Have someone met a strange behavior of Load-Displacement curve in a simple 2D indentation problem

I am modelling a very simple 2D contact problem between a rigid wedge indenter and a deformable squared-shape specimen (general steel material) in frictionless contact case. I used an implicit function f to describe the rigid indenter.



I implemented Lagrange multiplier method for this contact problem and the contact condition is: Inside the contact zone of the deformable specimen, the node set is active if f(node)<=0 and the Lagrange multiplier lambda<=0.



I also generated an exactly the same model on ABAQUS in order to make a code2code comparison of the ouputs. And these are the things I see:



a. Displacement: I must say it is 99% similar considering displacement distribution and value between my implementation and ABAQUS results. (There is only a little bit tolerance at the displacement value).



b. Load-Displacement curve: the ABAQUS curve is very reasonable with a linearly increasing. However, it is so strange in my implementation due to the fact that the load is equal to zero in 3 first steps of an increasing displacement u.



I attached my model, and the 2 load-displacement curve outputs: one is my implementation and the other is ABAQUS result, so that you can see how strange it is.



Have anybody had experience with this before? Could you please help me to explain about it?



Thank you very much,

Best regards,



Minh