How to model the effect of wind on a sheet metal panel? 1

[tlewis3348]

So most of what I've done in the past is analysis on static systems. However, there is currently an urgent need for some non-static analysis to be done where I work, and I'm the only one available right now that knows how to use ANSYS. What happened was that we had one of our systems that was supposed to be able to withstand high winds (i.e. hurricane force) being trucked down to Florida. On the way, one of the panels got ripped off by the wind from driving at ~65mph. This obviously caused everyone to go into a panic (or at least be significantly concerned) that we are selling things that fall far short of their expectations.

My question is, as someone fairly inexperienced in this field, I'm not exactly sure what kind of analysis is going to be required. It seems like it would be some form of a multi-physics simulation involving CFD, vibrations, and solid mechanics. In the end, we aren't sure if the panel was just ripped off by the wind from truck moving down the highway, or if effects of the vibration of the trailer and additional wind from the environment also played a role.

To say the least, this is very new territory for me, and although I'm well aware that the information I've provided above is fairly limited, I would very much appreciate any advice anyone might be able to provide.

 

[rickfischer51]

The first thing I would do is a forensic analysis of the failed assembly. First, you will go nuts trying to analyze the whole structure for failure from many possible failure modes. Second, it would be unfortunate if you revised the structure and increased its cost because a part wasn't heat treated or some bolts were not tightened. I say this as a finite element practitioner and as someone with some experience with failure analysis, fractography, etc. Maybe a load securement device came loose, maybe something got made out of the wrong material. Maybe somebody doesnt know how to weld. It would be difficult to determine something like that by performing a full banzai Ansys flog. Just my $2.0e-2.

Rick Fischer
Principal Engineer
Argonne National Laboratory

 

[L_K]

Agreed with the above poster. In my opinion, multi-physics (cfd, fluid-structure interaction) analysis should be the last resort.
If you have an idea where and why the structure has failed (based on forensic analysis), you could use static analysis (with some kind of estimation of wind load
included) to improve it.

 

[tlewis3348]

Yeah, after sleeping on it last night, that's basically the conclusion I came to. Having said that, it might be easier said that done to do that. I just got finished talking to a coworker who is more familiar with what happened about what's going on and he said that there was apparently a meeting yesterday that the Vice President of the North American operations attended, and two of the main takeaways from that meeting were that making assumptions about the cause of failure is bad, and simulations eliminate assumptions. Now obviously, that's not entirely possible (no model is possible without making some kind of assumption), but it is clear that they are going to want to see some sort of pretty picture. The good thing is that based on the pictures my coworker had and the explanation he gave, I think I have a pretty good idea what caused the failure. Let's just hope I can find a way to simplify the simulation sufficiently to make it possible and still be enough to capture what's going on.

 

[rickfischer51]

The pics should be of some help. See what the building and structural codes use for estimating wind load. And don't forget to google. Here's what I found in 10 minutes online:

https://ocw.mit.edu/courses/civil-a...gn-fall-2003/recitations/rec1wind_eqloads.pdf

http://nvlpubs.nist.gov/nistpubs/ScientificPapers/nbsscientificpaper523vol20p697_A2b.pdf

http://www.iitk.ac.in/nicee/IITK-GSDMA/W06.pdf

http://www.oas.org/pgdm/document/mhbdc/b1_text.pdf

Rick Fischer
Principal Engineer
Argonne National Laboratory