vibrations in structures

[wrxsti]

Examine setup below

0------------------------0
|....|.......|.......|........|
|....|....*.|.....*|........|
|....|.......|.......|........|
|....|....*.|.....*|........|
0------------------------0


* = Point load from vibrating equipment
0 = column
. = concrete slab
- = primary structural steel
| = secondary structural steel

I am trying to find natural frequency

proposing composite for secondary beams with slab
effective width span for interior beams

would not consider exterior beams as part of natural frequency

distributing point load as per Pa/L and Pb/L
in addition to mass
to get delta_beams

then delta of primaries

then do 0.18sqrt(g/(delta_Secondary1 + delta_secondary2 + delta secondary_3 + delta_primary_1 + delta_primary_2))

??

 

[WARose]

This thing basically screams for a FEA solution. For equipment constantly vibrating, localized effects will be important. Hand solutions (including capturing higher modes) just won't cut it in my opinion.

Also, in the time you'd spend messin' around with a hand solution....you could probably do it several times over in something like STAAD or RISA.

 

[jsmith234]

WARose what extent of modelling would you do in staad would you stop at the 4 perimeter columns
or model entire structure levels above and below?

 

[WARose]

WARose what extent of modelling would you do in staad would you stop at the 4 perimeter columns
or model entire structure levels above and below?

Depends on what is beyond (and how sensitive it is to vibration)....also on the amount of unbalanced force we are dealing with here as well as several other things.

But in general, if we are talking light unbalanced forces in a facility where low vibrations are no big deal.....yeah, I'd probably stop the model at the columns.

 

[jsmith234]

i am talking in terms just generating natural frequency through the software
to check for any magnification potential

 

[wrxsti]

Yes ofcourse WARose FEA would be ideal

Im just checking the approach or if

1/fn^2 = 1/f_slab^2 + 1/f_primary1^2 ..... + 1/f_secondary^2

would be more appropriate
as a simplified method to establish natural frequency