Hi everyone,
I am using FEA to find residual stresses in a 4-pass weld plate. I am using the Thermal Stress (Solid Mechanics/ Heat transfer in solids coupling) option.
I used 4 volumetric heat sources to describe the input energy of the weld and I created an analytic rectangle function that is described simply by an1(t) = rect1(t), defined between 0 and 1000. This is to allow me to "turn on" the heat source when the energy is being applied to a weld bead. When the weld is cool enough to call on the subsequent weld bead heat source, I use an1(t-300) if I wait for 300 sec.
Problem:
When I compute the solution and I check my plot for Temperature vs time at a point in the weld, there is only 1 heat source that is being applied while the others are not incorporated into the solution. Strangely, there have been times when I compute the solution without changing much and I might get 2 heat sources applied.
I'm new to FEA, but I believe it might have to do with the solver itself and the omission of certain time steps due to tolerance settings. Would it be possible for the solver to skip the time step when the heat source is being initiated? Would changing the relative tolerance be the issue? Originally at 0.01. This is going to be the next step I try. Is there anything else I should be looking at?
regards,
Dominic
I am using FEA to find residual stresses in a 4-pass weld plate. I am using the Thermal Stress (Solid Mechanics/ Heat transfer in solids coupling) option.
I used 4 volumetric heat sources to describe the input energy of the weld and I created an analytic rectangle function that is described simply by an1(t) = rect1(t), defined between 0 and 1000. This is to allow me to "turn on" the heat source when the energy is being applied to a weld bead. When the weld is cool enough to call on the subsequent weld bead heat source, I use an1(t-300) if I wait for 300 sec.
Problem:
When I compute the solution and I check my plot for Temperature vs time at a point in the weld, there is only 1 heat source that is being applied while the others are not incorporated into the solution. Strangely, there have been times when I compute the solution without changing much and I might get 2 heat sources applied.
I'm new to FEA, but I believe it might have to do with the solver itself and the omission of certain time steps due to tolerance settings. Would it be possible for the solver to skip the time step when the heat source is being initiated? Would changing the relative tolerance be the issue? Originally at 0.01. This is going to be the next step I try. Is there anything else I should be looking at?
regards,
Dominic