A Statistical Study of Mechanical Properties From Mild Steel Welds Deposited via Gas Metal Arc Additive Manufacturing (GMAAM)

Abstract

Welding has been used for decades to create materials from weld metal that were machined to form components and used as a substitute for long lead-time castings, plate and forgings. Past terms like “shapewelding” or “shape melting” have been replaced with “additive manufacturing” to describe the process, but there is debate whether it should be treated as an additive/subtractive manufacturing process or a welding process followed by machining. Welding procedure qualifications verify weld metal properties. The properties of qualified welds are quite predictable when the welding parameters (variables) are controlled. The range of variables to be controlled and the allowable limits vary based on the risk tolerance of the application. These variable qualification limits are covered in various welding qualification codes and standards. In the past, code rules for weld metal buildup were used to qualify additive weld metal, but tighter controls are demanded today. Because of this, ASME Section IX developed and published rules in Code Case 3020 [1] for welding procedure qualifications that are specific to gas metal arc additive manufacturing (GMAAM). The Code Case 3020 qualification rules require testing weld metal properties at the highest and lowest cooling rates to be used in production. Code Case 3020 rules also require testing the thinnest wall section and a thick section for each of those cooling rates. The rules also require that all of the essential and supplementary essential variables used for weld metal joining and operator qualification in ASME Section IX continue to be followed. An ASME supported research project was launched to validate these rules. The project included a design of experiments (DOE), created by subject matter experts and vetted by an advisory committee consisting of designers, fabricators, consultants and metallurgists. Approximately two tons of weld metal was deposited using 24 different sets of welding parameter input configurations as set out by the DOE. Over 300 tensile specimens and over 500 Charpy V-notch (CVN) specimens were taken from various orientations and tested to characterize the weld metal properties. The data was analyzed to determine if the variables and interactions provided statistically significant prediction of the weld metal properties. The degree of isotropy and the extent to which manipulated variables predict weld metal properties are key findings. This paper examines those results, and discusses the relationships as they relate to new and existing code rules.