Bending stress analysis

[flapjack1]

Hi everyone,

I have a question about a bending analysis I'm working on. The part in question is a flat glass plate with a series of holes. The glass manufacturer only provides one strength value - bending strength - which is 25 MPa. I've set up a simulation in ANSYS where I solve for maximum principal stress, which comes out to be 13 MPa. My question is, is it valid to compare those values directly? I.e. can I conclude that I have a safety factor of 25/13 = ~2? Or should principal stresses only be compared to UTS, and therefore do I need to somehow convert my results into a value that can be directly compared to the vendor's bending strength?

 

[rb1957]

you have 13MPa at the edge of one of your holes ?

"bending strength" is reasonably max tensile stress, so I think you're ok.

Or you could do a bending stress analysis ... maximum moment form the model, elastic bending (My/I = 6M/t^2, M from the FEM is moment per unit width) would be conservative (or reasonable for "brittle" glass). This would be closer to a "bending strength".

another day in paradise, or is paradise one day closer ?

 

[bugbus]

I have a feeling this is more of a fracture mechanics problem rather than a simple bending strength problem.

I found this info online:

The estimates for stress concentration factor assume a smooth surface. If the hole is created by grinding or drilling, thus creating a huge number of surface defects, the strength of the sheet could be reduced to zero. The relevant field is fracture mechanics. Glass has a relatively low fracture toughness.

 

[human909]

Glass is a different beast entirely. It is extraordinarily strong. But also almost always exhibits some micro fractures that can readily spread in the wrong condions. Add in internal stresses and it all becomes very complicated quickly. Normal approaches of analysis are hardly applicable.

Have a look here if you have time for a bit of enlightenment regarding some of the complexities:

https://en.wikipedia.org/wiki/Prince_Rupert's_drop

https://www.youtube.com/watch?v=xe-f4gokRBs

 

[Tomfh]

Agree with the posters above. It’s a complex problem.

I’d be breaking a number of samples to see the actual strength, then factoring accordingly.

 

[jjl317]

In the US, when designing glass for architectural applications, ASTM E1300-xx is typically used. This design approach is not based exclusively on an allowable stress, but also includes a factor to address the Probability of Breakage (number of breaks in 1000 lites). In E1300-12, in sections X6 & X7, there is additional information on stresses to be used with independent stress analysis. Additionally, the type of glass (i.e. annealed, heat strengthened, tempered), the makeup (monolithic, laminated) and load duration all play a significant role in the analysis also, especially if there are holes included in the glass.

Depending on the application, I believe that there is a glass specific finite element program (SJ Mepla) that might be of assistance, as they have incorporated some typical glass connection details into the program. I believe that the program is based in Germany or Austria, and may not be tied to ASTM E1300.

But overall, I agree with the previous commenters, as it has a number of unique design properties, that don't necessarily align other materials.