Designing ozone bubble columns: A spreadsheet approach to axial dispersionmodel

When designing ozone bubble columns, two major sources of uncertainties usu ally exist: (1) the measurement techniques and the estimation methods of th e various operating parameters; and (2) the application of the pertinent de sign model. This paper presents a simple and easy-to-use, yet accurate and reliable design model for describing the performance of ozone bubble column s for water and wastewater treatment applications. This model is a modified non-isobaric steady-sate one-phase axial dispersion model (1P-ADM). The 1P -ADM is different from the complete axial dispersion model, or referred to as the two-phase axial dispersion model (2P-ADM), in its simple use for pra ctical design and process control of full-scale contacting chambers. The 2P -ADM is represented by a system of two non-linear partial differential equa tions. In order to solve that system of equations, an elaborate numerical s olving technique is needed. On the other hand, the 1P-ADM is composed of a single non-homogeneous linear second-order ordinary differential equation r epresenting the liquid phase. Yet, this liquid-phase differential equation accounts for the countering effects of the gas bubbles' shrinkage and expan sion caused by gas depletion and absorption and reduced liquid hydrostatic head. The differential equation was solved analytically by the method of va riation of parameters. Expressing the 1P-ADM in terms of dimensionless oper ating parameters and with the available analytical solution of the differen tial equation, the model predictions of the dissolved and the gaseous ozone profiles along the column height were examined using a simple spreadsheet approach. Therefore, describing that analytical solution in terms of a simp le spreadsheet program facilitated obtaining the model predictions for any operating conditions represented by the model parameters entered into the s preadsheet program. Consequently, using the 1P-ADM for process design and/o r on-line process control becomes very feasible. The 1P-ADM was initially t ested to evaluate its predictions of the dissolved ozone profiles for water treatment conditions. The model provided excellent predictions of the diss olved ozone profiles along the bubble column for the counter-current and th e co-current flow modes.