General Mechanical

General Mechanical

residual stress distribution post welding

    • latnuk
      Subscriber

      I am trying to reproduce results from a research paper. But my results are strangely opposite. I am trying to simulate residual stress after welding. First, I found the temperature distribution in transient thermal, which is ok. After that, I used the temperature distribution as a load in transient structural. And here, the problem starts. It is expected to get tensile stress in the weld line's vicinity and compressive away from it. But my results are exactly the opposite.

      Here is the temperature distribution(I am using a dual heat source of Goldak type)

    • 1shan
      Ansys Employee
      ,nHave you checked the direction of stress reported in the paper (whether his Y axis and your Y axis are the same?). Generally, if a body is under tension in one direction, it gets compressed in the other perpendicular direction and vice versa due to the Poisson ratio.nRegards,nIshan. n
    • latnuk
      Subscriber
      Thank you for your reply.nI think the direction is not the problem here. The general trend is tensile residual stress along the weld line. But I am getting just the opposite.nAfter posting this question, I made some progress and found my simulation was not actually converging in transient thermal itself. But I don't know why Ansys was not showing an error. nI am not much experienced to interpret this plot and find a way out. Can you please help me out? Like what improvements are required, whether meshing is not proper or my number of steps have some problem, or my APDL command has some problem?nThank younLatnuk.n
    • latnuk
      Subscriber
      ,can you please look into it. nAdvance thank you to both of for your time. n
    • 1shan
      Ansys Employee
      ,nFrom the plot it looks like the solution is getting converged so that isn't the issue. Have you simulated the structural analysis post welding (t=20) or right from t=0 ? The thermal load should be from t=20, post welding when the plate contracts.nRegards,nIshan.n
    • latnuk
      Subscriber
      Thank you for your reply. n1) I haven't simulated the structural analysis post welding(t = 20). I will try and let you know the results after doing so. n2) I think I have figured out something more. Actually, I was not including bilinear plasticity in materials property(which is crucial in problems where stresses are exceeding yield stress). These things I was unaware of. But by including plasticity, I have some genuine force convergence issues.n3) I am not using temperature-dependent material properties for initial analysis. Once I get the residual stress pattern right, I will include them too.n4) Force convergence plotnJust after it, the simulation is throwing convergence error.n5) Displacement convergence plotnWhich is doing fine till now. n6) Newton-raphson residualn7) Ansys is showing some recommendations too.nnPlease, help me solving this convergence issue.nRegards,nLatnukn
    • latnuk
      Subscriber
      Finally, I solved my problem with my effort. Thank you all for commenting. It was an excellent learning experience. nIf anyone in future faces the same difficulty, I am always here. Just comment down; we will discuss it out. nRegards,nLatnukn
    • AndreMiranda
      Subscriber
      Hi Latnuk, nHave you solved the non convergence problem in the structural analysis?nI'm facing the same problem.nPs. I'm already considering the non linear temperature dependent material.nThermal analysis => OKnStructural analysis => errornnnn
    • latnuk
      Subscriber
      I am a beginner in Ansys. But within my understanding, I can suggest the following. n1) Try finer meshing near the weld line. You can use the 'sweep method' and apply a bias—something like below.nn2) I have observed that temperature-dependent plasticity properties sometimes causes non-convergence(reason not known to me). I suggest you use constant bilinear or multilinear plasticity first to check. After that, you can increase complexity.n3) Keep large deflections 'ON' in the analysis setting of the structural analysis.nnRegards,nLatnuknn
    • Shekhar
      Subscriber

      Hello Latnuk,

      I am also facing the same issue.

      Well, the temperature profile is correct with a maximum of 10% error between simulation and experiments.

      Besides the stress is completely opposite.

      Will you give me a hint about the right procedure?

Viewing 9 reply threads
  • You must be logged in to reply to this topic.