Hot Spot Stress Definition for FEA Analysis 2

[dcoppo]

There seems to be several criteria for "hot spot stress". During linear finite element analyses, small areas often appear which can be defined as "hot spot stress" and are excluded from limit state stress levels for design. One definition from a colleague from the American Bureau of Shipping defines the limit for a hot spot stress as either 2 X the yield strength of the material or the ultimate tensile strength of the material. I have heard that the ASME boiler and pressure code has another version as well as Det Norske Veritas (DNV). Does anyone know of another source for this definition?

 

[flash3780]

The assumption is that small stress concentrations will yield and strain harden. The best bet, especially when dealing with stresses over yield, is to look at a fatigue curve for the material. If a -3sigma fatigue curve predicts that the component will last through the service life, your part is probably fine.

Be sure to realize that the stress you're looking at isn't real (although it should work with a standard SN curve). Since you're above yield, you you either need to do an elastic-plastic FEA (best) or do a Neuber correction to find out where you actually are on a stress-strain curve.

Of course, you need to be certain that the FEA is accurate to begin with. Often stress concentrations involve small geometry that is difficult to adequately mesh. Sub-models are sometimes required.

 

[corus]

The definition for 2x yield is classifying the hot spot stress as a secondary stress, which is a little conservative. The usual definition is to describe the stress as a peak stress and assess it via fatigue design limitations. Most pressure vessel design codes, whether american, norwegian, british or whatever, will give a description. The question remains though is how big is a 'hot spot' before it's classed as something else. Some work by Seshedri has gone some way to answer this for pressure vessels with thermal hot spots.

Tata but not yet tara