Stress Classification due to Refractory Expansion

[corus]

A vessel is lined with refractory which expands and exerts a pressure on the outer shell. As this pressure is caused by thermal expansion of the lining, and as such is self limiting, is the stress in the shell classed as secondary, or as the stress is due to an external force from the refractory expansion (and not due to internal differential thermal expansion within the shell itself) is the stress in the shell classed as primary?

corus

 

[TGS4]

corus,

It has characteristics that are both primary and secondary in nature. It depends on the magnitude of the differential thermal expansion. If you classify the stresses as primary (a conservative classification), does your analysis pass? If it doesn't pass, then I would suggest that you perform an elastic-plastic analysis, and then no stress classification is required.

 

[corus]

Thanks for the reply TGS4, but I can only see the benefit of an elastic-plastic analysis if the stresses redistribute themselves after yielding, perhaps like at a discontinuity, and I can't see how that would happen.

I've checked the design standards but they only refer to stresses within the 'cladding' and strangely classify them as peak stress. Other references on-line say that if the load is external then it's primary, and some AISC document which referred to them as secondary. British standards don't seem to mention this case at all and ASME standards I don't have at present.

The case to be considered is a one-off situation and stresses have in similar situations exceeded yield, sometimes with no damage, sometimes causing rupture.

corus

 

[TGS4]

I've examined similar situations to yours before, and I can confidently say that an elastic-plastic analysis would solve your issues.

Let's say, for example, the combination of stresses due to internal pressure and the effects of the refractory cause stresses in excess of yield in the vessel wall. However, with a small amount of plastic deformation in the vessel wall, the refractory loads decrease significantly. Then, you could have sufficient margin against burst. Maybe even enough to pass an elastic-plastic analysis.

For the vessel wall itself, a 2-D axisymmetric analysis should be sufficient - probably a day or so worth of work.

I would also seriously investigate the crushing strength of the refractory. It may crumble well before it over-stresses your vessel wall. Or, you may find that it would be better to specify a very flexible insulation material (a kaowool or similar) between the refractory and the vessel wall, which will be squished by the differential thermal expansion, but not transit the load to the vessel wall...