| Abstract Scope |
This work introduces an advanced lattice Boltzmann method (LBM) incorporating an 8-wave magnetohydrodynamics (MHD) formulation for simulating anodic bubble dynamics and electrolyte flow in Hall-Héroult cells. The model uniquely integrates the 8-wave MHD framework, ensuring numerical stability, Galilean invariance, and physical consistency under variable energy input. Coupled with a multiphase LBM, it accurately captures bubble formation, coalescence, detachment, and transport beneath the anode.
GPU implementation using CUDA and JAX ensures computational efficiency and scalability for realistic cell-scale simulations. Initial validations demonstrate the model's stability, accuracy, and computational performance, establishing a foundation to explore spatial alumina transport and optimized alumina feeding strategies with adjustable Anode-Cathode Distance (ACD) to maintain thermal and electrochemical balance. Future work will extend the model to incorporate thermal-electrochemical coupling and machine-learning-enhanced reduced-order modeling, aiming to improve predictive capabilities and operational control. |