(The information on this webpage was presented at the SAMPE conference in Paris in April 1999)

Closed mould resin transfer moulding (RTM) techniques have been successfully applied for small products in large series [1, 2, 3]. This has proven the advantages of a closed mould infusion technique (better workshop conditions, less styrene emission from polyester or vinylester resins, better quality products and less labour costs). The application of a closed mould infusion technique for large products in small series (i.e. boats) is hardly practised, although the technology exists!

Boat builders and other composite manufacturers often want to switch to a closed mould (i.e. vacuum infusion). However, the risk of failure is often considered too high. This is mainly due to a lack of knowledge of the technology and subsequent difficulties in determining proper infusion strategies and a lack of process control tools. When treated properly the main practical production problems can be solved as shown at the Conyplex yard where the composite production implemented vacuum infusion fully. (Please see this webpage for the infusion of a Contest 55 hull [4]).

This paper will elaborate on the background and principles of infusion strategies. Before going into details about these principles a brief introduction into the resin infusion technology is given. The main variations in the technology and the most important processing parameters are discussed. Finally, the infusion method for the hull, the stiffeners in the keel and the deck of the Contest 55 is described in detail. It will show what kind of product characteristics can lead to disturbance of resin flow.

Vacuum infusion technology
RTM, VARI, SCRIMP, infusion, vacuum bagging [5, 6, 7, 8, 9]; there are many terms and many abbreviations that describe a process that is basically very simple:

(1)

With

j    : porosity of the reinforcement
k   : permeability of the reinforcement
h   : viscosity of the resin
l    : flow distance (length of the strip)
dP : applied pressure difference (constant during the infusion)

Equation (1) clearly shows how different parameters influence the filling time. The parameters are divided into material, product and process properties:

Material properties

Process properties

The Contest 55 hull is both a long (16.4 m), wide (4.5 m) and high (2.5 m) structure to inject in one shot. The height results in a hydrostatic pressure difference which results in a lower driving force for the infusion. There are three basic variations possible on the edge infusion strategy to distinguish:

The infusion downward is not preferred for two reasons. Firstly, air bubbles in the resin and laminate will be entrapped more easily and secondly, there is higher risk on the occurrence of dry spots due to racetracking of the resin through highly permeable runner channels. The infusion upward can only be carried out with sequential infusion channels (otherwise infusion time will be too long). This however, would result in about 10 sequential infusion channels each of which need to be controlled and opened at the right time. Based on the requirement to have a simple process there was a clear preference for an infusion with a grid of infusion channels in which the majority of the channels would allow for a sideways infusion. This would allow for one single resin inlet port (and consequently also one resin tank level to control). This grid would consist of a main infusion channel running from the stern to the bow (via the keel) and branches of channels from the main channel in the keel to the flange (with the deck). The overall infusion channel pattern is shown in figure 2.

Figure 2. The infusion strategy and position of the infusion channels.

Permeability differences in the keel

Based on the overall chosen infusion strategy we must verify that there are no local details, which can cause the formation of dry spots or else, disturb the infusion. The Balsa core was sufficiently permeable. Therefore there was no additional feeder material required on all laminates containing Balsa. However, the laminate used in the keel would have a low overall permeability. In addition there is a substantial thickness change. This could cause dry spots on the keel, which were clearly shown using flowfront simulations. Figures 3 and 4 show the simulations done without (figure 3) and with (figure 4) additional feeder material on top of the reinforcement. A small scale experiment was carried out to validate this simulation result. A small part of the hull laminate was injected on a flat plate. This part consisted of the keel laminate (partly with and partly without additional feeder material) and the balsa sandwich laminate shown in figure 5. It was injected using a main infusion channel at the edge of the keel laminate and an infusion channel branch in the middle of the laminate. The carefully registered flowfront proved the occurrence of a dry spot on the part without the feeder material and the absence of a dry spot on the part with the additional feeder material.

Figure 3. Part of the hull with dry spot formation at the location of the keel.

Figure 4. Part of the hull without dry spot formation.

Figure 5. Test infusion set up and resulting flow front.

The stiffeners in the keel
The stiffeners in the keel of the Contest 55 are designed to achieve good mechanical performance in combination with easy production, minimising space requirements and improving the accessibility of the hull for building the interior. This lead to series of wide stiffeners with gaps, sufficiently wide to allow for the bolts of the keel. The overall design is shown in the figures below (figure 8 and 9). To determine the infusion strategy of a part with many edges one must consider that every edge can lead to a highly permeable channel, causing resin racetracking. The most favourable situation will occur when the resin flowfront is progressing always parallel to the runner channel. This is shown in figure 6 and 7. When the resin flowfront is perpendicular with the runner channel the flowfront will be disturbed and the resin will flow much faster through the runner channel (shown in figure 6). In figure 7 it can be clearly seen that a runner channel, parallel to the flowfront will not cause disturbances of the resin flow front.
When using a simulation package to identify the filling pattern of a complex structure one must realise that runner channels can occur. This should be included in the simulation of the filling pattern.

Figure 6. Disturbance of resin flowfront due to a runner channel perpendicular to the flowfront.

Figure 7. No disturbance of resin flowfront due to a runner channel parallel to the flowfront.

The strategies of the infusion of the stiffeners are compared:

Figure 8. Filling pattern of stiffeners, infusion from the edge to the inside.

Figure 9. Filling pattern of stiffeners, infusion with main channels with branches.