Fillet weld fatigue assessment acc. to VIII-2 5.5.5

[Raul Villalba]

Hello,

I have some doubts about weld modelling and stress evaluation for weld fatigue assessment of a non-symmetric loaded fillet weld acc. to ASME BPVC.VIII.2-2021 5.5.5.

I have created a sub-model (weld included) meshed with solid elements. I have used a stress classification line (SCL) normal to the expected crack of the weld toe to obtain the linearized structural stresses that will be used in eq. 5.47-5.51. I have followed Figure 5-A.1 to place the SCL and restrict the element set to only elements belonging to the plate similar to Figure 5-A.11. However, stresses are slightly higher in this zone due to mesh contact and thermal stress.

Q1. Where should I place the SCL if the toe stress distribution is not homogeneous? Mesh contact between weld and components may locally increase stress.

Q2. Is it advisable to use another method for non-symmetric loaded fillet weld? This method seems strongly mesh dependent.

Q3. Could it be possible to assess a weld trough 5.5.3 if it is well defined with a fine mesh? (5.5.3.2 step 4(a) accounts for weld effect).

Thank you in advance.
Warm Regards.

 

[TGS4]

Your post and your attachment are very confusing and seems to be missing a substantial amount of context. Please provide that additional context.

Although the structural stress method is advertised as mesh insensitive, you still are required to have a mesh sufficiently adequate that you can say that the linearized membrane and bending stresses of the component stress normal to your hypothetical crack plane is independent of the mesh.

 

[Raul Villalba]

The equipment is a parallelepipedal double chamber pressure vessel which is assessed with ASME VIII-1 through App 46 - Design by Analysis (Div 2 Part 5). The main loads are the chamber pressure, jacket pressure and thermal. The main material is SA-240 316L. Tmax = 150ºC.

My thread involves two welds (see attached file):
- Weld 1: Similar to a fillet weld between the door plate (end) and a web reinforcement.
- Weld 2: Fillet between the inner chamber body and a web reinforcement used to connect body with jacket.
Those welds work as a fillet weld between the base and the web of an I-profile beam. Both weld throat sizes are around 4 mm.

Firstly, cyclic counting method is applied in the stress history (6 stress components) of a zone close to the weld of interest. Then, membrane, bending and peak is obtained acc. to 5-A.6 with a SCL normal to the hypothetical crack plane (weld toe and root) for each maximum and minimum in the cycle. Those values are then used for the following calculations acc. to 5.5.5

Let me clarify the previous questions:
- Q1: There is a significant stress distribution along the longitudinal direction of the fillet weld. For weld 1 the maximum stress is on the middle and for weld 2 is on the end. Where should I place the SCL? For weld 2, does it make sense to place it on the end? Would you recommend to sub-model and refine weld 2 to avoid singularities?
- Q2: I will verify mesh with a convergence analysis. Do the nodes of the weld and the plate need to be coincident?
- Q3: Could I use chapter 5.5.3 to assess those welds? What does it mean "controlled smooth profiles" in 5.5.5.1? If possible, how I may proceed?

Let me know if something is not entirely clear.