Scalable simulation approach for faster results

Simufact Welding provides the flexibility to select the proper approach for a given simulation goal.

How model simplifications work in Simufact Welding

A fully coupled transient thermo-mechanical simulation works in the way that in each time increment a temperature field is calculated. This thermal calculation is – compared to mechanical one – relatively cheap in terms of calculation time and scales nicely with the number of used CPU cores. If material with phase transformation is used, the phase proportions are also calculated during the thermal step, making it a thermos-metallurgical step. After each thermal step follows the mechanical one, where thermal strains are calculated out of the already given temperature field and acting as mechanical loads for distortion and stress calculation. This mechanical step is relatively expensive in terms of calculation time.

If the model is nicely converging and thermal calculation is delivering good realistic results, it is usually possible to replace the above described full coupling with a weak coupling, meaning, that not every thermal increment is followed by a mechanical one, but only every n-th increment, n being some number greater than 1. In converging models with sufficiently small time increments the loss of accuracy is usually relatively small while the calculation time decreases rapidly. The level of decoupling can be defined in Simufact Welding for each load case type (welding, cooling, heating and so on) separately for each process.

In addition to the definition of decoupling it is also possible to reduce the output data to coupled increments only to save the needed hard disk space. This approach does not change the model itself, only the level of coupling.

Further possibility to significantly reduce the needed calculation time is the usage of so called thermal cycle instead of a transient heat source. A transient simulation, where the heat source is represented by a moving heat flux, needs a lot of time steps in order to represent the movement of the heat source. It is however possible to replace the heat flux by a prescribed time-temperature-curve acting on the complete weld line at once. This approach shows good distortion results for problems where distortions are caused by shrinkage, but neglects most of upsetting effects, however reducing the calculation time significantly.

Further simplified approaches are also neglecting the welding sequence.

Single-shot solution automatically replaces all heat sources in the model with thermal cycle and also recalculates welding duration of each weld line as well as deactivation times of fixtures, applying all heat sources at once. Depending on the particular model the result quality can be sufficiently high, or rather low. For an already created model the remodeling effort is rather low: the user has to turn the single-shot solution on in the process settings and also select a thermal cycle table to be used for heating.

A further step is the Shrinkage approach, only usable for processes, where all weld lines are also having filler material. In this approach the non-linear material behavior and heat transfer between filler elements and base material is neglected. Filler elements experience a thermal load leading to their shrinkage, the shrinkage leads to deformation of the assembly. The result quality is highly dependent on the particular model, there are models that are well-suited for such approaches and there are models that will produce somewhat imprecise results. The activation of this approach works just by selecting the shrinkage approach in the process settings. Everything else is calculated automatically by the Simufact Welding GUI.

Recommendation: Where to use simplified approaches

Also rather huge models can be calculated within a rather short period of time.

For resistance spot welding Simufact Welding offers 4 different approaches to speed up the calculation. Beginning with most precise transient fully coupled approach to single-shot and purely mechanical approaches they offer the possibility to vary result quality and calculation times according to the simulation goals.

The table below allows the selection of appropriate simplification according to the goals to be achieved in the simulation.

Similar to transient approaches in thermal joining, transient coupled simulations in resistance spot welding are representing the complete coupled physics behind the welding process, including heat generation due to Joule’s heating as well as metallurgical and mechanical simulation with respect to welding sequence.

In general, the decrease of calculation time here is directly connected with the number of spot welds in the model. The more spot welds are present; the more calculation time can be saved. Distortions and other result values are very similar for each spot. Main differences to time-dependent model with respect to welding sequence are:

• The shunting effect is neglected
• All spots are welded on the undeformed geometry

The Mechanical approach, activated in the process settings (as shown in above figure), requires an already defined model. It neglects all coupling effects, only applying a force to electrodes and gluing the component surface areas touching each other between the electrodes. The welding sequence is respected. The approach shows nice results if it comes to overall distortion of the assembly, detailed results in weld spots differ because the material does not experience any temperature, so especially if it comes to evaluation of material properties and stresses inside the spots, this approach is not recommended.