Heat-flow simulation of laser remelting with experimenting validation
- PDF / 1,660,171 Bytes
- 9 Pages / 603.28 x 783.28 pts Page_size
- 42 Downloads / 224 Views
I.
INTRODUCTION
LASER remelting is of commercial interest because of its ability to alter with accuracy the properties of very localized surface regions without reprocessing the material as a whole. However, it is this scale of operation which makes in situ measurements of the process variables so difficult. Most information is obtained by studying aftereffects, such as the change to the microstructure and the mechanical or chemical properties of the treated surface. Therefore, modeling the physical process can yield much insight into the complex phenomena occurring within the region activated by the moving high-power laser beam. Furthermore, modeling work can reduce substantially the time required for process optimization, scaleup, and control. However, process modeling cannot be made in isolation from experimental work. The processing parameters which control the energy input to the workpiece influence directly the extent of the remelted region and thus must be determined with accuracy if the simulation is to give quantitative results. In particular, a special technique has been developed t~l for measuring the absorption, thus removing a considerable degree of uncertainty. The thrust of this paper is therefore in the systematic verification of a numerical model developed for the laser remelting process. Although a number of conduction-based heat-transfer models for laser surface treatment, such as welding and remelting, have previously been published, [2-5] only a few t4,5~ compare the simulation results directly with experiments. Those in Reference 4 consider only the width of the track for a single processing condition and this was overpredicted by 50 pct. Using the steady-state model which was the forerunner of the transient model presented here, Rappaz et al.[5] also calculated results which considerably overpredicted the melt widths, although the depths and the microstructure were accurately simulated. A.F.A. HOADLEY, Scientific Collaborator, M. RAPPAZ, Group Leader, and M. ZIMMERMANN, Graduate Student, are with Ecole Polytechnique F6d6rale de Lausanne, D6partment des Mat6riaux, Laboratoire de M6tallurgie Physique, MX-G Ecublens, 1015 Lausanne, Switzerland. Manuscript submitted September 22, 1989. METALLURGICAL TRANSACTIONS B
Since these publications, many of the authors have now considered the effects of Marangoni convection and the free surface shape, thus further extending their simulation capabilities, t6-1~ Again, only References 6, 9, and 10 make a direct comparison with the experimentally observed trace. Kou and Wang [6] achieved good agreement but for a single traverse speed of the laser beam of 4.2 mm/s which is at least two orders of magnitude lower than those speeds considered in this work. Furthermore, their measurement of the absorption by a calorimetric technique yielded an average value which did not distinguish between liquid and solid phases. Finally, the energy distribution was not known and a Gaussian distribution was assumed. Also, at low speeds, Paul and DebRoy t91 compared the melt p
Data Loading...
