Arch building stiffness - calibrate model with reality
Arch building stiffness - calibrate model with reality
(OP)
Well I've got a real interesting problem at work. Some testing was done on arches in an existing wood storage building (see attached sketch) to determine deflections under load, and I've been asked to make a FEM model that can match these and thereafter be used for reliable analysis (knocking out some conservative assumptions we'd make in design). I've matched deflections reasonably well for interior arches, modelling arch connections as pinned, and adding in the 5/8" plywood roof and 2x10 purlins @ 20" o/c. When I get to the arch with rod bracing on either side though, I'm getting deflections ~60% in excess of what was measured (measured deflections here are much smaller than for interior arches). I've modelled the bracing with pre tension (this does very little in the model and hand checks I did suggested that would be the case) and I've modelled the end wall as roller supports free in the long direction of the building. Curious to see what other people think of this problem. Deflections in the order of 1-3 inches all around FYI.






RE: Arch building stiffness - calibrate model with reality
Actually, the more I think about it, I guess whichever arch would be acting as the top chord of your horizontal truss would be in compression and might push that arch up giving the appearance of less deflection.
But I could just be talking out of my ass.
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
As I said I would have assumed the deflection was the same for all interior arches. I can only guess that the testing numbers revealed it deflected less than expected.
The other thing you need to keep in mind, old wood was super strong and highly variable. It is entirely possible that the one arch is made out of stiffer material than the rest or that perhaps the connections at the base are performing better.
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
Probably not as I've never heard of a modeling package with that as an option but what happens if you remove the pre-tension?
In any case it sounds to me like the rods are pulling down the arch in the model but not in reality.
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
As far pre tension goes, an increase to 50kip of pre tenstion stiffens the arch ~11%. Obviously this is not realistic and probably not where I'll be able to account for the measured deflection.
RE: Arch building stiffness - calibrate model with reality
This is an interesting problem. I've got a theory to submit. KootK confidence level = 7%.
You can think of your deflection as two separate components:
1) The component associated with arch action resistance. This should be very stiff.
2) The component associated with arch flexure. I'm guessing that this is where most of the action is.
I think (7%) that the tension rods, in combination with accidental compression struts formed by the plywood and the purlins, is creating "mini-arches" within the larger arch. And that is resulting in a stiffening because a greater portion of your load will be resisted via arch action as opposed to flexural bending.
It would be relatively easy to test this theory. Just model the accidental struts as such in your FEM model and see what happens. If you do that, I would love to see an axial load plot of the members of concern.
The greatest trick that bond stress ever pulled was convincing the world it didn't exist.
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
Kootk: Interesting theory. I've already modelled the purlins, the purlin nearest to the peak has 2.5 kip of compression in it. I've made a quick 2D model to test out these mini arches, they're actually in compression for the hinged arch I have (makes sense when you look at the deflected shape of the arch). I'll continue exploring the concept.
RE: Arch building stiffness - calibrate model with reality
Based on your recommendation on another thread I bought the book "The Stone Skeleton" and have begun reading it. Your arch-with-a-hinge-sketch looks like it could have come right out of it.
RE: Arch building stiffness - calibrate model with reality
Does the same thing happen at the other end of the building?
The rod braces are really tension AND compression members until the pretension is overcome. That means that the cross bracing can form its own arch capable of supporting a few kips.
The greatest trick that bond stress ever pulled was convincing the world it didn't exist.
RE: Arch building stiffness - calibrate model with reality
Agreed - I'm fine tuning my model to try and capture some of these effects. My 2D simulations done just for proof of concept show a much stiffer hinged arch system with mini arches, just a matter of trying to capture these effects in the more complex building model
RE: Arch building stiffness - calibrate model with reality
The greatest trick that bond stress ever pulled was convincing the world it didn't exist.
RE: Arch building stiffness - calibrate model with reality
The last (only other?) work of fiction that I know of that involved structural engineers in any way was that tv show "Prison Break" from a few years back.
RE: Arch building stiffness - calibrate model with reality
Doug Jenkins
Interactive Design Services
http://newtonexcelbach.wordpress.com/
RE: Arch building stiffness - calibrate model with reality
Update on my progress - I've got deflections only 25% greater than those measured by basically making sure the bracing more closely resembles what's on site in terms of location and placement, through better member offsets. I think this might be close enough.
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality
RE: Arch building stiffness - calibrate model with reality