I am simulating bolted plates using face-to-face contacts and pre-tension loads on bolts. Due to convergence problems on larger models, I have tested a lot of configurations in simple models. This was done to indentify which parameters influence the analysis time and results.
One of my findings is that when decreaseing the σ infinity parameter in the surface behaviour, the total time increases. I have done this to follow the advice that it should be 0.25% of the maximum expected stress in the model.
I realised that the normal stress on the contact surface and the contact pressure field are not affected by this parameter. As I understand it from the documentation, it is due to the truly bilinear pressure-overclosure relationship: The parameter c_0 being null, any negative overclosure will remove contact elements, allowing for some clearence to appear, and thus no resulting tension.
However, I connot explain why the calculation times are affected. As a test I increase the σ infinity parameter by a factor 100 and the calculation time was more than halved.
Does anyone have an explanation? It seams silly to me to enter a huge value of σ infinity thus to decrease my analysis time without a valid reason to.