Mat foundation using STAAD

[PAstructural08]

I have modeled a mat that is 115' x 83' in STAAD. I have modeled this mat using 4 noded plates and have placed spring supports on the plates using the subgrade modulus. I am also you using 'compression only' springs in the 'Y only' direction.

After running the model I have a few load cases that have warnings saying "Tension and compression did not converge after 10 iterations."

Does anyone have any suggestions on how to fix this warning??

On another note...just out of curiosity. I changed the 'compression only' springs to tension springs. And I don not get this warning at all...why is that?? It doesn't make sense to me that tension springs would work because that would suggest to me that the mat would lift up.

Thank you.

 

[slickdeals]

Soil is compression-only. You should not model using tension springs. When the mat has a moment on it, the program goes through several iterations eliminating springs that have tension force in them and redistributing it to the compression springs. This process repeats till there are no tension springs and the system is in equilibrium.

If you have a footing with a very large moment, then there is a chance that the results won't converge easily. Try to increase the number of iterations or make the footing larger (wider normal to the moment)

We are Virginia Tech
Go HOKIES

 

[ToadJones]

As Slick said, don't use tension springs at all.
I'm surprised that with a mat that large you are having problems with it converging.
I usually get instabilities when using the "Y Only" specification.
You can set the max number of iterations under Commands/Miscellaneous/Set ITERLIM

 

[SeanMD]

It might not be converging if you have the load combs set up wrong.

Are you using repeat load for the load cases instead of
load combinations?

If you havent already, try a "perform analysis" "change" and "spring compression xx to xx kfy" command directly before you start the repeat load cases.

 

[DaveAtkins]

I believe what is happening is that your footing is not big enough. Take the moment induced and divide by the vertical load induced. If the result is greater than one half of the footing dimension, you need to make the footing bigger.

DaveAtkins

 

[ToadJones]

Must be a tall building.
Be sure also that you are including all the dead loads and selfweight...sorry for stating the obvious, but that is a pretty large footing.

 

[PAstructural08]

I will set the max iterations to something other than 10 and see if that works.

SeanMD: STAAD is performing the analysis on the principal loads, these are the loads that are not converging. I am using load combinations and not repeat loads. Also, I did specify a "perform analysis" and "change" command after each of the load cases that do not converge...which doesn't help anything, but I wanted to do it for other reasons.

I notice a lot of people think this is a footing, but in reality it is a mat foundation. This mat has two steel towers on it, 8 storage tanks, and various other components. I am not able to make the mat any more wide than it already is.

I also do have all of the dead loads and selfweights in the model. I have also put 60% of the dead load in each of the load cases that do not converge. I have just done this so I'm not sure if this will work or not..I believe it should work. Since adding 60% of the dead load in each of those load cases, I have revised my load combinations to not account for the extra dead load.

 

[SeanMD]

IF your primary load cases dont converge you have a problem, and your fdn is not stable. Not enough DL, or not big enough?

Generally, to run load cases for a fdn with springs, you cannot use load combs. You will have different nodes going into tension (or zero force) for each primary load case, and each direction. You cant just sum them up for the load combs since you will incorrectly combine zero force springs. You have to use repeat loads so the primary load cases re-run combined.

 

[structuresguy]

OK, you probably have two problems with your model.

1. You don't want to run the analysis on primary loads. Only run it for load combinations. The reason is you probably have a wind (or seismic) only primary load, which will never be stable since you have no dead load to conteract the wind uplift/overturning.

2. You can't use normal load combinations for anything that needs to iterate. You have to set up primary load cases, and crete "combinations" using the REPEAT LOAD command. Doing this, you won't need to use the change command, which isn't going to do you any good in this case anyway.

 

[structuresguy]

Oops. Now I see I said almost exactly the same thing as SeanMD. I started typing my response, then got called away, before seeing your response. Oh well. at least we are on the same page.

 

[ToadJones]

Now that I think back, Sean and Structures are 100% correct.
You must make load combinantions.
Generally when I use STAAD for soil pressures, I determine the worst case combination by hand for overturning, then just make a separate load case for it.

 

[SeanMD]

Here is an example from a past project. My primary lateral load cases did not include dead loads. all primary load cases converge.

