It is a testing example for our methodology for

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: objects. It is a testing example for our methodology for constructing the initial set-up for a SPH simulation. Neighbour Search C-shaped Mould The summations in the SPH equations are over all particles within a radius 2h of ra (the position of particle a). One of the basic requirements of an SPH computer code is the identification of the neighbouring particles of a given particle. In contrast to grid-based method, where the locations of neighbouring grid-cells are found directly, SPH is dependent on fast techniques for finding the neighbouring particles that contribute to SPH summations. Without an efficient method, it will degrade into a direct summation and computational time will scale as N2, where N is the number of particles. One approach is to use a searching grid with linked lists to store the particles within each cell. This method works well when a constant smoothing length is used and is described in detail in Hockney and Eastwood (1988). Figure 1 shows the geometry of the die used in previous experimental and numerical simulations (Ha, et al., 1998 and Cleary, et al., 1999). It is 50 mm long, 20.9 mm high and 20 mm deep. It is connected to the runner by a gate of width 2.9 mm. The width of the vertical sections is 4 mm and the width of the connecting horizontal section is 4.9 mm. A resolution of 2.6 particles per millimetre was used giving a total of 208,054 particles for the simulation. The fluid speed through the gate was 0.62 m/s. Two perspective views of the filling pattern at 6 different times are shown in Figure 2. The fluid particles are shown coloured by their speed. The flow in the xy-plane is broadly similar to that predicted by the 2D SPH simulation and observed in the 438 vertical cylindrical cut-outs. After collision with these cut-outs and central cylindrical core, the fluid begins to flow vertically up into the upper part of the die cavity. At the time the upper part of the runner is filled by a reverse flow reflecting from the surface above the gate back towards the shot sleeve. Note the rapid acceleration of the fluid particles as they pass through the gate produced by the converging surfaces of the runner. matching experiments (Ha, et al., 1998). In particular, the separation from the right angle bends and the back filling down the left side of the first vertical section are reproduced. gate mould runner Visualisation of particle systems in 3D is quite difficult with the three dimensionality of the flow difficult to perceive when using just the particles. An alternative is to construct iso-surfaces based on the underlying particle data. A substantial difficulty is that the data is nonstructured with no inherent geometrical relationship between particles. In principle, a given particle can be anywhere in the computational domain. Figure 5 shows a rendered surface mesh coloured by fluid velocity which has been created from the particles and shows the fluid distribution as the fluid front passes through the gate and enters the die cavity....
View Full Document

This note was uploaded on 10/30/2013 for the course ENG 101 taught by Professor Cheng,m. during the Fall '13 term at Nevada State College.

Ask a homework question - tutors are online