Time History Analysis in Caesar II

Dear Richard / Dave ,

I have a query regarding Time History analysis for water hammer .In Caesar II we have two options by which we can enter the Time Force profile ( Time History).
1. I can enter data in form of time force profile directly through the "Enter/Edit Spectra Data Tab" .
2. In second option we can feed data by converting the data from the form of Time force profile to Frequency -DLF profile with the help of DLF spectrum generator.

Which method is suitable for Time history analysis ????.
I think only first method(through ENTER/EDIT SPECTRA DATA TAB) is valid for time history analysis. As in the second method the Time-Force profile is converted in to DLF -Frequency profile . During time history analysis Caesar II consider it as FORCE multiplier-Time profile data i.e the frequency as obtained in response spectrum is considered as time in milli-second by Caesar II for time history analysis .
I think this is wrong it will represent the wrong results because frequency by mistaken is considered as time .

Awaiting valuable response ,

Dear All ,

Please suggest which approach is valid for time history analysis of water hammer .

Awaiting valuable response

You can do this either way. If you want to perform a time history analysis, then you follow option 1 above. If you want to simulate the fluid hammer in a Force Spectrum Analysis, you follow option 2.

Both methods are acceptable, however:

- Time History requires you to be more precise with the input data.
- Time History allows you to walk the pulse down the system.
- Time History takes longer to run

- The Force Spectrum is performed in the frequency domain, therefore there is no sense of time or phasing.
- You can only look at the pulse in one location per load case, so if you want to walk the pulse down the system, each leg must be a different load case.
- The results from the Force Spectrum run should be conservative when compared to the Time History run.

Thanks A lot Richard for your valauble response !

Hi Trips,
just model a simple loop anchored at both ends. then apply the some arbitary dynamic force at loop in time history profile using edit/enter spectrum data .
After this enter dynamic force using DLF generator.

Observe both the results.
you will see the difference in results.

i feel DLF spectrum give the correct result.

I also working on same problem and have not come to any conculsion yet.

Hello Learner2011 ,

The response from Richard is correct . If we are using Time History analysis for Impact load , we have to enter the Time-Force waveform through only "edit/enter spectrum data" .

And if you are inputing the wave form data through DLF/SPECTRUM Generator then the Impact load analysis will be done by RESPONSE SPECTRUM METHOD . Which is already used in the example HAMMER.C2.

The TIME-HISTORY analysis is more accurate than the Response spectrum method but it requires a refine & well defined TIME-HIstory profile to get the accurate result .

Hi,

If we have 2 unbalance forces (in this case the unbalance forces are surge forces) with same magnitude but opposite directions in 2 adjacent elbows, we need to consider F1 and F2 for each elbow respectively to avoid the forces cancels out each other in static analysis.

In dynamic time history analysis, do we need to set the "force set" to #1 and #2 respectively for the 2 loads?

I am referring to Caesar II 2011 User Guide, page 7-21. But in this reference it is explained for PSV reaction force.

Thanks.

If these two forces are caused by the same surge passing through your system it will affect (to some extent) each elbow-elbow pair throughout your system and not just this single run.
You could enter two forces - the first starting the imbalance and the second ending it but I think a single load based on the pressure change between the two eblows is simpler. Just model the change in pressure rather than actual pressure in the run.
I believe you will find a useful discussion of this in the time history example in our old newsletters (available through "support" at COADE.COM)

Thanks Dave.

I will look into the example and discussions.

Cheers!