Nonlinear analysis of stress concentrations in creep realm

[McT178]

I am running a nonlinear analysis of a stainless steel structure at 1450 F, and need some guidance (I am new to nonlinear FEA). The design is governed by the ASME allowable stress which is lower than the yield stress; so my stress concentrations after the nonlinear analysis are lower than the yield but are higher than the allowable. How do I interpret these results? Are these areas susceptible to creep failure or will creep act as yielding and redistribute the loads?

 

[TGS4]

You're definitely in the creep regime, depending on the material. Over time, creep will act to redistribute the stresses, but during the redistribution, creep damage will occur.

In the creep regime, I would not ignore creep effects, particularly for longer time periods (>100 hours).

 

[McT178]

Do you think a fatigue analysis is required to determine the creep effects?

 

[TGS4]

Fatigue and creep are different failure mechanisms. For starters, throw away the focus on the types of analyses you are doing. Start fresh with a concentration on the failure modes. Think about all of the possible/potential failure modes, how they act, and what drives them. Then direct your analyses into the evaluating those.

 

[McT178]

How are stress concentrations or localized stresses analyzed for creep failure? I can understand how a stress through a cross-section would cause failure due to creep, but it seems that stress concentrations would act like localized yielding and redistribute the stress as creep progresses. Is there an asme guideline for this?

 

[normet]

For creep hot spots, the risk is a localized fracture. You need to do a fracture mechanics analysis to determine expected life. You can use the method in API 579, or British Standard PD7910. I would assume the determined locallized stress value for my analysis as my experience with creep analysis is that it is always not conservative.

 

[TGS4]

Stress concentration locations ARE where creep damage initiates in complex geometries. If you perform an assessment in accordance with API-579, using Project Omega data for your creep model, then you will find this. Code Case 2605 has a procedure, also, to do this, but not for your material.

 

[McT178]

Thanks normet and TGS4; I will look into API-579.

 

[mcclain]

"so my stress concentrations after the nonlinear analysis are lower than the yield but are higher than the allowable"

Are the stress allowables applicable for stress results from an elastic - plastic analysis?

Regarding stress concentrations, remember that with an elastic plastic analysis, stress concentration goes down in the plastic regime but strain concentration should increase. You need a rigorous model of the notch to address plastic strain concentration. Initiation of rupture will be due to the concentrated plastic strain whether its fatigue or creep.

 

[McT178]

"Are the stress allowables applicable for stress results from an elastic - plastic analysis? "

After thinking more about this, I do not think that they are. I think I need to do Viscoplastic analysis, but I do not know where to get those properties from. It seems very hard to find material properties at elevated temperatures.

"Initiation of rupture will be due to the concentrated plastic strain whether its fatigue or creep."

Does this cause crack propagation which migrates out of the concentration?

 

[normet]

For clarification, permitted stress values are applicable regardless of the method employed to determine stress in a material as the permitted stress values are a Code consideration for safety. The Code also recognizes "shakedown" which is localized plastic flow. What the Code does not permit is gross deformation that causes a structure to deform outside of permitted tolerances.

Creep cracking will propagate only is the material toughness is less than the stress intensity caused by the cracking. As you are operating at a very hot temperature, toughness is probably sufficient to resist crack propagation unless you have material embrittlement.

 

[TGS4]

For clarification, permitted stress values are applicable regardless of the method employed to determine stress in a material as the permitted stress values are a Code consideration for safety.

Well, not really. Allowable stress-basis design is applicable to elastic analysis. When you start into limit load or elastic-plastic, there are different bases. The design margins, in the latter, are considered in the load factors, rather than on the stress side.

Furthermore, creep cracking is a different failure mode than cracking otherwise. Creep fracture will initiate and propagate based on creep damage. It not really a factor of the toughness.

 

[McT178]

What would be a general outline for analyzing a stress concentration in the creep regime from an elastic analysis? Is it outlined well in APC 579? I have not been able to get access to this code.

 

[normet]

API 579/ASME FSS-1 is in part a cook book standard. For simpler analysis, you follow a step by step procedure that is well laid out. The step by step approaches are necessarily conservative and do not cover all circumstances. API 579 spells out the limitations of each procedure.

TGS4 is correct in the comment about creep fracture propagating based on creep damage, however, when the extent of creep damage is localized to a hot spot, further propagation depends upon the fracture toughness of the material (or possibly other damage such as fatigue). Both API 579/ASME FSS-1 and PD7910 give methods for handling all.