Predictive tools for the cooling rate-dependent microstructure evolution of AISI 316L stainless steel in additive manufacturing

Abstract

This study presents a thorough evaluation of advanced predictive tools for the solidification microstructure of AISI 316L stainless steel, taking into account both chemical composition and cooling rate. In order to achieve a broad range of cooling rates comparable to prominent additive manufacturing processes, autogenous laser welds were carried out on the steel plate by means of powderless directed energy deposition (DED) and selective laser melting (SLM) machines. Computational analysis of cooling rates and metallographic investigations revealed that DED welds solidify at moderate rates, exhibiting a primary ferrite (FA) solidification mode. Conversely, SLM welds solidify at significantly higher rates, showing dual solidification modes: initial austenite (A) at the fusion boundary and subsequent ferrite (F) at the interior. The WRC-1992 constitution diagram predicts the FA mode in accordance with the findings in the DED welds. An implicated Schaeffler diagram, proposed for a comparable cooling rate, also predicts the same mode. An artificial neural network model, referred to as ORFN, consistently predicts the variation of ferrite content within the DED welds based on the cooling rate variation. The dual mode observed in the SLM welds aligns with established knowledge regarding the alteration of stainless steel solidification under extremely high cooling rates. While the occurrence of the A mode at the fusion boundary is anticipated according to implicated Schaeffler diagrams for extremely high cooling rates, the transition to the F mode is not addressed explicitly.

Keywords

• Stainless steel
• Solidification
• Mechanistic model
• Constitution diagram
• Additive manufacturing