End Cap Pressure Simulation

[PetkovStoyan]

Hi all, I'm new to the forum and SW Simulation so please help me on this.

In our company we use very often custom made end caps for hydrotest purposes of boiler equipment. I've attached a cross section for the most common type we use. We have to calculate the thicknesses of these end caps so they won't fly away during the hydrotest with pressure reaching 300 bar. This partucilar end cap can withstand 250 bar (tested in practice!). The thickness is calculated by hand and it's working. I want now to use solidworks simulation to optimize and upgrade our calculations. I've modeled the cap and started study. The cap is fixed on the right open face and I've applied pressure type external load 250 kgf/cm^2 which is about 250 bar. The results show max. von mises stress sagnificantly above the yield strenght of a typical boiler building materials. What I've noticed is that the area with the highest stress is on the surface (the section view shows no red areas in depth). Even though the stress is above yield strenght and therefore the cap should fail (which it doesn't in practice) according the simulation results. So what am I doing wrong? Has anybody expirience with similar models?

 

[corus]

You need to assess this against pressure vessel design code limits. For this case your maximum is at a stress concentration feature and would be classed as a peak stress and would require a fatigue assessment if the loads were repeated. Normal yield criteria apply to the average stress across a section and the linearised bending stress, which excludes the peak stress component. The design codes have further details and what limits would apply for a hydrotest case.

 

[PetkovStoyan]

Thank you for the reply. One of the tutorials in SW simulation was for pressure vessel design and in this design study is used von mises stress. That's why I tried this way. Otherwise we use EN12952-3 : Design and calculation for pressure parts.

So is there a way to plot the average stress across a section? I tried with directional stress values along axis "SZ", but I'm not sure this is what I'm looking for. I'm asking just to get some experience with SW simulation.

 

[corus]

Typically you define a SCL (Stress Classification Line) to derive these nominal membrane and bending stresses. Here's a link that explains it

http://www.pveng.com/FEA/FEANotes/NutsBoltsStress/NutsBoltsStress.php

. ASME also has guidelines on where these SCLs should be placed.

I would also recommend checking for mesh convergence as the contours in the model seem a little haphazard for axisymmetric results that should produce smooth contours in the circumferential direction.

 

[PetkovStoyan]

Thank you very much for the helpful guidelines. This is more than enough to get me started. I'll seek for your advice if needed.