Physics-Based Predictive Model of Lack-of-Fusion Porosity in Laser Powder Bed Fusion Considering Cap Area

This work proposed a computationally efficient analytical modeling strategy to calculate the product porosity in laser powder bed fusion (LPBF) induced by a lack-of-fusion defect, with the consideration of cap area in solidified molten pools, influence of powder bed characteristics on material properties, and un-melted powders in the lack-of-fusion portion. The powder packing pattern and powder bed void fraction were estimated by an advancing front method and the technique of image analysis. The effects of powder bed characteristics on the material properties were considered by analytical models with solid properties and powder bed void fraction as inputs. A physics-based thermal model was utilized to calculate the temperature distribution and molten pool size. The molten pool cross section in transvers direction was assumed to be dual half-elliptical. Based on this assumption and molten pool size, the geometry of the molten pool cross section with cap area was determined. The overlapping pattern of molten pools in adjacent scan tracks and layers was then obtained with given hatch space and layer thickness. The lack-of-fusion area fraction was obtained through image analysis of the overlapping pattern. The lack-of-fusion porosity was the multiplication of the lack-of-fusion area fraction and powder bed void fraction. The predictions of porosity under different process conditions were compared with experimental results of 316L stainless steel and showed a better predictive accuracy than the predictions that did not consider cap area. The proposed analytical modeling method has no numerical calculations, which ensures its low computational cost. Thus, the proposed model can be a convenient tool for the fast computation of lack-of-fusion-induced porosity and can help the quality control in LPBF.

Automated summary

This study proposed an analytical modeling strategy to calculate product porosity in laser powder bed fusion, taking into account various factors and utilizing a range of methods and models for computation and prediction.