|
The electrostatic precipitator (ESP) is an efficient device for removing fine particles from large volume flows, such as flue gas from a power plant, to meet the low emission standards prescribed by environmental legislation.
The particles are removed from the gas by the strong electrostatic field between rod-like negative emission-electrodes placed between grounded collector plates that form the side-walls of ducts on which the dust accumulates. Being usually dielectric, particles tend to keep their charge, hence stick to the walls. Periodic rapping of the plates causes the dust to slide down and be collected in bins below. It is the Coulomb force acting on a particle that makes it drift to the plates to precipitate. Each particle has received the charge from the gas by processes of field charge and diffusion charge, because the gas itself has been charged by corona discharge from various points along the negative emission electrodes.
Since the gas carries charge, it is also acted on by Coulomb forces determined by the complex three-dimensional electrostatic field and charge distribution between emission electrodes and plates. This means that the field induces secondary flows and turbulence in the gas, superposed the otherwise uniform bulk flow through the ducts, and these disturbances affect the efficiency of the precipitator, hence its environmental impact.
Although ESPs have been built and used on an industrial scale since the 1920s, the traditional semi-empirical design criteria are insufficient today. Legislation continues to tighten standards for emission to the atmosphere, particularly of fine particles, so the design of high-efficiency ESPs relies increasingly on advanced CFD-models to compute 3D electrical fields, velocity and turbulence fields, and particle charging, transport and deposition. The parameters of an optimal design include electrode geometry, particle properties and loads, etc.
Testing of the validity of the CFD-models calls for reliable experimental data on both global and local quantities, such as efficiency versus particle size, current density distributions, and in situ 3D distributions of velocity and turbulence. This has been the main objective of the research at MEK, using optical methods such as stereo PIV and LDA in a laboratory-scale test facility (Fig. 1). Results show the nature of secondary flows as axial rolls (Figs 2 and 3), and the way in which their strength and level of turbulence increase with increasing current density and decreasing bulk velocity, expressed by an inverse electrical Froude number (Fig.4). When dust (here Rollovit particles with a mean diameter of 1.9 mm) is precipitated on collector plates, the compaction of the dust layer varies with the local current density, forming an open and thick dendrite structure at low current density. This feature is used to check the accuracy of electric field calculations (Fig. 5).
The physics and fluid mechanics of the ESP are interesting and challenging. Given the electrode geometry and potentials, the electric field strength and charge density can be calculated from the Maxwell equation of the E-field and the charge conservation. The distribution of charge density in the gas is part of the solution and depends on the details of the corona process. Since this is difficult to model, we use the measured total current as an additional boundary condition to adjust the charge density at the corona points so as to give the right current at the collector plate. This electrical problem can actually be solved without regard to the gas motion because the currents associated with these charge-flows are negligible compared to the conduction currents in the charged gas. Thus, while the electrical field problem is decoupled from the flow, the gas flow depends on the field in terms of the Coulomb body force appearing in the momentum equations and in the turbulent kinetic energy balance. Conceptually interesting, the electric field may be perceived to act on the gas through a distribution of vorticity sources in the volume, which drive secondary flows and generate turbulence.
The research was supported in part by EFP-2000 in collaboration with FLS Airtech A/S and Force Technology, and in part by a Ph.D. stipend from DTU. Two M.Sc. students also contributed by their thesis work.
Fig.5. Computed and measured current density at collector plate, and observed dust pattern (U-100-50 electrodes).
|
Fig. 1. Test section (0.2´0.2´1.0 m) with 7 electrodes and stereo PIV setup, giving 3 velocity components in plane perpendicular to the bulk flow (duct rotated 90° for optical access).
Fig.2. Cross section of test section showing area investigated by PIV and the measured secondary flows in the form of axial rolls driven by the electric field (see also Fig.3).
Fig.3. Secondary flows in lower half of test section. Upper. from PIV (color contours give axial velocity). Lower: from CFD calculations.
Fig.4. Measured mean strength of turbulence (Tu) and secondary flows (<VW>) over lower cross-section versus inverse electrical Froude number.
|