Vinsamlegast notið þetta auðkenni þegar þið vitnið til verksins eða tengið í það: http://hdl.handle.net/1946/2931
The investment of an aluminium smelter casthouse is a very long and expensive process, therefore it is crucial to carefully model it before the beginning of construction. There is not a lot of research aimed specifically at modelling for aluminium casthouses and this thesis is aimed at fulfilling that need. The outcome is a three-way approach to model a casthouse and using this approach will help management teams to focus on the right path during the modelling process and therefore save money by choosing the right model for them at the outset and use the model selected to avoid costly errors during the construction phase. In the first step research is done on the aluminium market, equipments available, and the quality of metal. That information is used, along with management constraints, to create a binary integer program which is used to select four models. The next step is to simulate those four models and in the simulation the models are tried with different inputs and efficiency on the machines. The sizes smelters that will be looked at are 180,000 – 450,000 tons a year and the efficiency on machines range is from 75% - 95% on the machines. The last step is then to choose the best model of the four, using the AHP method and the information obtained from the simulation. The variables selected in the AHP method are cost, value, metal backlog, throughput, utilization of machines and product mix.