Final Report Summary - PCPLAS (Prediction and control of porosity in laser welding of non-ferrous metals)
Executive Summary:
The PCPlas project aimed to improve the fundamental understanding of porosity formation in the laser keyhole welding of non-ferrous materials (an example shown in Figure 1), of importance to the light-weighting of transport components.
Figure 1 Example of laser weld porosity. Longitudinal section of a non-optimised weld in an aluminium alloy.
To do this, the project:
• Developed numerical process modeling of selected materials and
• Validated these models with respect to experimental results and observations, in an effort to devise means of reducing porosity contents and improving weld qualities in the future.
The work carried out and the results achieved by the PCPlas project include:
1) A literature review on numerical modeling of the laser welding process to date, to indicate the steps by which modeling could be advanced by the current project.
2) A selection of aluminium alloy AA5083 and titanium alloy Ti-6Al-4V as examples of non-ferrous materials for the project, given their widespread use in industry.
3) The development of a first-stage model (excluding keyhole formation) applied to the laser welding of aluminium alloy AA5083, to take account of, and identifying the effects of, significant factors including heat input, convective heat transfer, and latent heat. A Gaussian-based heat source model of the laser beam was developed for this model which, after validation by selected experimental results, could then be used to model partial and full penetration weld shapes and fluid flows over a wide range of conditions. One example is shown in Figure 2.
Figure 2 Example of agreement in weld pool shapes from first-stage numerical model and experiment.
4) The development of a more complex second-stage model (now including an evolving keyhole), applied to the laser welding of titanium alloy Ti6Al4V. A volume of fluid (VOF) method was used to trace the keyhole surface during model run time, and a second Gaussian heat source developed to model the laser beam, of a depth that varied as modelled keyhole depth changed dynamically with time. An example is shown in Figure 3.
Figure 3 Example of keyhole evolution predictions (from keyhole opening to closure) from second-stage numerical model.
5) High speed camera video imaging was used during actual welding of AA5083 and Ti6Al4V was used to monitor the actual shapes, dimensions, and flow patterns of the keyhole. An example is shown in Figure 4, with the keyhole and weld pool boundaries marked. These observations were then used to validate the second-stage model, as well as to gain an insight in to the actual welding process under different sets of conditions.
Figure 4 Example of high speed video image of laser welding
6) The trials showed that porosity contents of full penetration melt runs in AA5083 decreased with increasing laser power, or decreasing welding speed. By comparison to differences in the predictions of the simple model (without keyhole), it appeared that these porosity levels were more controlled by weld pool dimensions than any differences in fluid flow speeds (albeit distant from any keyhole) between different sets of welding conditions.
7) The porosity in Ti6Al4V melt runs decreased with increasing power and/or power density (and correspondingly increasing welding speed). Second-stage (with keyhole) model predictions failed to predict keyhole instabilities as the cause of porosity. Instead, lower power conditions led to predictions of fluid flow vortices and separations (or voids forming) behind the keyhole, which appeared absent when using higher powers. This situation appeared analogous to a Reynolds number controlled turbulence. In such cases, the use of measures to reduce keyhole diameter (such as a more focused laser beam) were postulated to then be effective in reducing porosity content.