I hope you have time to help me out to understand using contacts in SOL101. I am trying to understand when SOL101 is not sufficient for contact analysis and when you need to move on to SOL106 or SOL601,106(?). I I have made simple case with a beam and 3 contacts as shown on sketch. I have search the forums but could not find anything
Lets say you have a beam with 3 contact definitions as shown on the sketch. The following contacts are defined:
Contact no. 1: Beam is resting on support. No gap. Allows for sliding and uplift but is supported in negative Z direction.
Contact no. 2/3: Horizontal support with a gap of 5mm. Since there is a gap the beam needs to bend before contact.
Lets assume there is defined friction in the contacts.
To make things more complicated a very low temperature is applied on the beam (supports not affected) so the beam will shrink and there will be sliding on contact no. 1 beacue of the contraction of beam in X direction as shown on the sketch. Lets assume only gravity is applied on the beam and will create a load on the contact area. I would assume this to be a static linear problem and SOL101 could be used to solve this problem?
But lets say that the sliding plates are 100x100mm and the sliding distance is 30mm because of thermal effects. So the distance between nodes (assume coincident before loading/thermal) after loading is 30mm. Is this to much for SOL101? Large sliding distance relative to size of sliding plate?
Lets also combine a horizontal load (Y direction) and because of thermal effects the gap (5mm) is now larger and load is applied (Y direction) the beam will move towards the contact and make contact (somewhere) on the support. Need also to account for shrinking in the X direction of the beam. Can SOL101 be used or does this need to be considered as nonlinear contact problem?
Lets assume that the loads are applied slowly (static) and there are no geometrical or material nonlinearities in this problem and no ridig body errors.
As with any assumptions made in a finite element analysis, this is an engineering judgement call.
Strictly speaking, when contact is added to a solution, the solution is no longer linear. In SOL 101, the nonlinearities are strictly related to contact - there is no global stiffness update (geometric nonlinear), no material nonlinearities, etc. Given that, SOL 101 "Linear" contact would typically be valid for conditions where the contact pair regions are initially touching, where there is little or no sliding and where the contacting meshes are relatively stiff. A good example is contact between a cylinder head and engine block. These are stiff cast components being bolted together with very little deformation. There is no sliding because dowel pins keep everything aligned.
In general, "linear" contact is OK as long as the sliding is small enough that the contact pairs remain valid and as long as the deformation does not alter the contact normal directions significantly. If either of these occur, then the contact elements themselves need to be reformed. This only happens in a nonlinear solution sequence.
Thank you for reply. Basically what you are saying is that SOL101 is only valid for contacts where there are no or little sliding. If I use my example again:
- Contact no. 1: Has (assuming) 30mm sliding. Lets say the beam is 15000mm long. The sliding is small compared to length of beam, but I would say it is large compared to sliding area if the sliding area is approx 100x100mm.
- Contact no. 2/3: Open gaps before loading. No initial touching.
- Assuming gravity (distributed) load on the beam. The shape would look like a banana and this will probably create small gaps on contact no.1. Before loading all is touching and after loading the contact area will probably be reduced.
Based on your reply I would treat my example problem as a nonlinear contact problem and SOL101 would no be sufficicent to solve this problem correctly? I am still assuming no geometrical and material nonlinearities.
Basically SOL101 is ok to solve typical problems where you have bolted flanges with bolt preload. No sliding is expected but may have small sliding if the loads overcome the friction and preload?