Biomass accumulation in a biofilter treating toluene at high loads – Part 2: Model development, calibration and validation Antonio David Dorado a,⇑ , Javier Lafuente b , David Gabriel b , Xavier Gamisans a a Department of Mining Engineering and Natural Resources, Universitat Politècnica de Catalunya, Bases de Manresa 61-73, 08240 Manresa, Spain b Department of Chemical Engineering, Universitat Autònoma de Barcelona, Edifici C, 08193 Bellaterra, Barcelona, Spain highlights " A rigorous model predicts satisfactorily performance and biomass accumulation. " Model is based on a few numbers of parameters and simple system of equations. " Pressure drop, bed reactor porosity and biomass weight are described by model. " Model considers neither a uniform biomass nor a constant distribution in biofilter. " Model predicts how biomass accumulation is favored up to clogging for tested conditions. article info Article history: Available online 17 August 2012 Keywords: Biomass growth Kinetic parameters estimation Modeling Toluene abatement High loads Biofiltration abstract In this work, a dynamic model describing volatile organic compounds abatement and the corresponding biomass accumulation is developed, calibrated and validated. The mathematical model is based on detailed mass balances which include the main processes involved in the system: advection, absorption, adsorption, diffusion, biodegradation and biomass growth. The model overcomes common assumptions considered in classical biofiltration models such as uniform, constant biomass distribution. The model was calibrated and validated using experimental data obtained from a biofilter packed with clay pellets during its operation from inoculation to clogging. The model was able to predict satisfactorily experimen- tal data by calibrating only a minimum number of parameters such as the half-saturation constant for toluene and the volumetric maximum growth rate of microorganisms. Kinetic parameters were fitted by means of an optimization routine using toluene concentration profiles along the bed height of the biofilter. A confidence interval for each parameter was calculated based on the Fisher Information Matrix procedure. The model was satisfactorily validated during the operation of the biofilter under different process conditions. Biomass accumulation permitted to predict macroscopic, critical operating parameters such as the pressure drop through the bed. The model may help predicting energy consumption requirements as well as biomass clogging episodes due to excessive biomass growth. Ó 2012 Elsevier B.V. All rights reserved. 1. Introduction Although biological gas treatment has become an effective and economical alternative to traditional systems, biofilter modeling has been scarcely developed due to the complexity of describing the fundamental processes and due to the lack of reliable physical, chemical and biological parameters. Although more than 200 experimental works focusing on VOCs biofiltration have been pub- lished in the last 5 years, less than 40 studies dealing with mathe- matical modeling of toluene removal by biofiltration can be found in literature in the same period of time. Amongst them, simple and complex models have been employed to model biofiltration under both steady-state and dynamic operating conditions [1–6]. Dy- namic models are generally more interesting since they could serve to predict biofilters performance and for process control and optimization [7]. However, most dynamic models in biofiltra- tion rarely incorporate biomass accumulation processes, even if experimental observations do not correspond with this common assumption. Instead, a constant and uniform biomass distribution is commonly assumed. Then, important scenarios in full-scale bio- reactors operation such as bioreactors startup, starvation periods and clogging episodes are not well-predicted if biomass accumula- tion processes are not considered in the model. In addition, since biomass density is strictly related with the availability and concen- tration of pollutant in the gas-phase, predicted results could differ 1385-8947/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cej.2012.08.019 ⇑ Corresponding author. Tel.: +34 93 877 72 36; fax: +34 93 877 72 02. E-mail address: dorado@emrn.upc.edu (A.D. Dorado). Chemical Engineering Journal 209 (2012) 670–676 Contents lists available at SciVerse ScienceDirect Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej