August 2025 Volume 7

FORGING RESEARCH

This concept is exemplified in Figure 11, wherein the geometry is a demonstrative example of a simple part with flat edges and a thick center., This can be produced via three different autonomous incremental forging paths. Path 1 has the Agility Forge incrementally form the part with dies hitting increasingly deeper into the workpiece at a single spot until the target thickness is reached, then the dies progress along the length of the workpiece to finish the forming. Path 2 has the dies hit to a certain depth, then progress along the length of the workpiece, then return to the original spot and hit deeper, etc., until the target thickness is reached. Each of these paths produced a part wherein the thick section, which experiences very little deformation, retains essentially the initial microstructure of the stock material, in this case, a 316L material with initial average grain size of 50 microns. Thus, in order to induce deformation sufficient to cause recrystallization throughout the entire part, including the thicker center region, a third path is proposed wherein we start with thicker stock material of the same average grain size and essentially “cog” it down to the size of the initial workpiece modeled for Paths 1 and 2. However, because it has already been worked to the point to induce recrystallization, the starting initial grain size is smaller (approximately 30 microns in diameter on average) this time. Then, Path 2 is applied to this workpiece. The resultant grain size from “left” to “right” of the part along the center line is then graphed in Figure 11.

Figure 9: Examples of forged parts made just in the last few months since the first prototype Agility Forge was built. Figure 10 (top) demonstrates how grain size as a function of recrystallization can be modeled via a Johnson-Mehl-Avrami Kolmogorov (JMAK) curve, which predicts grain size as a function of strain, strain rate and temperature—during and after processing. Depicted in Figure 10 is metadynamic recrystallization, or MRX. Once average grain size is known, a prediction about flow strength as a function of grain size can be made (Figure 10, bottom).

Figure 11: Three different process paths to form the same geometry via incremental autonomous forging. Figure 12 shows the predicted grain size from “left” to “right” through the center line of the part. For Paths 1 and 2, we see a slight reduction in average grain size for Path 2. For Path 3, though, for which the workpiece experienced additional deformation, we see a much more uniform and smaller average grain size. From the simple Hall-Petch relationship, we see that this path would result in a predicted increase in flow strength from ~300 MPa to ~400 MPa - all achieved with simply a few extra minutes of autonomous incremental forming prior to this finish forging path (Figure 10). Figure 13 superimposes these average grain sizes as a function of the selected process path.

Figure 10: Example microstructure evolution (metadynamic recrystallization post deformation) top; and effect of grain size on yield strength (bottom). By modeling microstructure evolution during manufacturing, grain size before (i.e., preheating), during (i.e., deformation) and after (i.e., cool-down and heat treatment) can be predicted as a function of thermomechanical processing. Since multiple different TMP paths may produce the same part geometry but different final grain size distributions throughout the part, it is possible to model which TMP path will produce the grain size that will result in the best flow strength. Thus, we can include a new design parameter in our forging process: properties as a function of microstructure. Processing leads to microstructure, microstructure leads to properties and properties lead to performance.

FIA MAGAZINE | AUGUST 2025 72