Modelling, simulation and optimization of a copper electrolysis cell group

Abstract

Ultra-pure copper is produced in industrial scale electrolysis cell. The copper electrolysis process consumes large amounts of energy. Disturbances in the cell – like busbar contact failures, anode and cathode positioning inaccuracies and short circuits – increase the energy consumption, decrease the production and reduce the quality of the purified copper. These disturbances are inherently stochastic; predicting their location, severity and occurrence time includes significant uncertainty. The disturbances cause undesirable non-uniform current distributions in an electrolysis cell group. On the other hand, the formation of short circuits is affected by the current distribution. Hence, understanding the relation between the disturbances and the current distributions in the cell group is of great importance.

This thesis deals with computationally intensive means for product development. The 'product' in question is an electrolysis cell group and the 'means' are simulation and optimization. To simulate, one needs a model. In the modelling part of this thesis, a finite element model of an electrolysis cell group is presented. The model is multiphysical; it includes the electrical conductivity and heat transfer. Moreover, the model is augmented with stochastic sub-models for typical sources of cell disturbances: busbar contact resistance, electrode positioning and the short circuits. To optimize, one needs an optimization problem. In the optimization part of this thesis,a multiobjective optimization problem is presented and different settings – hereafter referred to as optimization settings – for solving the problem are introduced.

The main contributions of this thesis are as follows: (i) the development of a multiphysical FEM model of an electrolysis cell group, (ii) simulation results revealing the importance of taking stochasticity into account, (iii) a novel current distribution to be used instead of the cathode current in calculating the current uniformity of an electrolysis cell group, (iv) the formulation of a multiobjective optimization problem for the electrolysis cell group design and (v) the optimization settings, revealing that the design of an electrolysis cell group – defined by topology and sizing – can be optimized. In addition, the work resulted a novel current distribution system, the Outotec TwistBar (patent pending).