| Description |
In the aluminium industry, the Hall-Héroult process is used for smelting of aluminium using carbon electrodes such as anodes. The baking process of anodes is required for their maximum efficiency during this electrolysis process. The anode baking process contributes up to 15% costs in the aluminium industries and therefore, has been proven to be an important field of research since the 1980s. The process consists of various interdependent physics such as turbulent flow, combustion, radiation, and conjugate heat transfer. The ideal anode baking process should attempt to optimize energy consumption, reduce NOx and CO2 emissions and improve anode quality. The focus of this project is to understand the parameters that affect thermal NOx production in the anode baking furnace and thereby, finding optimum values of these parameters. In this paper, a 2D reactive turbulent flow model is developed using COMSOL® Multiphysics finite element software. The effects of radiation are elaborated for a 2D model. The temperatures calculated by the model are compared with the temperatures measured in the furnace using an infrared thermal camera. The comparison shows that the temperatures obtained by the model are in the range of measured temperatures. The combustion modeling is mixing dominated. A 3D model provides more accurate mixing behaviour. Therefore, an extension of the 2D model to 3D using COMSOL® Multiphysics software is developed. In this paper, results of 3D non-reactive turbulent flow results are discussed. Initially, mesh sensitivity of the 3D results is analysed by comparing three mesh refinement levels. Subsequently, the results are compared with another simulation environment, IB-Raptor code. The comparison shows that the differences in the results are mainly observed near the fuel pipe. These differences can be attributed to the dissimilarities in mesh size and structure. However, these differences are within a 10% range and therefore, the two codes can be considered to provide comparable results. An improvement in the comparison with the two tools can be achieved by increased consistency in meshing. Furthermore, a 3D non-reactive flow would be improved by implementing other important physical phenomena similar to the development of a 2D model. |
