Experimental and numerical investigations of mould filling of thin wall horizontal aluminum castings

The application of Computational Fluid Dynamics (CFD) in the foundry practice is considered an essential tool to achieve a better understanding of the occurrence of casting defects introduced during mould filling of a casting. CFD is believed to be advantageous in predicting possible defect locations, pouring times, and the lowest possible casting temperature without the risk of cold shuts. Furthermore CFD may be used to optimize pouring system geometries. Contrary to experimental studies in, e.g., cold water similarity models, computer simulations allow for the flexible modification of the casting geometry and the pouring system, and for the inclusion of physical effects such as solidification and heat transfer. Computer modelling may lead to an improved recovery in cast shops without expensive trial-and-error procedures. CFD simulations with the commercial FLOW-3D software were done on the mould filling conditions and gating systems for the manufacture of a horizontal horseshoe. These simulations were validated against data obtained with various experimental techniques. In the paper the following cases are considered: flow patterns, i.e., the transient velocity field during mould filling, metal-air interface shapes, i.e., the position of the free surface and its shape due to surface tension; freezing lengths, i.e., the prediction of cold shuts in horizontal thin-wall casings. Predicted flow patterns and interface shapes are overall agreement with experimental observations obtained from video registration and Particle Image Velocimetry (PIV). Freezing lengths that are experimentally found to depend linear on the pouring temperature, can be predicted with an accuracy of about 15%.