Artificial Intelligence for Shrinkage Porosity Prediction

The wider adoption of artificial intelligence in strategic industrial sectors can help to improve the invested time in optimizing manufacturing processes such as aluminum casting. The latter is one of the most used processes to form metals like aluminium. A casting part might contain several defects that can affect its resistance. Shrinkage porosity is one of the major anomalies that should be avoided. For this purpose, rounds of numerical simulations should be performed with operating on a selection of parameters in order to minimize shrinkage porosity in the casting part. Depending on on the complexity of the study case and the needed accuracy, these approaches are generally expensive in terms of time. In this work, two solutions are proposed to minimize this time-cost. The main purpose of these procedures is to assist users at optimizing porosity distribution according to their application using artificial intelligence. The two proposed methodologies are based on Decision Trees (DT) coupled with Model Order Reduction (MOR) and Optimal Transport (OT) combined with shape morphing. Our approaches are tested on two study case and promising results of shrinkage porosity prediction are presented. These solutions can contribute to open perspectives for more data-driven solutions that optimize the time-cost in the design stage.