Predicting the leak-rate of seals

When the space shuttle Challenger exploded 73 seconds into its flight on 28 January 1986, the fault was soon traced to a faulty seal in one of the vehicle’s solid rocket boosters. Engineers had noted previous problems with these seals but predicting how bad it was likely to get would have been near impossible anyway.

Seals are simple devices for making a joint leak-tight. But although simple, predicting their performance is surprisingly hard, says Bo Persson and Boris Lorenz at the Institute of Solid State Research in Julich, Germany. The problem is that leak rates depend on the contact made between the two surfaces in the seal and this depends on the surface roughness which may vary over wide range of scales.

So predicting how liquid may percolate through a seal depends on how well this surface roughness can be modelled and until now, the various theories attempting to do this haven’t worked well.

Persson and Lorenz say this is mainly because the models haven’t taken into account the elasticity of the surfaces, which obviously influences the contact patterns that form between the two surfaces.

When this elasticity is taken into account, everything changes. And the German pair prove it today with a neat experiment in which they measure the flow rate through rubber seal in contact with sand paper. Their results exactly match the predictions of the new theory.

All of which may make it possible to design better performing seals. And who wouldn’t want that?

Ref: arxiv.org/abs/0904.2988: Leak-Rate of Seals: Comparison of Theory with Experiment