylene removal using biotrickling filters: part II. Parameter estimation and mathematical simulation

Mathematical models of biofiltration often encounter uncertain parameters characterizing mass transfer, microbial degradation, biofilm growth, and biofilm detachment. The genetic algorithm, which is one of the most reliable methods for optimization although it has rarely been addressed in biofiltration models up to now, was utilized to estimate the unknown parameters using given experimental data. This study combined genetic algorithm and biofiltration equations to obtain simulated ethylene (C2H4) removal efficiencies with estimated parameters. Sensitivity analysis of each parameter was assessed to observe the significance of each parameter.As a result, the simulation well characterized C2H4 removal efficiencies for most of the reactors. The large difference in removal efficiencies among reactors could be mostly explained using the mass transfer parameters. Perlite biotrickling filters with low continuous liquid flow tended to increase C2H4 removal efficiencies, due to a large active surface area of biofilm facilitating C2H4 transfer from the gas to the biofilm phase. Conversely, most of the other reactors underwent relatively low C2H4 removal because of high liquid flow that generated a severe mass transfer limitation. The low C2H4 removal in the biofilters with discontinuous liquid recirculation flow, in spite of the lowest liquid flow rate was, probably caused by a low active microbial growth condition.