A high strength-to-weight ratio makes polyurethane foam ideal for moulding in many engineering applications, but the unpredictable nature of foaming makes them difficult to work with.
In a bid to ensure that moulded items are as perfect as possible, two researchers from German chemical company Bayer are employing computer simulation methods.
Peter Wulf and Bolko Raffel are working on the behaviour of polyurethane foam — used widely as an insulating material and for impact-absorbing parts in the automotive industry — as it is injected into a closed mould. The behaviour of the foam as it is simultaneously pushed into the mould and expands is very complex, and can lead to problems.
‘A foaming polyurethane system not only expands its volume considerably during the reaction time, but also radically changes its properties,’ said Wulf. ‘At the beginning it behaves like a liquid, but it becomes more and more viscous and then solidifies.’
This can lead to severe production problems. For example, if the foam expands past an obstacle in the mould, it can form air bubbles as the viscous fluid tries to close behind it. These are pushed along with the front of the material, but the bubbles remain when the foam solidifies. This reduces its strength and impairs appearance.
Wulf turned to computational fluid dynamics (CFD) to solve this problem, using the system to model not only the behaviour of the polyurethane foam as it flows and expands around the mould, but also of the air displaced by the polymer. ‘We wanted to know at which points the air was entrapped and where it was transported to,’ he said.
The simulation relied on a detailed knowledge of the polyurethane’s properties, which Raffel investigated. His team mixed together in cardboard tubes the two component parts of polyurethane, along with the various additives used in particular applications.
Ultrasonic sensors provided information on how fast the various blends expanded, while thermocouples inside the tubes told the team how much heat was released. this indicates the progress of the polymer and foam-forming reactions, and correlates to the expansion rate. He also developed a sensor linked to a vibrating paddle which measured the pressure exerted by the foam, which showed how the flow properties altered as the material changed from liquid to viscous foam to solid.
Wulf then used Raffel’s information to set up a simulation for the behaviour of the foam, linking 16 computers to process the huge mass of data. To test this, the pair built a transparent mould and filled it with polyurethane, filming the progress of the expanding foam as it flowed into the enclosure.
They then compared the film with an animation derived from the simulation program — and found that it predicted where and when air bubbles formed in the mould.
Wulf believes that the simulation could prove useful for manufacturers using polyurethane foams to optimise their mould designs before going into production. ‘The program can make a valuable contribution to reducing costs, particularly in the automotive industry,’ he said. ‘A mould for manufacturing polyurethane parts can cost tens of thousands of euros. In such cases, it makes good sense to avoid design faults right from the very beginning.’