Vertical Multiphase Flow in Electrocoagulation Reactors: Laboratory and Numerical CFD Modelling

Electrocoagulation is a water treatment technology that consists of generating coagulant species in situ by electrolytic oxidation of anode materials triggered by electric current. It removes suspended solids, heavy metals, emulsified oils, bacteria, colloidal solids and particles, soluble inorganic pollutants and other contaminants from water, offering an alternative to the use of metal salts or polymers and polyelectrolyte addition for breaking stable emulsions and suspensions. The method essentially consists of passing the water being treated through pairs of consumable conductive metal plates in parallel, which act as monopolar electrodes, commonly known as "sacrificial electrodes". Electrocoagulation reactors can be open (with free water surface) and closed (pressurized). The present work focuses on open electrocoagulation reactors with ascendant and descendant flow components. Given that the electrochemical process generates gas, three interacting states of matter are present in such reactors: liquid (the water being treated), solid (the particles to be removed), and the gas itself in form of bubbles. Little is known about the influence of this multiphase flow, and particularly the gas production, on the hydraulic head, flow streamlines, floc generation, the mass transfer, and the contaminant removal efficiency. In order to create recommendations for the electrocoagulation reactor design and optimization, a prototype laboratory scale model of such reactor was built at the Mexican Institute of Water Technology, and tests were carried out on it for different number of electrode plates and separation between them. Particle image velocimetry was applied to visualize and measure the instantaneous gas bubble velocity, their geometric characteristics, density and distribution, and interaction with the medium. A monochromatic high speed (8,000 frames per second) digital camera was used. The registered videos were disaggregated by the ImageJ® software, applying some filters, and the resulting images were processed by the PIVlab1.32® (time resolved digital particle image velocimetry tool for Matlab) algorithm, providing multiple flow pattern parameters and visualization. The hydrodynamic process is fundamental for the reactions that take place at the electrode boundary layers, and their interaction with the coagulant and flocs. Dead spaces and bypassing should be minimized. Tracer tests were carried out on the laboratory model, and processed by a three-parameter model developed by the authors, to determine the type of flow through the reactor and the portion of eventual dead space and bypassing. The tests showed predominantly plug flow behaviour, with small bypassing and dead space. A two-dimensional multiphase computational fluid dynamics (CFD) model of the reactor was implemented using the ANSYS Fluent software, for five geometry configurations, based on a first order implicit Eulerian scheme and k-epsilon turbulence model, using spatial discretization with first-order upwind scheme for the momentum and the turbulent dissipation rate, and second-order upwind scheme for the turbulent kinetic energy, and the PRESTO (PREssure STaggering Option) scheme with discrete continuity balance for a "staggered'' control volume about the face to compute the "staggered'' (i.e., face) pressure. Interchange between phases and flotation forces were considered. The CFD model provided the detailed velocity distribution of the water and gas phases inside the reactor, that agreed well with the laboratory model and tracer test results. The portion of dead space inside the reactor was obtained to be very low (from zero to 2%) in the tests without electric current applied, and increased with applying electric current, attributed to the influence of gas bubbles generated during the electrolytic reaction. The gas presence also increases the velocity in the ascendant part of the reactor. The velocity increase may be an important factor for the treatment efficiency as it reduces the time needed for the particle mixing. In all studied cases, the mean hydraulic residence time was very similar to the theoretical hydraulic residence time, and the concentration generated by the electrochemical reaction was higher than the concentration predicted by the theory. The results obtained were processed statistically by the Statgraphics software to obtain the effect of each variable, which was compared with the known equations for reactor design. These results can be used as tools for design and scale-up of electrocoagulation reactors, to be integrated into new or existing water treatment plants.