I am trying to run a foundation slab with multi-linear springs, representing below soil conditions, but running into "not converging" issue:
1. As a test run I modeled 2-ft thick foundation slab that is 180-ft x 60-ft. The only applied load is self-weight of the slab for a test run.
2. The soil condition below I am trying to model is a case where soil is liquefied and settles 3-in down towards the center. I am trying to apply multi-linear springs to imitate this condition where center of the slab sees no soil over first 3-in of depth but bears against the soil if the slab deflects beyond 3-in. This condition is interpolated from center of the slab towards the two ends in the long direction. I tried to do this by assigning a small stiffness value to the initial curve of the multi-linear springs then actual soil stiffness once displacement reaches the expected soil settlement.
3. After running the model I am getting warning message saying that analysis did not converge. It appears the program reads the model being unstable if any of the slab joint displacement lands within initial curve of the multi-linear springs, where the stiffness is value is very small, and the total applied load does not match with the total reaction.
Is there a command or a way to make this analysis happen and converge? In the real world, if there is gradation pattern of the soil settlement, I am expecting either slab deflects until bears against the settled soil or span between the higher elevation of the soils if the slab is stiff enough.
