| Applies To | |
| Product(s): | AutoPIPE |
| Version(s): | ALL; |
| Environment: | N/A |
| Area: | Modeling |
| Original Author: | Bentley Technical Support Group |
| Date Logged & Current Version |
Feb. 2017 11.00.00.22 |
Problem 1:
Why are AutoPIPE's Fluid transient loads different from other programs?
Example:
An in house software application was developed that address reflection, losses and phase change. But when modeling with this in house application the piping assumes the wave is reflected back. Also, small branch lines that are connected are not considered. Some folks in our organization believe this would produce erroneous and conservative results. People argue that when one includes all the branches and so forth, then the reflected waver is much weaker and the resulting load will be much smaller. Their argument is based on a study performed by another company some 35 years ago where they had decided not to consider the reflected wave in their designs. Unfortunately this study was never published.
We compared AutoPIPE generated fluid transient forces with our in-house program results and we found that our program results are 3 to 4 times higher than AutoPIPE generated values. why?
Solution:
Hard to answer questions when Bentley development cannot review / validate calculations performed by another application. AutoPIPE's Fluid Transient module has documented calculations used by the Fluid transient module. If there are differences between the two, it could indication that those limitations of our module are important in this system or that the other method is too conservative. No difinitative answer can be made with out a thorough review of the other application.
The reality is that each branch will reduce the effect of the wave 'chopping it up' into smaller waves, but each time this happens there will be a reflection that AutoPIPE does not consider. These reflections will be much smaller than the full reflection which the other software may or may not be considering, Therefore, it isn't hard to imagine that the real answer is somewhere in between.
Problem 2:
I am running a force-time history analysis on an onshore pipeline. Transient hydraulic analysis is being done in Third party software from which force-time history data is generated, as well as pressure & mass flow rate (as function of time).
For comparison, I also used AutoPIPE's fluid transient module by adjusting the mass flow rate input to give the Jowukowski pressure equal to that calculated by Third party software (as per the AutoPIPE manual). However, the resulting forces are massively different to that from Third Party software.
Now, I can also play with the rise time (make it shorter) as this doesn't affect the Joukowski pressure but does affect force magnitudes. But not sure if this is suitable.
Yes, I could use the Third Party software results directly but
a) requires pre-processing to get it into AutoPIPE format
b) wanted secondary check. So just exploring using AutoPIPE only with necessary tweaks initially.
Can you please confirm whether the Fluid Transient method only computes the unbalanced forces at bends etc. due to the change in pressure? i.e., it does not account from the effect of fluid momentum change. This may be the cause of the difference. Or, does the pressure change account for this & so should be captured...)
Solution:
AutoPIPE considers only the pressure differential when calculating the generated forces. Many factors accounted for in a detailed fluid transient analysis are not included in AutoPIPE’s method. As a result, the approach is approximate and generally conservative. AutoPIPE’s Fluid Transient methodology primarily evaluates pressure‑induced unbalanced forces and does not model fluid momentum (convective inertia) effects in the same manner as more advanced hydraulic transient solvers.
See Also