Transient-state strategies for the removal of toluene vapor in a two-liquid phase biotrickling filter: Experimental study and neural network analysis

Abstract

Abstract The performance of a two-liquid phase biotrickling filter (TLP-BTF) packed with pall ring and pumice in the presence of silicone oil, as the organic solvent (10% v/v), was evaluated for the treatment of toluene vapor under various transient conditions. The experiments were done at EBRT of 1 min and inlet loading rates (ILRs) of 150-180 g m−3 h-1 and comprise of shock loading, intermittent loading for 10 h day-1, and aeration without toluene loading during shutdown periods. The TLP-BTF removal efficiency (RE) attained steady state (RE = 66%) in less than 40 min after loading initiation in different operating strategies. The TLP-BTF was subjected to a 5.5-fold increase at ILRs from 34 to 186 g m−3 h-1, however, RE declined only 10% within 2 h after shock loading. Nevertheless, RE decreased to 44% due to dominating the kinetic-limitation regime and silicone oil saturation. The variations of TLP-BTF performance were analyzed by using artificial neural networks (ANNs) modeling by means of casual index (CI) values. Lower discrepancies for CI-values in TLP-BTF (-8.05 > CI > -10.68) in comparison to the single-liquid phase BTF (-9.19 > CI >14.04) indicated the role of silicone oil for decreasing the negative impact of different intermittent loading types on the BTF performance.