Contact analyses

Hi :slight_smile:
First of all, I would like to say that I am new to the field of FEA. I am not an engineer, but only a CNC programmer who wants to deal with FEA privately. I bitter therefore for some indulgence if I do not understand something directly.

I have in my attempts again and again objects where contacts occur. All the videos I found about PrePoMax (and also those I found about Abaqus where the contacts are defined similarly) use the type “hard”. But I am not sure if “hard” is the right way 99% of the time. Therefore I wanted to ask if someone can explain to me when I use which type. So for “hard”, “linear”, “exponential”, “tabular” and “tied”. In my books which I have read on the subject of FEA is only minimally addressed to contacts so I’m not sure when I have to use which type.
A second question is where do I get the data needed for the other types except “hard”? I have no way to test or measure anything myself.

I would be very happy if one would help me as a total beginner here :innocent:

You should check the CalculiX (and maybe also Abaqus due to the similarity) documentation first. Those contact types are described there. But there aren’t many general tips (you can find some here and on the CalculiX forum though) since the choice is highly case-dependent and usually requires some testing for a particular model. When changing contact stiffness you should check the results like contact pressure and penetrations.

One of the helpful threads on the CalculiX forum: Contact stiffness setting - CalculiX (official versions are on www.calculix.de, the official GitHub repository is at https://github.com/Dhondtguido/CalculiX).

Even after reading both documentation on the subject of contacts, I can not say to have understood in which cases I must apply which type. Also I don’t know where to get data for other types. Material properties can be found on the Internet, but I do not know where to find, for example, “Slope of the pressure-overclosure curve”.

Edit: However, with the link you attached, I also notice that it is not as simple as I thought it would be and the values seem to be estimated.

As I said, there are no general rules apart from the fact that contact stiffness is usually specified as a multiple of Young’s modulus of underlying materials. In each particular case, you should carry out a few tests if you want to change those settings from defaults. Usually, there’s no need to change them but you may want to get a better agreement with experiments or obtain a better (more regular) contact pressure distribution. Then using “softened” contact might be a good idea but you have to watch for excessive penetrations that may occur because of that softening.

To understand how those types of contact work, it’s best to take a closer look at the pressure-overclosure curves referenced in the documentation.

Usually, there’s no need to change them but you may want to get a better agreement with experiments or obtain a better (more regular) contact pressure distribution. Then using “softened” contact might be a good idea but you have to watch for excessive penetrations that may occur because of that softening.

Ahh, sounds like hard contacts are usually used. Probably for me then the type “hard” are sufficient. As I said, I’m a beginner and I don’t have any data from my tests, so there’s probably no reason to use a softer type for now :slight_smile:

indeed, it’s ideal contact condition and recommended. However, these type not easy to achieve and may have convergence problem. So, a linear type can be used instead of hard or switching to Mortar type.

also, keep using surface to surface type of contact when a quadratic element in model, not using node to surface types. Some disadvantage is consequently in computational effort become large.

2 Likes

Thank you, this is a good and simple explanation. I’ll take this as an orientation :slight_smile: :+1: