Structural glass design without FEM 1

[fp23]

I'm being asked if I can design a glass railing for a client. I'd like to learn glass design and I don't mind eating it on my fee to break into it, but this project doesn't warrant buying an expensive software like RFEM. My project would be a glass panel system with 2 circular glass clamps at each stair tread tread and a steel plate top rail. I think the steel plate top rail would provide enough stiffness to act as a support but i'd like to check it as a cantilevered system and keep the top rail as redundancy if I can. I've attached a photo of what the base connection would look like but my railing would have a continuous steel top plate with corners. Based on the research I've done, it seems like I would want to build an FEM model of a glass panel to be able to look at stress concentrations at the glass clamps.

My questions are:

1) For those of you who have done glass design like this, do you agree that an FEM model would be required to accurately check for stress concentrations at the glass clamps? If the glass was simply spanning between the top rail and the tread do you think an FEM model would still be needed to check for stress concentrations at the clamps?

2) I typically use Risa 3d but it's my understanding that Risa can't analyze brittle materials so even if I created a material with glass properties I shouldn't trust an FEM analysis for glass from Risa. Does this seem correct?

3) Any other affordable softwares or design guides out there for something like this? The most affordable option I've found so far is a 1-month license for RFEM at $600.

 

[KootK]

Lots of useful stuff out there but I would look at the guide below by NCSEA as a starting point. It's a decent road map for the discerning glass tourist.

 

[fp23]

Thanks Koot! I appreciate the starting point. It's definitely a little overwhelming to try and figure out where to start with glass.

 

[SirScole]

I have a client for whom we analyze ornamental handrails like this on a regular basis. We typically make a FEM model in RISA (I haven't personally heard anything about the brittleness issue that you raised before, but would be interested to learn more about it) and then compare the output stress to an allowable stress using an omega of 4. The stress concentration around the clamps is the biggest issue here so I would always recommend an FEM when using these kind of button bracket clamp systems. Then we use the reactions from RISA to analyze the button bracket clamps.

 

[Euler07]

I've done a lot of these and can provide the following comments:
- For residential, the answer will most likely be 12mm thick heat soaked toughened glass. For any higher loading use 14.28mm SGP. Your local code may now require laminated (SGP) for certain heights above ground level (for example, over 5m).
- FEA analysis around bolt holes using FEA is highly dependent on the assumptions and modelling conditions. For example, using a fixed surface for the whole clamp area will probably result in a pass at the glass edge. In which case the dominant stress area will be at the edge of the clamp.
- If you're thinking of RFEM because of the "glass" module, then I can tell you that you will not be using this. Instead, you will just use the default FEA plate analysis. In which case, any software that has plate analysis is sufficient. The benefit of RFEM would only be in the large deflection 2nd order analysis option, which might slightly reduce the stress in the glass.
- Glass balustrades (excluding along diagonal stairs) are more often than not based on testing data. The fixing supplier should be able to give you testing reports. You can then model the glass in the testing report, then compare this against the diagonal glass, to say whether there is greater or smaller stress for the diagonal.
- If I remember correctly based on my own modelling, stress in the glass at stairs is less than along straight lengths. This is because the moment lever-arm is larger due to vertical distance between treads.
- The top rail will give additional support for point loads in the top corner of the panel (load sharing to the next panel), but will not count as a support in the top. It needs to be designed as cantilevered.
- Looking at your 3d model, the top landing needs vertically paired stand-offs instead of a single line of stand-offs. This is the main issue. The fixings at the stair appear sufficient, based on my experience.

 

[shacked]

Interesting topic. I am in the same boat, since I am a sole proprietor and the cost of FEM software is just too much to justify the purchase considering on how many times a year I would use it.

 

[MRob909]

I've designed an almost identical staircase in the past, and I used all hand calculations to get the job done.

For the glass design, my reference was this textbook: https://www.istructe.org/resources/guidance/structural-use-glass-buildings/

Which covers hand calculations for balustrades etc. A bit simplified, but with some research on 'typical' solutions, you won't be far from the recommendations that @Euler07 is making.

I would add, for the design of the stringer don't try to skimp on the size - deflection is not your friend when glass panels are involved. You will also want to carefully consider the detailing around the connection of the glass to your treads to isolate or limit deflection etc.

 

[RenHen]

1) For those of you who have done glass design like this, do you agree that an FEM model would be required to accurately check for stress concentrations at the glass clamps? If the glass was simply spanning between the top rail and the tread do you think an FEM model would still be needed to check for stress concentrations at the clamps?

I would probably always use FEM for point supported glass. For a two side simply supported condition, you could do that by hand with the ASTM E1300 procedure (assuming US), but the point supports at the tread won't behave as a continuous support and the top cap will not behave as a support at all unless it is attached to something at the ends and sufficiently stiff to transfer load from the edge of glass back to those supports, which isn't something I see much. You will get some benefit of load transfer between adjacent pieces of glass through the top cap for a concentrated load case on one piece of glass, but then you'll have to consider concentrated load on the end pieces of glass where they don't have an adjacent piece on both sides to transfer the load to.

2) I typically use Risa 3d but it's my understanding that Risa can't analyze brittle materials so even if I created a material with glass properties I shouldn't trust an FEM analysis for glass from Risa. Does this seem correct?

That is not my understanding though I'd also love to learn more if there is a concern there. My main concern with RISA would be the lack of a way (that I'm aware of) to account for the variability in interlayer properties and the geometry of the fitting through the thickness of the glass, but for an interior handrail where temperature and load duration are not going to be anything interesting and with clear holes in the glass, I can't think of any reasons I'd be too concerned with RISA. Appropriately defining the boundary conditions will be trickier than with the more specialized options, but it's doable. JEI Structural Engineers has a few videos on YouTube comparing SJ MEPLA and RISA for glass handrail design, probably worth a gander! I would also definitely review the ASTM E1300 on effective thickness (for laminates) & load share factors (for asymetric laminates) if those are going to be relevant.

3) Any other affordable softwares or design guides out there for something like this? The most affordable option I've found so far is a 1-month license for RFEM at $600.

Not that I'm aware of but it's worth it for my company to just pay for all the specialty stuff for this so I haven't dug around. The standard ones I see for glass are AWGD (doesn't do point supported glass, basically just ASTM E1300 as a program), SJ MEPLA (I believe this one is more expensive than RFEM), and RFEM. The ASTM E1300 (or local equivalent if you're not in the US) isn't a design guide for this but will have information that is useful for modeling/analyzing this correctly.

 

[KootK]

That is not my understanding though I'd also love to learn more if there is a concern there.

My understanding matches yours:

- No brittleness problem with RISA.

- What RISA lacks is the ability to model laminar plate things.

- Laminar modelling only matters when it matters. Mostly IGU and any laminates where differential shear stiffness of the layers is so significant that it cannot be reasonably captured using an effective EI etc.

 

[fp23]

Thanks for all the great responses. This is all super helpful. I've got a lot of reading/research to do now. I'm also waiting for a call back from Risa and will let you know if I find out anything informative.