| The flow of melt couples with electric current, temperature and magnetic field in aluminum reduction cells. Their interaction has significant influence on the current efficiency and the cell life. Therefore, it is important both in theories and practices to understand turbulent behavior of metal pad flowing, to choose the appropriate turbulent model and to increase the numerical calculating accuracy of flow field, which will result in great improvement of the technology of the simulation and optimization of cells as well as rapid progress in aluminum business.In this paper, the motions of melt in cells have been analyzed and the model of electromagnetic hydrodynamics been discussed. Based on the conception of hologram simulation, multi-field coupling comprises turbulent fluctuating correlation, time-averaged correlation, circular correlation, bilateral and multi-phase correlation. The coupling mechanism has been discussed in detail.Based on the commercial software CFX 4.3, the subroutines, such as the input of magnetic and current distribution, the calculation of the source term in momentum equation (electromagnetic force) and interface shape of metal-bath, were developed. The magnetic and current fields obtainea by numerical simulation were taken as known conditions. The segregated method of multi-field in the cell was employed. The flow fields of metal melt were numerically calculated in three different prebaked cell designs, i.e. 82kA cell with three anode risers, 156kA cell with two anode risers and 200kA cell with four side anode risers. Furthermore, metal velocities were measured using iron rod dissolution method. The results are of great help to verify the mathematical models and study electromagnetic hydrodynamics models.The influence of grid density on the calculating accuracy in different models was studied quantitatively. The results show that grid density in the boundary layer has less influence on high Reynolds' number model and has greater influence on low Reynolds' number model. Therefore, a finer grid scheme in the boundary layer must be used if low Reynolds' number model is employed.According to the definition of turbulent Reynolds' number, the turbulent intensity of metal melt in cells was analyzed. Reynolds' numbers in above three different cells are of the order of 103. So, the turbulent flow of metal pad is of high Reynolds' number. Moreover Reynolds number is also an important number to be applied to determine the turbulent intensity of fluid. It is proposed that metal depth should be taken to be the characteristic length for calculating Reynolds' number at metal melt zone. Then, Reynolds' numbers in above cells are of the order of 104.Standard k-ε model, low Reynolds' number Jones-Launder k-e model and RNG model were adopted to simulate the flow field of metal melt in above three different cells. The results were discussed and were compared with measuring data. The conclusion that the turbulent flow of metalpad is of high Reynolds' number was further confirmed. So, the preliminary conclusion was got, i.e. low Reynolds' number model is not suitable to calculate metal melt flow in cells if the current is great than 80kA.The function that electromagnetic force inhibits turbulence of metal melt motion was analyzed quantitatively by means of Hartman number. Hartman numbers in above cells are of the order of 104. The simulating results showed that electromagnetic force has a stronger inhibition to the metal pad flow and turbulent motion of metal in cells is approximate to that of low Reynolds' number, if Hartman number is near to 2 X 104.
|
PDF Full Text Request