Aluminium smelter cell energy flow monitoring

Abstract

A method of measuring the duct heat losses from the top of an aluminium reduction cell has been developed. The duct gas flow rate was measured using a nozzle whose design is based on that of a venturi meter. This has the advantage of not becoming blocked with particulates, which other devices are prone to. The duct gas temperature, as well as the ambient air temperature, was measured in combination with the flow rate to give the duct 'heat losses. From the duct heat profile, it was possible to distinguish process operations in the cell such as alumina feeding, anode changing and metal tapping as well as conditions such as anode effects. These events had strongly characteristic duct heat profiles, and individual parts of the operation, e.g. removal of hoods and breaking of the crust, could be observed from the profiles. The method gave a clear indication of hot cells, with a good correlation between the duct gas and electrolyte temperatures. Therefore, as well as indicating events at a cell, duct heat losses can also give an indication of cell condition. It was found that increases in tile duct gas flow rate, while reducing the temperature of the exhaust gases, led to an overall increase in the top heat losses. This impacted on the heat balance in the cell and led to changes in the temperature and chemistry in the cell. These changes were measured and were able to be related to the duct heat losses. A simple model was able to adequately explain the observed changes. The impact of a change to the duct heat loss on the cell condition highlights the need to include all components when performing an energy balance around a cell. An energy balance boundary was used, which surrounds the entire cell and considers all material and energy flows across the boundary. The energy components of ambient and exhaust air, as well as anode cover material and the usual raw material inputs, alumina and carbon, all need to be considered. Heat balance changes affect vanous aspects of the cell, including the superheat and sidewall heat transfer. A study was made of the heat flux in a cryolite-based electrolyte. This was done in a laboratory cell by immersing cold objects in a superheated electrolyte. The heat transfer coefficient was determined by measuring the temperature within the probe. By varying probe characteristics such as mass and initial temperature, as well as electrolyte conditions such as superheat and velocity, the effect on the heat transfer coefficient and heat flux was able to be determined. The measured heat transfer coefficient, Aluminium smelter cell energy flow monitoring at around 1000 W:rrl2K-1, was within the typical range used by some smelters and determined by other researchers. It was found that the heat flux in the electrolyte could be predicted by recording the remelt times during a number of measurements. However, this was limited to a single electrolyte composition, and more work needs to be done at lower heat fluxes to make the work more useful. The effect of different electrolyte composition was to change the remelt time, as well as the proximity of the remelt temperature to the liquidus temperature of the electrolyte. The frozen layer that forms on the probe was found to physically detach before it had completely remelted. The time at which this happens depends on the electrolyte composition. Alumina additions increase the remelt times and temperatures whereas aluminium fluoride additions have the opposite effect. This phenomenon helps explain the discrepancies that have been demonstrated for some liquidus sensors and illustrates the potential shortcoming of other sensors where the same problems have not been highlighted. In short, a method of measuring the duct heat losses from a cell has been developed and the usefulness of this shown by both measured data and an energy balance. Determination of the heat flux within the electrolyte has also been studied, with a dependence on composition being observed.