Initially Transient heat transfer analysis welding was conducted using ANSYS software by SOLID70 element. Next, thermal histories are given as a input for mechanical analysis to predict Distortion and residual stresses using SOLID185 element. Displacement and stress results are predicted only for single time step and not able to predict for all the time steps.
Guide me to conduct transient mechanical analysis using thermal histories output from transient thermal analysis.
Thermo- mechanical + welding processes are under developing in all FEM solvers more or less better or bad. You can write own procedure in Fortran for solution a problem but it hard. You use ANSYS - probably because it is available for you - but it can’t be the best solution at all. From many years experiences the most appropriate welding software is SIMUFAC WELDING. Using the software you are focused on welding problem not on Fem modelling problem in subject how to model this or this n ANSYS or other software – answer take software when the problem is solved or write own.
You can use ANSYS Parametric Design Language (APDL) in ANSYS. You will find how to use APDL in the software help package.
You can use SYSWELD software for Thermo-mechanical analysis of welding. In the tool box you will find help file each step by step. It involves both steady state and transient analysis of thermal, distortion and residual stress analysis.
If you need more help how to use SYSWELD you can mail me.
For complete thermo-elastic-plastic analysis there has to be provision of supplying the results of thermal as well as mechanical analysis from the previous step in to the next step along with the temperature distribution in the next step. I had used ANSYS long time back and do not recollect whether there was such provision. As suggested by someone else you can try using ANSYS parametric language if it helps. One way of testing it could be to use a very simple example (say a restrained plate subjected to uniform heating but rising (or falling) temperature with time) and to test whether both thermal and mechanical output of previous step could be supplied as input for next step. You may also have a look at the following article for an alternative approach for elastic (only) FEM analysis to predict distortion. It can be obtained through Google search. More relevant references are available in this article.
Development of a Methodology for Correcting Distortions
Produced by Welding Processes in Naval Structures by Adan Vega Sáenz
Universidad Tecnológica de Panamá COPINAVAL 2013
Bear in mind that it is typically very difficult to get meaningful residual stress predictions from thermo-elastic-plastic modelling of most welding processes. Realistically, you'll need elevated-temperature flow curves and dilatometric data for all materials involved as a minimum, or else it's just a case of 'garbage in = garbage out'. This kind if modelling can also be very time-consuming.
If you have never performed FE analysis of welding processes before, I'd advise thinking about what you need to get from it and why. For example, if you want to know the residual stress distribution for the purpose of assessing the structural integrity of the part, then for certain common types of weld you can often use a 'conservative' estimate (eg. as detailed in assessment standards such as BS 7910). If you instead want to use your FE model to estimate weld distortion then in many cases it is much quicker and more accurate to do a physical experiment.
If you are able to use APDL it is not so difficult.
The most important commands are 'lsread' and 'lssolve' which could be helpful to solve your problem.
- Badges
- Science topic