How to model sloping diaphragms 3

[jbedre]

Wouldn't it be fantastic if diaphragms did not have to be flat (XYZ) planes? Anyone have any creative ways to model sloped diaphragms (roofs) that do not adversely affect the rest of a model?

 

[lav322]

I have the same question. In a perfect world all diaphragms would be flat and perpendicular to the wind loads. A perfect world wouldn't need structural engineers either.

When I asked RISA about implementing sloping diaphragms they say it will be released in the next release. Two releases, maybe three, and still no sloping diaphragms. I guess I won't hold my breath.

 

[rwpe]

Yes...that would be cool.

However, as a point of clarification from my own experience, you need to be careful of what kind of diaphragm you are trying to model. It is my understanding (from a project a few years ago) that the "diaphragm" feature of RISA is limited to rigid diaphragms only. It is my experience, at least in the southwest, that the majority of your roof systems (residential and commercial)will be flexible. In which case, I do not believe the RISA diaphragm feature could be used here.

 

[Indianola]

I just put a message in the 'wood/residential' thread that pertains to this. (See it for more info.)

While I was addressing wood in particular, any flexible diaphragm with fasterner slip (plywood, or steel deck), has the same requirements for getting an accurate model.

As soon as you get fastener slip in your diaphragm, your "G" for the plates becomes lower than Risa will allow for the "E" of the system. The work-around that still gives accurate results is to use "G" with made-up "E" for the shear panels. You have to be sure to model all the boundary elements and drag struts, and use the correct "E" for their material (not the "E" from the diaphragm). If you miss a boundary element or strut, your answer is no longer reasonable.

If you try to model for the "E" of a flexible diaphragm, your results will not be realistic. You can only do that with a stiff diaphragm without slip, such as a concrete deck.

Scott Beard, PE, SE

 

[McSEpllc]

Regarding modeling sloped diaphragms: I have added weightless "diagonals" of certain stiffness between framing members to model the impact of diaphragms in the past.

ASCE 7 says untopped steel deck and wood sheathing can be idealized as a flexible diaphragm, while concrete slabs and concrete on metal deck with a span to depth ratio of 3 or less without horizontal irregularities can be idealized as rigid. Other conditions can be evaluated and considered flexible were the computed in-plan diaphragm deflection is more than twice the average story drift of the adjoining vertical elements.

RISA says:
"You may alter the stiffness of the diaphragm, though this value should almost never be changed. Arbitrarily changing the diaphragm stiffness without understanding the ramifications on the stiffness solution can produce solution results that are inaccurate. Having said all that, the stiffness of the diaphragm may be adjusted from the Diaphragm spreadsheet by clicking the button on the Window Toolbar. The default value is 1.E+7. The exponent of the internal stiffness of the diaphragm is twice this displayed value (1.E+14 in the default case) This value should only be adjusted for 2 reasons.

The first reason is that the lateral force resisting elements in your model are so stiff that they are causing the rigid diaphragm to behave semi-rigidly (i.e. the rotations are not all the same for all joints on the diaphragm.). In this case you could try increasing the diaphragm stiffness to 1.E+8, however the internal diaphragm stiffness of 1.E+16 starts to encroach on the boundary condition reaction stiffness which is internally modeled as 1.E+20. A diaphragm stiffness value greater than 1.E+8 can produce unpredictable results that are characterized by ghost reactions, meaning some of the joints in the diaphragm begin to behave as Reaction boundary conditions.

The second reason that you might adjust the diaphragm stiffness is because your dynamics solution will not converge. In this case, you will want to reduce the diaphragm stiffness to 1.E+6 or even lower if necessary. As you lower the stiffness, you will need to watch the joint rotations for joints on the diaphragm to insure that you getting, or at least approximating, rigid diaphragm action. The joint rotations should be the same for all joints in a rigid diaphragm."

The world is not black and white as our math and the codes are. I usually play around with stiffnesses of the diaphragm to see what the impact is on the various structural members.







Eric McDonald, PE
McDonald Structural Engineering, PLLC

 

[JoshPlumSE]

Lav322 wrote:
"When I asked RISA about implementing sloping diaphragms they say it will be released in the next release. Two releases, maybe three, and still no sloping diaphragms. I guess I won't hold my breath."

Guilty! But, if it is any consolation RISAFloor has added sloped floors for the October (or early November) release.... that doesn't yet include sloped diaphragms yet, but it does bring us one step closer.

There is actually a way to do sloped rigid diaphragms, but it is a decent amount of work. Therefore, we never added it into the Modeling tips section of the program.... If you use a series of weightly rigid links (with pinned end releases) to form a sloped truss then this will get you the same behavior as a sloped diaphragm.

 

[JoshPlumSE]

Just an FYI:

RISA entered into a development agreement with the WPC. Part of this agreement was for RISA to incorporate flexible wood diaphragms into the programs. Technically, this is supposed to happen by the end of the year.... While we will be hard pressed to meet that date, I don't think we would miss it by much.

What we are planning on doing is analyze the wood diaphragms as ideally flexible or ideally rigid. Then run the APA deflection calculations for the type of diaphragms specified (with nailing patterns and such considered) and compare the APA deflections to the code assumptions about rigid and flexible diaphragms.... Letting the user know if their assumed behavior matches the analysis assumption for the diaphragm.

The first release of this will, of course, only do this for wood diaphragms.... But, after that is complete then we will be open to user suggestions about adding this for other types of diaphragms (metal deck or concrete).