文摘
Two-stage bioreactors, i.e., a biotrickling filter (BTF) and a biofilter (BF) were subjected to transient state operations. The removal efficiencies of methanol (REM), hydrogen sulphide (REHS) and α-pinene (REP) were modelled using artificial neural networks (ANN). ANN topologies were: 3-4-2 and 3-3-1 for the BTF and the BF, respectively. REM and REHS in the BTF were affected by the presence of α-pinene. REP in the BF was affected by the inlet concentration of hydrogen sulphide.