In the industrial aluminium smelting process, hundreds of cells are placed inside the potroom building. Ventilation of this space is primarily provided by natural convection where the electrolysis cells are the heat sources. Occasionally, the ventilation is also affected by winds accessing the building through its openings and windows.

In previous works, the simulation of potroom ventilation has been shown to be a difficult task, particularly regarding the appropriated choice of the turbulence model, which greatly affected the flow pattern outcome. Difficulties to find a turbulence model that can correctly represent the thermal plume that forms above the cells and rises towards the roof have been reported. Two equation turbulence models such as k-epsilon and k-omega have been shown to be inadequate. The Reynolds flux model was found to give good results, but this turbulence model is no longer supported by the major commercial codes, allegedly due to its limited range of flows applicability.

In this work, the physical model studied by Dupuis is revisited. The buoyant flow numerical simulations of the physical model are compared with the experimental results. Both Large Eddy Simulation (LES) and Detached Eddy Simulation (DES) are intrinsically transient turbulence models and both model results are shown to be representative when compared with the experimental work. Finally, DES is used in an industrial potroom ventilation simulation. DES proved to be the most suitable choice of turbulence modelling for industrial potrooms while maintaining results accuracy, because it requires less time and space discretization than LES.

Cell energy balance is typically treated independently from both magnetohydrodynamics (MHD) and pot-to-pot busbar design, even though all three aspects are intertwined in many ways. Of particular interest is the fact that the downstream (DS) sidewall of typical side-by-side, side riser reduction cells is usually hotter than in the upstream (US), suggesting that the MHD behavior impacts the distribution of heat losses around the shell. This has been investigated by different authors and it was found that the flow of the liquid metal and bath, the reduction of anode slot depth with anode block consumption and the metal pad heaving all contribute to the asymmetry of the potshell temperature distribution. Moreover, during the 41st International ICSOBA Conference in 2023, it was suggested that hotter DS cathode busbars would also contribute to the hotter DS sidewalls observed in actual practice; specifically, the DS busbars are hotter because they have smaller cross-section as the means of balancing the US-to-DS current split. To address this intriguing assertion, this work investigates the radiation heat transfer between cathode ring busbar and potshell, which has not been included in the past models. The radiation heat transfer between the potshell, pot-to-pot busbars and ambient was implemented in the modernized ANSYS-based cell energy balance model presented earlier [1, 2]. Key conclusions are illustrated by means of numerical results obtained for a fictitious 375 kA reduction cell.

MHD stability is known to limit aluminium reduction cell energy efficiency. Normally cell stability is achieved by designing the cell with an optimized magnetic field and electric current distribution in the liquid metal. Modelling of electric current distribution requires a detailed 3D representation of the cell cathode assembly coupled to the liquid metal zone. The modelling software MHD-VALDIS is an established tool for MHD stability investigation and cell design. The recent update is described which permits includes variable contact resistance along collector bar and carbon, temperature dependent electrical conductivity of the collector bar, ledge profile along cell walls, etc. Cathode voltage drop (CVD) and horizontal current density in the metal pad, obtained by MHD-VALDIS are compared with those obtained with 3D ANSYS software. The impact of reducing both JY and JX on the MHD stability are then analyzed using MHD-VALDIS software.

The adaptation of the control strategy to be used during the life of the aluminum reduction cell is a critical issue. The evolution of the dynamic behaviour of the cell leads to significant changes in the thermal balance and current efficiency, which require such adaptation to maintain efficient process. To do so, a good understanding of the parameters affecting the cell behaviour is required. One of those parameters is the metal mass in the cell which affects, among others, the thermal balance and hence the protective ledge. During its life, the erosion of the cathode leads to an additional mass of metal in the cell, which may also change this equilibrium. To guide the control strategy, a calibrated thermoelectrical ¼-cell model (ANSYS™).is used to investigate the effect of the metal mass via the variation of the metal pad height and/or the eroded cathode profile. The results obtained from the simulations allow the estimation of the energy input correction for a given metal mass in order to assure proper protective ledge/thermal balance. This estimation is based on the minimisation of the difference between the obtained and calibrated ledge profile. Such procedure and results give critical information for further optimisation phase