Karloss12
Mechanical
- Nov 18, 2010
- 44
I have a regular full pen nozzle in an ellipsoidal head. It has been designed to pass PD 5500 nozzle reinforcement calculations comfortably. It is subject to internal pressure.
I have some queries about calculating the Stress Range for the fatigue assessment in PD 5500 Appendix C. I’ve used different commercial software, and do it yourself methods which have all resulted in very different stress ranges and therefore very different fatigue life.
In Appendix G there is a clear process of calculating rho (a non dimensional geometry constant) and the membrane stress in the head and then looking up the appropriate figure to find the stress concentration factor ‘SCF’ and hence the stress range to use in the fatigue calc. This Method results in a direct principal stress of 180MPa and a life of 28000 cycles. This is the method that I have used and PVELite also gives the same result.
But Appendix C does say FEA can be used to find this stress range. (For use with complex geometry etc.)
So I also decided to use the Compress FEA nozzle Add-on to calculate the stresses. It uses 2D shell elements. As I understand it, these shell elements aren’t much good for calculating peak stress so as a result the stress it returned was 140MPa. As it turned out the programme then went on to do many fatigue assessments to various codes. For PD 5500 Appendix C it says the nozzle is good for 400000 cycles.
I also modelled the head and nozzle in ANSYS. I used quadratic tetrahedral elements with mesh refinement etc. I modelled the weld as a smooth radiused corner. No matter what size the radiused corner is I always get a high localised stress at the toe of the weld of roughly 400-500Mpa or about 4000 cycles according to PD 5500.
Soooo.....
there is PD 5500 Appendix G that returns a stress maximum principal stress of 180MPa and life of 28000 cycles. (PVElite calculates this with detailed working.)
There is Compress that does an FEA analysis that returns 135MPa and a PD 5500 life of 400000 cycles (With no working at all).
And there is my manual FEA that returns a principal stress of about 450MPa and a life of 4000 cycles.
The commercial Appendix G (PVElite) and FEA (Compress) returned very different stresses and fatigue results. The nozzles were modelled the same, so I don’t see this as a case of rubbish in = rubbish out. Does anyone have any comments on this? Which one is the most correct? Are these ‘Nozzle Pro’ and ‘Compress Nozzle FEA’ etc actually any good?
With respect to my FEA attempt see attach image, is the high local stress on the smooth radiused weld a peak stress or a secondary bending stress? Peak stresses aren’t used in PD 5500 so if it is a peak stress, how do I separate it so that the primary + Secondary stress can be extracted from the FEA result to be used in the Fatigue calc or in other PD 5500 calculations?
If it is a valid primary + secondary stress then by using 3 different analysis techniques I am calculating fatigues lifes of between 4000 and 400000 cycles!!?!!
Any thoughts would be apprieciated?
I have some queries about calculating the Stress Range for the fatigue assessment in PD 5500 Appendix C. I’ve used different commercial software, and do it yourself methods which have all resulted in very different stress ranges and therefore very different fatigue life.
In Appendix G there is a clear process of calculating rho (a non dimensional geometry constant) and the membrane stress in the head and then looking up the appropriate figure to find the stress concentration factor ‘SCF’ and hence the stress range to use in the fatigue calc. This Method results in a direct principal stress of 180MPa and a life of 28000 cycles. This is the method that I have used and PVELite also gives the same result.
But Appendix C does say FEA can be used to find this stress range. (For use with complex geometry etc.)
So I also decided to use the Compress FEA nozzle Add-on to calculate the stresses. It uses 2D shell elements. As I understand it, these shell elements aren’t much good for calculating peak stress so as a result the stress it returned was 140MPa. As it turned out the programme then went on to do many fatigue assessments to various codes. For PD 5500 Appendix C it says the nozzle is good for 400000 cycles.
I also modelled the head and nozzle in ANSYS. I used quadratic tetrahedral elements with mesh refinement etc. I modelled the weld as a smooth radiused corner. No matter what size the radiused corner is I always get a high localised stress at the toe of the weld of roughly 400-500Mpa or about 4000 cycles according to PD 5500.
Soooo.....
there is PD 5500 Appendix G that returns a stress maximum principal stress of 180MPa and life of 28000 cycles. (PVElite calculates this with detailed working.)
There is Compress that does an FEA analysis that returns 135MPa and a PD 5500 life of 400000 cycles (With no working at all).
And there is my manual FEA that returns a principal stress of about 450MPa and a life of 4000 cycles.
The commercial Appendix G (PVElite) and FEA (Compress) returned very different stresses and fatigue results. The nozzles were modelled the same, so I don’t see this as a case of rubbish in = rubbish out. Does anyone have any comments on this? Which one is the most correct? Are these ‘Nozzle Pro’ and ‘Compress Nozzle FEA’ etc actually any good?
With respect to my FEA attempt see attach image, is the high local stress on the smooth radiused weld a peak stress or a secondary bending stress? Peak stresses aren’t used in PD 5500 so if it is a peak stress, how do I separate it so that the primary + Secondary stress can be extracted from the FEA result to be used in the Fatigue calc or in other PD 5500 calculations?
If it is a valid primary + secondary stress then by using 3 different analysis techniques I am calculating fatigues lifes of between 4000 and 400000 cycles!!?!!
Any thoughts would be apprieciated?
