Hi, I do have a question regarding the pressure wave reflection time from my simulation result, it seems double from 2L/a that obtained from calculation. Does this means that my mode setup is incorrect?
Based on time histories and profile result, the complete wave reflection (back and forth) is around 15s (normal load rejection 100s second closure Pelton needle valve). However, my preliminary calculation shows, 2L/a = 2* 3308m / 892 m/s = 7s.
Amazingly when I run for rapid closure (0.1s needle valve closure), the wave reflection time is almost the same as 2L/a. Hence, I am confused with the discrepancies, as fas as I understand 2L/a is constant as physical conditions are not changing.
Hope to get your advice on this. Thank you.
Sin, the characteristic time is more of a physical property of the network since the length and wave speed are essentially fixed physical parameters. A transient wave during a transient simulation in HAMMER will travel based on the wave speed. Note that the wave speed or length can be adjusted by HAMMER if your timestep is too large and/or pipe lengths are too short. Check the wave speed or length adjustments in the pipe Flextable. See more here: Understanding length/wave speed adjustments and their impact on results
As for the timing you are seeing in the results and the reason for the change, it might be better to visualize them with a profile animation, covering the complete pipe length. Animate the profile results to get a better understanding of the wave propagation. Depending on what is happening in your system, the interactions may be more complex. Additional boundary conditions could be forming such as a vapor or air pocket, or more complex interactions with reflections from branched connections.
If this does not help, please provide a copy of the model and reply here with details steps to reproduce and observe the specific results you are looking at.
Regards,
Jesse DringoliTechnical Support Manager, OpenFlowsBentley Communities Site AdministratorBentley Systems, Inc.
Hi,
Sure. I had uploaded two scenarios for your reference. Normal closure and slam shut case for Pelton turbine.
Thanks for sending the model.
If you animate the profile path while keeping the HGL time history graph open at the same time, you will see that the characteristic time is reflected not between the HGL peaks but rather from the peak to the low point. For example for the s1a "normal closure" model if you start animating at 141 seconds, you will see the wave travels to the end of the system by about 145 seconds and then reflects back and reaches the origin at around 148 seconds (at a low point in the HGL graph) for a total travel time of 7 seconds as expected (again, based on the pipe length and wave speed).
Waves will reflect with a different sign depending on what they reflect off of. See: HAMMER Transient wave reflection behavior for different boundary conditions
In the "s4b" model with the instant closure, the conditions are more chaotic (as expected) but if you animate the profile you can again observe the wave taking 7 seconds to travel from the transient origin to the end of the system and back (between 100 and 107 seconds).
Here are some other observations:
Hi Jesse,
Thanks a lot for the comments and recommendations.