The results in my Spectrum dynamic analysis (water hammer type) look strange because all my maximum responses occur in missing masss mode. I have exytended the frequency cut off way beyond the max. table frequency of my spectrums; the included mass output shows a 98% force contribution of the Spectrum and yet the missing mass modes persists in my output. can anyone explain why this happens?


second question: on a different note, does it matter at all to define the (+/-) direction of the acting force in the spectrum analysis?

First question - not sure. You might want to delete the dynamic results and rerun.

Second question - +/- will matter if you have multiple concurrent forces, whether they work in conjunction or opposition.

What was the cut-off in your table frequency?

For a fluid hammer analysis, you may need to take the table frequency and the frequency cut-off out to at least 100hz, perhaps more depending on the system.

I set up the table frequency on 100 HZ and I went up to even 300 HZ for my frequency cut off range.

@Michael Fletcher: Had I done Time History analysis, I would have agreed with you but, in the spectrum analysis, I just do not see any difference if I change the force vector direction.

The water hammer loads typically act axially on the long piping runs; unbalanced load due to time lag. The axial stiffness of a pipe is very high k= AE/L and thus the natural frequency is high. In general, the pipe natural frequency may be estimated as f = w = (1/2pi)* SQRT (g/delta) and delta = PL/AE. Thus, for a unit load f= (1/2pi)* SQRT (gAE/L).

I do not have a past situation handy (it's at the office), but it was a dynamic slug evaluation for a 24" to 36" line. Good practice is to use a line stop on all long straight runs, thus the dynamic frequency will be high and will not easily be captured by CII as can be seen for an upper bound case ignoring elbow flexure and anchor stiffness. For a 100 foot run (30m) the "Column" axial frequency is quite high at 2980 hz (24"x0.5" wall) and 3660 hz (36" x 0.5" wall).

I may run a test case to verify and come back and edit this post, but you get the idea.

This arose while using CII dynamics in 1994 (used many other tools prior) where it was obvious that applying a 50,000 lb (11 kN) dynamic impulse load and only getting out a few thousand pounds. By the way doing a dynamic analysis with the missing mass option yielded a MAXIMUM slug response load of 1.3 * SLUG load versus the standard 2.0 DLF if done by hand.

Using more applicable data or a more rigorous analysis can lead to more practical designs.