Torsion in Beams and Columns for Concrete Design in RISA-3D?? 1

[silvabkenny]

I have a doubt about concrete beams and columns design in RISA-3D. When I show the Detail Report of some of those elements it prompts a message like this: "Beam (or Column) Design does not consider any 'T' & 'My' Moments, nor 'A' & 'Vz' Forces". I made a same model of a structure in RISA-3D and SAP2000. When I designed the structure in SAP2000 the 25 cm x 25 cm concrete columns failed and only passed with 30 cm x 30 cm. The support beams of the flat slab failed with 25 cm x 30 cm and then I changed to 30 cm x 40 cm to avoid failure. In the other side, RISA-3D surprisingly passed the bending check and shear check with 25 cm x 25 cm columns and 25 cm x 30 cm beams. I used the same Code for the design (ACI 2005) and the exactly same load combinations in both programs. I suppose the appreciable difference between the two programs is RISA-3D doesn't consider torsional effects on the structure elements. My fundamental question is ¿What level of trust can I have with the design of concrete structures in RISA-3D?

 

[JoshPlumSE]

You did not mention how much steel was assumed in the SAP model vs. the RISA model. It seems to me that this is the most important part of determining if the member passes or fails a concrete code check.

In RISA-3D, that warning text regarding A, T, Vz, and My is just warning you about something fairly straight forward. When you designate a member as being a "beam" the concrete code checks and rebar design look at strong axis bending and strong axis shear. The program does not consider axial+moment interaction diagrams. Nor does it consider the effects of biaxial moment interaction.

For columns, RISA-3D will consider everything except torsion. However, it is also worth noting that bi-axial column interaction is pretty complex and two different methods (exact integration vs. PCA Load Contour) may yield different results. If you are comparing the program to hand-calcs, then you should probbly use the PCA load contour method, otherwise, the more accurate solution is usually the exact integration option.

 

[silvabkenny]

I didn't assume steel reinforcing for the concrete members in SAP2000 or RISA-3D. I know I can do that in RISA-3D creating a Rebar Layout and then assigning to the design property of the concrete members, and in SAP2000 in the concrete reinforcing information of the Definition of the Concrete Section Members. I didn't do that. I only create the same model in both programs with the same material property, the same sections, the same applied loads, the same load combinations and the same design codes (LRFD Second Edition and ACI 2005). Then, I analyzed and designed the model to obtain how much steel reinforcing it was neccesary for every concrete members, and of course, to know if the section passed the respective code checks. I like very much RISA-3D program, but I'm worry about the way RISA-3D make the code checks of the structure concrete members. I don't know if SAP2000 consider torsional effects for the design, but I suppose an affirmative answer because in the design report of the failed concrete members it prompts a message like this: "Torsional stress combined with shear forces passed the allowable value". I attached the RISA-3D file that contains the model I made. Please, check it and tell me if I'm doing something wrong that makes that RISA-3D pass the code checks with that sections while in SAP2000 fails. Thanks.

 

[JoshPlumSE]

You post is pre-disposed to thinking that RISA and SAP need to give the same code checks for the model. But, that may just be an incorrect assumption. Possible sources of differences:

1) Stiffness / Cracking: Are RISA-3D and SAP assuming the same stiffness for the concrete members. RISA assumes a certain amount of stiffness reduction due to cracking for beams and columns. If the stiffness of the members is different then they will get different forces in one model from the other. If the forces are different then how can the code checks be the same?

2) Reinforcement: How much reinforcement is placed in the model in SAP? How much reinforcement is placed in the model in RISA-3D? If they are not the same, then the two program will arrive at different results. When you ask RISA to place the bars directly, then you run the risk of the program reporting a dramatically different number of bars from your existing SAP model. Additional, the location of these bars could be different as welll. RISA always trys to put an equal number of bars per face for columns. Does SAP try to place a different number of bars in each face based on which local axis receives more moment? If the bars end up being different then how could the code checks remain the same?

3) Moment interaction equations: The interactioni diagram for biaxial bending + axial force for concrete columns is relatively complex. There are all kinds of different ways to treat this. In RISA, you have chosen to use a parabolic stress block and the exact integration method. That should work fine it's probably the most accurate choice.

But, what has SAP chosen to use? Are they doing a full interaction surface, or are they using an approximate PCA Load Contour type method? I strongly suspect that they're using the standard / rectangular stress block. If so, you could switch the Global Parametes setting in RISA to match this result.

4) The torque issue is a possible source of the difference. However, when you look at your columns it appears that columns have very small torques relative to the strong and weak axis moments. Therefore, I'm not sure that this can be considered the major source of your differences.

 

[silvabkenny]

Hello, my friend JoshPlum. I didn't write early because I was reading the RISA-3D and SAP2000 user's guides. After that I'm very disappointed about RISA-3D concrete design. First, I verified without any doubt that RISA-3D not consider torsion in beam or column design. Torsion is very important in building structures and in seismic design and SAP2000 take into account torsion, flexion, and shear in the concrete design. The shear design is made by RISA-3D only by the simple formulas of the ACI Code without Seismic Considerations. SAP2000 in the design of special moment resisting concrete frames (seismic design), the shear capacity of the beam is also checked for the capacity shear resulting from the maximun probable moment capacities at the ends along with the factored gravity loads. This check is performed in addition to the design check required for Ordinary moment resisting frames (wich is made in RISA-3D). Second, RISA-3D not take into account the joint design (this is very important in a structure design). SAP2000 program performs a rational analysis of the beam-column panel zone to determine the shear forces that are generated in the joint. The program then checks this against design shear strength and also determine the effective area of joint that resists the shear forces (RISA-3D don't make this). I don't wanna say that RISA-3D is a bad program, it makes very well its work in its scope, but not take into account some design criteria very important in the design of concrete structures. If someone decide to use RISA-3D like a main structural analysis and design program must take care of check the torsion, capacity shear and joint design requeriments of the ACI Code (or another Code). This can be made with a well designed spreadsheet or another kinds of auxiliary programs. Thanks for your help my friend.

 

[JoshPlumSE]

silvabkenny -

Well, that certainly explains it. The reinforcement that RISA gixe is based on the main parts of ACI code and the reinforcement that SAP chooses is based on the seismic detailing provisions.

That being said, the tone of your message is overly harsh.... The documentation in RISA should be very clear about what the program does and what the program does not do. What in the RISA input or output suggested that seismic detailing provisions were accounted for?

I wouldn't mind hearing more from you about torsion and what you think should be done for concrete frames. Most of the time, I would think, you will just want to check "threshold torsion" to see if close loops are required in your beams or whether you can get away with plain stirrups. Am I correct? Or, were you looking for something more detailed regarding torsion?