Mesh Size Does Matter: FEA Errors from Mesh Sizes

Are you sure that every FEA analysis ever ordered is accurate? Mesh sizes are the biggest source of error in an FEA simulation. Today I demonstrate that mistakes in FEA mesh sizing may be larger than you anticipated.
Get full details in the article: wp.me/p3pMQV-4rO
View more tips and helpful articles at www.dmsonline.us/

[@omgitzsteg]

You're a funny dude, thanks for the video. Learned about mesh independence. Thanks.

[@luiseduardojatarmontano1983]

that is a great explanation!!! Congrats/ and how could you deal with stress singularities in complex models?

[@gia_minh_]

you are a diamond in the forest of Indian guys :))

[@gregeconomeier1476]

Great videos! On another matter, it would be interesting to hear the pros and cons of the different bow designs currently used in marine architecture.

[@shwetaattarde4062]

Very informative.thanks sir👍

[@myparadiseonbantayanisland9030]

You should first explain what "mesh" and FEA are.

[@MrAbah105]

Amazing channel

[@andreaspoppe3124]

With the "right" size of meshing you can "adjust" the result if you want to :-)

[@mohitbisht8503]

What if we keep getting result value lesser & lesser even beyond the accurate real value with further mesh size decrement.
For e.g. 6mm mesh size give the result equal to real result but any lesser than 6mm mesh size give result way different from the real result.

[@DatawaveMarineSolutions]

Then it isn't strictly an issue with mesh size. It could be that your mesh restructures below 6 mm, resulting in poor element quality. Or possibly part connectivity issues.

[@akshaymohan2287]

thank you sir❤

[@Freakybananayo]

this is pretty entertaining.

[@Bassoomamor]

Thanks.

[@JaZoN_XD]

Question is, how exactly do you find this convergence mesh size?

[@DatawaveMarineSolutions]

There are various techniques for a mesh convergence study. I still need to create a video about mesh convergence.

[@3-DtimeCosmology]

I'm really awesome. 😎

[@arionfar]

Wrong! below 50mm results change a lot. they actually change more than above 50mm. It continuously goes down!

[@DatawaveMarineSolutions]

Not what my data shows. I would love to see your data that supports that argument.

[@arionfar]

@@DatawaveMarineSolutions Your data is constantly decreasing. how can you judge based on that?

[@DatawaveMarineSolutions]

@@arionfar I'm not sure what you mean by constantly decreasing. Can you please refer to a specific graph or part of the video.

[@arionfar]

@@DatawaveMarineSolutions I mean the element size verses stress. how did you conclude about the mesh?

[@DatawaveMarineSolutions]

@@arionfar As stated in the video, that graph showed that below 50 mm, the data showed very little variation in the Y-axis, which was the Von-Mises stress. I don't normally use statistics to analyse numerical methods, because I don't like to apply numerical methods include any random variables. But for a comparison in this case, the coefficient of variation (COV) for the bottom five data points was only 0.0051. Where the top four data points had a COV of 0.0715, 14x larger. Or if you compare the bottom four data points to the entire data set, the entire data set had a COV of 0.0616, 12x larger than the bottom five data points.

[@DerekWoolverton]

Color correction does matter. Either that, or you were suffering from some severe case of Venus flu.

[@DatawaveMarineSolutions]

Yes. That was one of my older videos when I was very new to video editing. (Still new to video editing, but at least I know about color correction now.)