Introduction: Advanced low alloy steels have regained interest due to their high strength and toughness combined with low cost. AF9628 is one such steel with potential applications in a range of industries due to its good balance of strength, hardness, toughness, and affordability.
Performance optimization is needed to meet the growing need in large format AF9628 additive manufacturing (AM) for a range of feature scale and reduced qualification costs. The welding metallurgy of AF9628 AM deposits and heat affected zones was not well understood. A large program is underway to develop metal direct energy deposition (DED) build models for powder laser (PL) and gas metal arc (GMA) processes that maximize affordability based on feature complexity and scale. A goal is to evaluate whether temperbead approaches can be used to optimize AF9628 build properties with lower cost heat treatments. A computational design of experiment (CDoE) platform for process-feature-microstructure-property optimization is being developed for DED AM of AF9628 steel components. A key element of this platform is the establishment of quantitative process-property-microstructure relationships for a range of DED deposit sizes, heat inputs and interpass temperatures. Data generated from DED samples and physical simulation experiments will be used to validate and refine computer simulated process-property-microstructure relationships. Experimental Approach: To develop AF9628 DED technology, research was performed to develop consumables, metallurgical properties, DED parameter models, and computational models. Using samples removed from AF9628 forgings, continuous cooling transformation (CCT) diagram were developed using the Gleeble for thermal cycles representative for GMA-DED and PL-DED deposits. AF9628 powder consumables were produced by atomization of AF9628 forging materials. AF9628 metal cored wire were developed by two suppliers to support commercialization of AF9628 DED. Powder laser and two modes of GMA, Cold Metal Transfer (CMT) and gas metal arc pulse (GMA-P) DED were used to make build samples for property tests. Preheat and interpass temperature tests were varied and used to eliminate hydrogen cracking susceptibility for single-pass per layer “walls” and multi-pass per layer “blocks”. The interpass temperature was robotic controlled using infrared sensors that evaluated each deposit. Thermal histories of CMT and GMA-P DED processes were recorded in-situ to relate the multiple reheat tempering thermal cycles to the resulting build properties. The effects of post-build heat treatments, as post-bake for hydrogen desorption, stress relieving, and post weld heat treated (PWHT), were also evaluated. Hardness maps were developed that covered 6 to 7 DED layers along the Z-build direction. Both X- and Z-tensile specimens were extracted normal and parallel to the build direction, respectively where Z- specimens had 9-10 layer gauge sections. Digital image correlation (DIC) was used during tensile testing to understand microstructure behavior under load. DIC allowed quantifying local mechanical property variations along the build direction, generated by bead tempering effects of the GMA-DED processes.
Metallurgical characterization with light optical microscopy (LOM), scanning electron microscopy (SEM), and electron beam scattering diffraction (EBSD) is being performed to identify the effects of multiple reheats on the builds microstructure. Results & Discussion: Both PL and GMA property builds were made to evaluate tensile, Charpy impact, hardness maps, chemical analyses, and metallurgical characterization. The martensitic start temperature ranges from 341 to 356 °C depending on the cooling rate. The initial goal was to use preheat & interpass temperatures of 125 °C that were below the martensitic finish temperature, ~150 to 160 °C to ensure transformation and subsequent temperbead response. However, a higher interpass build temperature, 250 °C was necessary to eliminate hydrogen cracking in highly retrained block builds. This increased the complexity of the temperbead models to include effects of delayed transformations due to higher interpass temperatures. The thermal histories measured in the CMT and GMA-P DED processes were compared to understand tempering effects in as-deposited condition. The GMA-P build deposits were found to experience a greater number of tempering reheats compared to the CMT deposits. The GMA-P process generated two subsequent austenitizing reheats followed by 6 tempering reheats below the Ac1 temperature, 829 °C. In contrast, the CMT process generated only one reheat above the Ac3 temperature, 905 °C, and 3 tempering reheats. The smaller bead size and greater heat input in the GMA-P process led to increased number of tempering reheats. Both processes generated a banded hardness pattern with hard and tempered regions along the build direction. Mechanical property (tensile, Charpy impact, hardness) of PL, CMT and GMA-P builds was in-progress and will be presented. As-deposited properties will be compared to stress relief and full heat treatment properties of DED builds. DIC will be used to relate the banding in hardness to variations in strain and strain distribution of each sample. Further support will be provided by advanced metallurgical characterization, relating hardness and mechanical property variations to the local microstructure produced in the AM process. All this data is being incorporated and used to develop a computational model that will determine properties over a large range of process-feature-thermal conditions. Conclusions: Consumables and DED procedures have been developed that produce deposits that are resistant to hydrogen cracking, meet AF9628 composition requirements, and offer a range of feature scale capability to build large format structures with the most affordable process. CCT diagrams have been developed that accurate predict on-heating and on-cooling transformation properties. Computational models are under development that integrate AF9628 metallurgical properties with DED process conditions that will accurately determine properties for a wide range of process-feature-thermal (deposit size, heat input, interpass temperature, cooling rate) conditions. Post-weld stress relief and heat treatments will be developed to tailor properties of specific structures. The discovered relationships from this work will help further develop feature-microstructure-properties relationships for AF9628 over a range of DED processing conditions. Additionally, transmission electron microscopy (TEM) will be used to characterize carbide formation in AM microstructures and further develop the process-microstructure-property relationships. Key Words: Directed Energy Deposition, DED, Additive Manufacturing, AM, Build Models, Property Models