Why is it that after an entire day of trial and error, the meshing starts to act up? What was previously coming along nicely suddenly stops working the way it is supposed to. This has happened for the last three days, I'm debugging the various fatal errors that show up, do minor adjustments to the mesh and other things, then towards the end of the day things start to bug.
In this case I have a large assembly (1m x 2m x 0,1m roughly) consisting mostly of various sheet metal parts that I idealize as shell elements (one for each part) where I define all planar faces as quadrilaterals and all non-planar faces as triangles. The first time I tried it worked like expected. Then later on, re-meshing the model causes the faces defined as triangles not to be meshed at all in ONE part. Trying to figure it out I keep working with it, and after a while I attempt another re-mesh, with the same result but on all the shell parts.
I also notice that the if I run the simple mesh dialogue it runs, but doesn't generate any mesh. If I open the table and select all the entities it does mesh, but not with the triangles. Only the quads. Another peculiarity is that changing the mesh size in the simple dialogue doesent affect the values defined in the table.
Deleting the analysis, creating a new one and just copying the different idealizations, manual contacts etc. worked the first time, on tje first mesh attempt, but trying to mesh a second time renders the mentioned result. Retrying this process, renders the same result on the first mesh attempt.
Along the way here, the graphics disappears, and I have to hide the mesh to see the model, and then turning it back on. Somtimes all I see is some weird black line across the screen.
The only thing that apparently really lets you start over fresh is creating a new assembly and setting the model up from scratch. During the course of all this not one single error message appears. This is mighty frustrating.
My overall goal here is to simulate reality as closely as possible by running a non-linear steady state heat transfer with a difference in temperature if about 1000°C between the hot and cold side of a door and using the results in a non-linear static afterwards to determine how much the model will bend as a result of the heat expansion on the hot side. The intention afterwards is to do various modifications to the model to see what will inhibit the displacements most effectively etc.
How do I keep this thing stable long enough so I can continue my path of trial and error with the errors I get trying to solve the analysis, instead of starting over all the time? Is there some folder somewhere that acumulates junk as I go along that causes issues like these, or is it something else entirely?
