Measurement and Prediction of Turbulence in Flotation Cells

Abstract

Turbulence and its distribution are of great importance in flotation cell design as they affect suspension of particles, air dispersion and particle-bubble collision rates, which in turn determine flotation performance. The ability to model the turbulence distribution in flotation cells would be of great use as it would enhance our ability to optimise the flotation process. However, there are no mature techniques to measure turbulence in three phase (liquid-solid-gas) systems such as in flotation cells to validate modelling results. This is because the environment in flotation cells is highly aggressive and abrasive, and the system is opaque. In this thesis, the author presents two new approaches suitable for measuring turbulence distribution in three phase systems and has validated their applicability. The first technique uses a piezoelectric vibration sensor (PVS) to measure the standard deviation of the fluid kinetic energy and the second technique applies electrical resistance tomography (ERT) to measure conductivity variation in a flotation cell. The PVS technique is based on the piezoelectric effect in which the mechanical vibration of the piezoelectric material caused by the turbulence of the fluid is turned into an alternating voltage output signal. The frequency response model of the system determined by calibrating the PVS can be used to calculate the force applied on the film sensor from the electrical signal. By comparing the force applied to the PVS sensor and Laser Doppler Anemometer measurement results, it was validated that the PVS measures kinetic energy fluctuation of the turbulent fluid. When tested in a 60 L flotation cell in one and two phase environments, the PVS was shown to be capable of measuring how the turbulence changes as the cell operation is varied. The PVS was also tested in a 300m3 industrial flotation cell and gave consistent results showing how the turbulence profile changes when the cell was operated at different conditions. The ERT sensor is based on voltage measurements. When a current is injected at a pair of electrodes, the response voltages are measured at all electrode pairs. Two methods were developed to analyse the measured data produced by an ERT probe. The first method established a model to calculate the kinetic energy fluctuation in the fluid from ERT measured conductivity variance data. This method had to use the piezoelectric sensor measurement data and therefore cannot be used independently. The second approach was based on the Green-Kubo relations, which is normally used to describe thermal and mechanical dynamic transport processes. In this approach the fluid velocity fluctuation was calculated from ERT voltage measurement data. To validate the two approaches, the ERT probe was applied in a 60 L two phase laboratory cell and the measurement results were compared to that of the PVS. Results showed that the first approach worked well when air flow rates were low, but the comparison between the results deteriorated as the air flow rate increased. The second approach, however, worked well for both low and high air rates. Therefore, it has been demonstrated that ERT has the potential to be applied for measurement of turbulence in the flotation environment. To demonstrate the capability of the sensors to develop models for predicting the turbulence in a flotation cell, the PVS sensor was applied to a JKMRC 3M3 pilot flotation cell operating with two phases and a Metso 3M3 flotation test rig operating under three phase conditions. By interpolating the localised PVS measurement results, the shape of the turbulent volume in the flotation cells was obtained, from which volumes of different turbulence intensity zones were calculated. Models were then established to predict the turbulence zone volumes from the cell’s operational parameters, e.g. impeller speed, air flow rate, cell level and viscosity. Using the models, the influence of the cell operational parameters on the turbulence distribution could be evaluated. This turbulence distribution information coupled with models that relate turbulence to flotation sub-processes can potentially provide a meaningful way to control and optimise the flotation process. The ERT measurement tool was also used to measure turbulence in the Metso 3m3 cell and found to be capable of distinguishing different turbulence intensities. However, the ERT produced measurements were not accurate enough to be used for the modelling work. In future work, the ERT measurement tool can be further developed to extract other parameters from the rich information provided by the sensor, e.g. the bubble size, the bubble trajectory and velocity and the mineral slurry properties. More turbulence measurements can be conducted using both sensors in flotation cells of different design, operated under a range of different conditions and types of feed material. The datasets collected can be used to establish and validate turbulence distribution models for the measured cells. There would also be value in determining how the turbulence distribution can be incorporated into a flotation rate model to predict flotation performance.