Abstract

A result of the energy transition in Germany, widely known as ’Energiewende’, is a volatile supply of energy available to consumers. These fluctuations can be compensated by synchronizing industrial processes, named ’Demand Side Management’. Aluminium production plays a key role, here. On the one hand, the energy consumption is significant (more than 1 % of German electricity consumption), on the other hand, the process itself offers a great potential for flexibility. Generally, electrolysis cells are operated with a constant current. An adaptation of the process control to fluctuating input power is required. The three submissions named ’Modeling of aluminium electrolysis process: part I-III’ each show approaches to solving individual challenges that arise as a direct result of this. This work deals with the prediction of the bath temperature depending on future current modulations. The temperature is a key parameter for a stable operation of the cell. The realization of the future prediction model is implemented in Matlab®/Simulink®. The model represents all relevant physical and chemical phenomena involved in aluminium production. Furthermore, interventions in the process, such as the changing of anodes, the removal of aluminium or the feeding with aluminium oxide can be simulated. In order to generate precise predictions, the model is adapted to the specific temporary properties of the investigated cell. This is done using an optimization algorithm that mainly takes thermal properties into account. It is shown that a prediction of the temperature in the cell is possible with a deviation of less than 3° C for a current modulation of up to 10 kA after 24 hours. By means of this model, a statement about the process stability in the case of possible future current modulations can be achieved. By simulating different operating modes at high current modulation, the maximum possible current modulation can be determined.