WELDING RESEARCH A Argon B Vacuum C D Fig. 9 — 3D CT images and volumetric porosity plotted as a function of weld depth in Ti for the following: A, B — The HP laser weld made in Ar; C, D — the HP laser weld made under vacuum conditions. mospheric pressure. Also note that the fit to the Weibull plot (log-normal distribution) has some important implications about the distribution of the data, or pore sizes in our case. Recent work has studied the significance of the upper tail of the distribution that deviates from a log-normal distribution, showing how the distribution with a long tail will underpredict the importance of large pore sizes, while an upper distribution with a short tail will overpredict the importance of large pore sizes (Ref. 15). The combined data for all the welds shown in Fig. 6 indicate the log-normal distribution doesn’t fit the data all that well, where the actual pore size data curves over the best fit line, and the long upper tail indicates many larger pores than would be expected from the Weibull parameters alone. This type of fit indicates there are most likely multiple porosity mechanisms active in the different types of welds studied here, and that a grouping of all the data together may not be the best representation of porosity in each individual weld. Analysis that follows will take a closer look at the pore size distributions of the different types of welds and materials. A B 426-s WELDING JOURNAL / NOVEMBER 2016, VOL. 95 Laser Welds in Nickel The porosity distribution statistics are useful in quantifying the higher number density of small pores in the welds, but additional information can be acquired through visualization of the pore locations within the weld from the 3D CT renderings. Figure 7 compares CT rendered data in Ni for the porosity that formed in the HP laser welds made in Ar and vacuum conditions. The weld made in Ar is shown in Fig. 7A, and has the highest porosity of all the welds made in this study. The porosity is located in the lower part of the weld and has a globular morphology that is very similar in appearance to the porosity observed in laser welds made in Ar in a previous study with a different laser (Ref. 1). Another way to illustrate the porosity distribution is to integrate all of the porosity on a given CT slice through the weld, and then plot the porosity contained in each slice from one side of the weld to the other. This is easily done from the 3D CT data, and can be performed in any orientation relative to the weld. Figure 7B plots the porosity for the laser weld made in Ar from the top surface (x = 0 mm) to the root of the weld (x = 10 mm). This plot shows there is little porosity in the top 1 mm of the weld, and that the remainder is concentrated between 1 and 3 mm below the surface. Comparing this to the metallographic cross section in Fig. 5A confirms the location of a large pore about 2.5 mm from the top surface, located near the center of the keyhole portion of the weld. Fig. 10 — Histograms showing the pore size distribution in the Ti welds for: A — Laser welds made with Ar shielding gas; B — the laser vacuum welds. The inset plot in (A) shows the Weibull curve fit to the data for the welds made in Ar; however, the low porosity in the laser vacuum welds did not yield enough data for fitting.
Welding Journal | November 2016
To see the actual publication please follow the link above