SUPPORTS
xx TO xx PLATE MAT DIRECT Y SUBGRADE xx PRINT COMPRESSION

LOAD 1 LOADTYPE Dead TITLE DEAD LOAD
MEMBER LOAD
(enter all primary load cases)

PERFORM ANALYSIS
CHANGE

LOAD 7 D
REPEAT LOAD
1 1.0 2 1.0
(enter all repeat load cases, dont use load combs)

If you are still having problems, I wonder if you have a connectivity issue with the conc plates you are using for the mat?

 

[SeanMD]

I do all the load combs in staad, thats the beauty of it, its so easy to just run them all. You can print out pictures with the base pressures to include in your calcs. Once you get used to it, I think staad is pretty good for mat fdn design, and I prefer it over PCA mats or similar.

I do an overall hand calc for OT as a sanity check also.

 

[ToadJones]

Sean-
Do you use your plate stresses for concrete design?
To this point, I have not felt comfortable doing that. I usually only use STAAD to get soil pressures in complicated footings or mats.
If you use the plate stress for concrete design, do you care to explain briefly how you approach that?

 

[PAstructural08]

Thanks for all the posts. I will try and use the repeat loads and not the load combinations. Might take me a day or so to make the changes because my model is so large.

SeanMD: When using the repeat load in your example...load 7 D, would you use the coef. that you would use in the load combinations?? Say you are using LRFD so for the dead load it would be 1.4, 1.2 or 0.9.

 

[iv63]

ToadJones: To design concrete I use STAAD’s DESIGN SLAB/ELEMENT feature. It is very handy and I did not have any problems. Element design is performed for the moments MX and MY at the center of the element. Regards,
IV

 

[rowingengineer]

From what I have read, PSU08 you have a rather large model for a building foundation. From what I can see you have entered this into your software without doing any hand calc’s, creating a large model, that would be very difficult for you to understand your results, errors or similar.

Do you see anything wrong with this? How about some hand calc’s, rules of thumb, past projects. The idea’s above are giving you good advice on how to possibly handle your current problem, whatever it may be! You should however go see your mentor before wasting a few more hours trying to get your large model into working order.


An expert is a man who has made all the mistakes which can be made in a very narrow field

 

[SeanMD]

ToadJones,

You can get concrete bending and shear stresses directly from staad. Report - plate results - plate stresses. On simple mats or footings you can do a quick hand check pretty easily on some stuff to see if it makes sense. On large combined mats it's really hard to determine the stresses, so staad is really nice for that. I dont do the reinforcing design in staad, I just take the plate stresses and determine my reinforcing with separate hand calcs. Probably old school, but thats the way I do it.

PSU08,

Yes, you enter factors just like a load combination. The only difference between "load comb" and "repeat load" is that they model re-runs all the included primary loads for each load case. Always used when doing tension/compression only springs or members.

 

[SeanMD]

Also, check all the nodes for reactions, if any show any negative values (tension) you have something wrong.

 

[structuresguy]

sendMD: thanks for all your help here. I am agreeing with everything you have said so far. This should help the OP solve his problems.

I too use Staad for bearing pressure, deflections, and overall stress calculations for mats. I then, like you, calc out my reinforcing by hand. One reason I do this, is you end up with very high concentrated moments under your loads, particularly if you have heavy columns modeled at only one node. So I average out these stress concentrations a bit, to try to account for the size of the column relative to the size of the mesh elements.

I also have Mat3D, which I have used for some things, mostly pile supported mats. Compared to staad, it gives a more complete design, calculating uniform reinforcing over the entire mat and also checking punching shear. But what I don't like, at least with my version which is a couple years old now, is that it does not give me a bearing pressure or stress plot, like I can see with staad. When you have an over-beearing failure, it is not always obvious where is the best place to enlarge your mat.

We recently also just got Staad Foundation. I like that I can export my Staad frame analysis results to it to solve my foundation design. It is pretty good, but still only on first major release version, and needs some things improved. Namely, if your mat is too small, you can't simply add more width on one or more sides. You have to delete the mat and start it over again. All you loads stay behind, but still it is a pain. Just let me edit the nodes of the corners of the mat. That is all I need. Otherwise, I really like the output calculation sheets, which show all the equations being used. It definitely spits out some of the best calcs sheets I have ever seen from a program. Great for submitting to a building department.