I am not sure, but I think if you select a bunch of nodes and then apply the same force to them, instead of applying the force pointly or just on one node, it'll work out ;since, you are distributing the force into the proximity nodes too this way , instead of applying all the force on one node, which logically should result in a high stress concentration in that point.
I think Arash is right! I also can suggest to increase the mesh density around the points of force application and to use an incremental increasing mesh size around the points.
The solution is to apply the force over an area rather than at a single node. Increasing mesh density will only make the stress appear higher, as the effective area covered by the single element the force is applied to will become smaller with increasing mesh density.
If you want to know what the stresses are like in the area where the force is applied then change the load from being applied to a point to being applied to an area, but if you're interested in the stresses in another part of the model then you could decrease the mesh density in the area where the force is applied, which will make the stress concentration appear smaller.
I agree with Barry, may you applied point load. that is why Stress at that region is more. If you want to reduce stress concentration increase the area of application of load (or)decrease the mesh density at that location.
1. Redistribute the load along lines (if possible) or number of adjunct points, and check the final solution contour distribution, is it logical or not?
2. Divide the total applied forces for specific length that the final solution will be per distributed load other than point loaded..
3. You may need to check the effect of thermal change in the tip due to high friction effect and this effect may produce thermal wave that redistribute the mechanical and thermal stresses...