Modelling the mass and energy balance in a compost biofilter

A biofilter model was developed using the mass and energy balances in the gas, liquid and solid phases, which related the biofilter performance to the water content in the packing material. A key simplification of the model was that the concentration gradients in the biofilm were neglected by treating the biofilm/water layer as well mixed and in instantaneous equilibrium with the gas phase. Thus, the biofilm geometry and density parameters were lumped into the overall degradation term. The solid phase was treated as a separate well-mixed layer but solid phase dynamics were accounted for by using the Linear Driving Force (LDF) mass transfer model. The mixed form of Richard's equation together with experimentally obtained unsaturated hydraulic conductivity and water retention curves for compost were used in the continuity equation for the liquid water phase. This approach produced a model where all parameters could be potentially independently determined. The model was used to test suitable irrigation strategies for a biofilter system degrading toluene subject to different operational conditions. Under this approach both unidirectional and directionally switched biofilter configurations were tested for a 1 m long column. The unidirectional schemes incorporated both open and closed loop irrigation schemes, where the latter was based on commonly used on-line moisture measurement techniques. All schemes were evaluated based on the removal efficiency achieved and the leachate produced. Simulations under a constant irrigation rate of 5.46x 10⁻²g/m²s for a mass loading range of 13-60 g/m³h yielded removals ranging from 88%-26%. An order of magnitude drop in leachate under the high loading indicated severe drying in the system. For a high mass loading of 60 g/m³h, directional switching with a one-day frequency yielded a removal of 33% Vs 26% in an up flow scheme with similar leachate rates. Feedback control on water content provided an improved removal of 84% as compared to 73% under constant irrigation, when both schemes were subjected to load and inlet air step disturbances from 13 g/m³h to 62 g/m³h and from 298 K to 283 K respectively. A sensitivity analysis indicated that the model was most sensitive to the relationship between moisture content and degradation, which was also reflected by the high sensitivity of the model to the kinetic parameters in the degradation term. A novel batch recycle reactor was thus developed to investigate the effect of water content changes on the degradation rate in low water content systems such as biofilters. The reactor tightly controlled the water content of the unsaturated packing material by using the principle of a suction cell. Experimental runs were performed with toluene as the contaminant using unamended compost at a constant temperature of 30°C. Matric potential in the compost was maintained at values between -6 and -36 cm H₂O and the gas phase was monitored by sampling/gas chromatography. A soil water retention curve relating matric potential to gravimetric water content was generated for the compost. Periodic dry weight analyses of reactor samples together with the water retention curve verified moisture content control. Degradation results demonstrated a biologically limited region followed by a non-linear region at lower concentrations. Elimination capacities were obtained along the wetting and drying curves and changes in the water content affected the removal rates in the linear region ranging from 155 g/m³h to 24 g/m³h over the matric potential range investigated. Repeatability studies indicated that moisture content was the most likely parameter that influenced the changes in performance. Batch scale experiments were also performed using microbially inhibited compost, which provided linear sorption isotherms for toluene on compost at concentrations between 0-1000 ppmv and temperature values of 25°C and 35°C. The simulation model developed here provides a useful tool to implement and evaluate various operational schemes under different irrigation strategies. This is achieved by way of greater flexibility in incorporating the various schemes into the base model and the comparatively low simulation time to obtain the relevant results