I am working on a piping system where I also have to calculate the impacts of the slug force. In CII ver5.3 the spectrum generator is a bit different. could anyone tell me why and how the generated spectrum is always based on Frequency/Force? if I change the range types in the first tab of time history analysis into Time/Force, do I need to edit the spectrum again? something is seriously wrong in my calculation since a 10in elbow goes bust even with 1000N of slug force. (piping has sufficient supports in every direction.)

First you have to decide what type of analysis you're performing: (a) time history analysis is performed in the time domain and your pulse must be defined in terms of time/force, (b) response spectrum analysis is performed in the frequency domain and your pulse must be converted to a DLF, frequency/magnitude. You cannot mix the two types of loads.

If your evaluating your slug situation using the response spectrum method, you do want to use the "DLF/Spectrum Generator" and you DO NOT want to leave the spectrum definition as Frequency/Force.

Richard,

thanks for the reply,
well, I forgot to mention that I chose time history analysis. my question still remains "how come the entry of time/force in the spectrum generator/user defined time history, turns into frequency/force?"

regards

That's what the spectrum generator does - it creates force spectra and one of functions is to convert a user-defined time history into a force response spectrum (DLF vs. frequency). This force response spectrum can then be used in a force spectrum analysis.

But as Rich points out, do not use that spectrum data as input for a time history analysis.

By the way, I find that force response spectrum data to be very useful in determinung the frequency content of the time history, if only for resetting the frequency cutoff for the analysis.

Dave,

do you mean it would be best if we set the cut off frequency at the point when the oscilation almost fades away at DLF=1 ?

regards

To be conservative, at least use all modes below the peak response and include missing mass. As you go beyond the peak response, the DLF used for the missing mass will drop. So including modes out to your DLF=1 value would be more accurate, it may also take much longer to calculate.

Dave,

one last question; all of my frequencies are calculated under "out of core decomposition" , have you got any suggestion to get rid of all singular matrixes? I set my tolerance on 1E13 (the highest) and I got no change.

thanks

If you have a singular matrix you have rigid body motion (some part or parts of your model are not restrained). Changing the tolerance won't help this - you have to correct your model.

Richard,

my piping system is relatively short and there are some flanged conections in it. I can not really improve my system in that respect. do you mean having frequencies under the "out of core decomposition" calculated would potentially give me wrong answers?

regards

Whether the solution runs "in-core" or "out-of-core" is irrelevant. You shouldn't have rigid body motion. Something in your model is disconnected or otherwise free to move in at least one direction.

I see, I have got a free end vent line inside a sulg catcher vessel....

A free end is fine (i.e. a cantilever). However, if you removed the anchor from the wall and substituted a pin, you would have rigid body motion, and therefore a singular matrix.