Non-Stop Solution Calculation - RISA is in a non-stop cyclical loop of solving my model? What's up?

You have P-Delta instability most likely. Turn off the P-Delta analysis, run the load combination you're stuck on, and look for issues. You can turn on the deflected shape and animate; that will often lead you to the problem. That's one possible cause.

Thanks UcfSE. I turned P-Delta off for that specific LC and I guess it Solved. But I did get the Warning message 1078 that says the analysis was run without P-Delta.
I looked at the deflected shape and sure enough I've got a Node taking off into the ground. It's a node at the end of a monorail beam. this rail extends about 6 feet outside the building, I've got 2 horizontal braces supporting it and a third member attached to the top of it, kinda like a strut, or knee brace at the top. there's 7.5 kips at the end of the monorail (test load) and it's flying off the paper. I could put a small spring there because I know in real life it will hold up but it's not in the model.

Any suggestions about putting a Spring at that Node/Joint?

Btw, my boss wants all models solved using P-Delta so I've got to find a way to get it to Solve using P-Delta.

Thanks.

I would absolutely keep the P-Delta on. We're just troubleshooting right now, so you need the model to run so you have something to work with. RISA will give you a warning about turning off P-D, so no surprise there. How you approach from here will depend a lot on what you're modeling and its purpose (so that your approach to "fixing" your model does not lead you away from where you want to end up). New design or existing?

I'd check your boudary conditions for accuracy (member fixity/releases for example), or maybe your beam simply is not stiff enough. A spring may be an option if you can define or at least bound the spring stiffness you have.

Thanks, yeah I've checked all my releases and boundary conditions. I'm applying springs to different nodes as required and it's like chasing a greased pig.

I also tried changing member sizes but it didn't work. This is like a Cantilever condition with 2 horizontal and one vertical brace attached to the free end.

You could post the model and let us see if we can spot the problem. Even a screenshot with all the member end releases highlighted would help.

Like UcfSE mentioned, make sure your beam is turned the strong way. Could be something that simple.

How do I post/attach the model file?

Thanks,

Below the "Reply To This Thread" is a place for attachments. Click on "...or upload your file to ENGINEERING.com", browse for the location of your file, and "upload".

Node number 76 is the problematic node. Currently the only way I can get the model to solve for All lC is to put a Pin connection at that node.

Lunch Time!

You have an instability. Your knee braces and hangers are all free to rotate around the axis of the beam (M38).

If you fix the ends of the two vertical members you'll see that it will solve all load cases. Basically it is a bad arrangement. You probably shouldn't tie the knee brace back into the middle of the beam either, in my opinion.

You have too many members that are basically free to swing. I see now why you wanted so many springs. You can keep the springs, of course, but I would take a look at what it is you're modeling and see if you can get the structure to support itself. Springs still mean you have to consider force transfer and deformation compatiblity wherever you are attaching to something, and need to be representative of the stiffness of the connection. Some things to consider: give your node labels a prefix, such as "N" or "J". Makes it easier than having more numbers to track later on. I would also clean up your member spreadsheet. That makes a good habit. Consider adding a prefix to existing and new members so you can identify, sort and exclude/hide them easily. What are the tiny area loads for?

Sometimes you can get an instability error when you have too many releases at a particular joint. In that case, pick one member end to be fixed. If all the other members connecting at that joint are free, the joint will still be hinged.

All three members connecting to node 76 are listed as "Euler Buckling" members. Just looking at this model, I would suggest that you set them as regular members which are capable of taking both tension and compression. Remember, "Euler Buckling" members are normally intended to be slender rods, cables or singles which are not expected to take ANY compression. We set them to Euler (as opposed to a regular tension only member) only to help with convergence. If this is a member that would cause an instability if it were a pure tension only member, then it should NOT be specified as an Euler member.

I really appreciate ALL the input and help guys. Some of it I knew and some suggestions were very helpful. I am a Staad veteran and have only been running RISA for about 3 months now, I like it a lot and a lot of the stuff is the same, just different syntax and nomenclature. As far as this structure being a "bad arrangement", I have found thru the years that sometimes, not all the time, but sometimes there are things like cyclone ducts, fire water piping, interior light hangers, steam piping, cable trays, other existing structural members, the inability to weld because of a highly explosive area (which limits the types of connection you can have), the inability to field drill holes, I can go on but basically those are some of the limitations I have with this structure. And, what you don't see in this model also, is a dual 3 Ton monorail that is hanging from this new support steel (it's a TC/American patented track monorail) that splits and becomes one monorail. Designing this "non-traditional" structure is what keeps me coming to work every day.

Again, thank you all for your input. The model is solving now. I may add a couple more members if I have the space. I agree, it looks very "funky".

Peace!

Oh yeah, UcfSE, those tiny loads are 10 psf for collateral/utility loads and 20 psf for roof loads.

Ok. Check your units then. You have them as -0.01 psf, not -0.01 ksf.

YIKES! LOL. Good catch. Got it. Thanks.