I'm trying to simulate an over-reinforced concrete under 4-point bending test and I aim to simulate the compressive crushing process of concrete. However, when I expect to get the compressive and tensile damages (DAMAGEC & DAMAGET) somewhere between two concentrated loads(where the bending moment is maximum), I unexpectedly notice that it is begun below the loading plates and when the load is further increased, over-stressed elements are also seen around the supports as well. It seem when the plates penetrate in the concrete the damage is started. This process disrupt all the damaging mechanism and delivers the results unlike the experimental tests. I have come up with a multitude of workarounds as below to solve the problem:
-Defining surface-to-surface contact between surfaces(Hard contact & Frictionless).
-Trying Tie Constraint.
-Using Cylindrical shapes instead of the cube ones.
-Suppressing all the supports and applying the boundary condition directly to the beam.
But unfortunately the problem has still remained unsolved. Is there any solution to prevent damage around the plates?
Please use full elastic elements in these regions and you will see better results from that.
Dear Sir,
This case called stress concentration (local damage). It happen under concentrated loads (applied loads and reactions). To avoid this process, assume a new material, which is behave as an elastic non-damage material. Use this new material below the applied loads and above the reactions.
Best wishes.
Thanks for your responces,
I'm well-aware of this case(sress-concentration) and the approaches that I pointed out in the question were to deal with this issue. I already thought of an approach same what you proposed but I'm afraid that this time local damges emerge in the elements close to non-damage elements.However I will try your suggestion and will get back to you.
In the real world, we actually don't have a ''point'' support (other than the knife edge support used in precise weigh balances). In reality the support will have a finite width, which, in FE terms, may be spread over a couple of rows of elements. I'ts because we assume it to be concentrated along a node line that we are getting such unrealistic stress concentrations.
Rahul Leslie
- Badges
- Science topic