Digital Technologies Drive Industrialization of Large Scale DED-AM

The potential to build very large components with a high deposition rate is what makes metal direct energy deposition (DED) additive manufacturing (AM) more appealing than many other AM disciplines. The DED AM process deposits molten metal, layer upon layer, to build up a part. Most systems use a laser, plasma, electron-beam or an arc to generate heat and either a wire or blown powder to add material.

DED process simulation results in Simufact Welding. (All images provided by Hexagon)

Theoretically, DED AM parts are nearly unlimited in size and can take many intricate forms based on the energy source and method of feedstock delivery.

Nonetheless, the very nature of the DED AM process introduces challenges in the part’s form or thermal stress, strain and distortion. These factors can make it difficult to build parts in a single printed iteration that meet engineering drawing requirements. Until now, these issues have been mitigated through experience and old-fashioned trial-and-error. When a single build costs tens of thousands of dollars and takes days, weeks or even months to complete, a fresh approach is justified to minimize the number of builds required to achieve an acceptable outcome.

Simulation Enables the Virtual Tryout

Process simulation software for welding and additive manufacturing applications has become increasingly popular in the past few years. Engineers use the software to evaluate thermal history and resulting stress, strain and distortion of the structure during the build process. This enables the all-important virtual tryout.

Before conducting any physical DED builds, the user can predict the outcome of the planned AM process and diminish undesirable outcomes prior to attempting any physical builds. However, simulation is not without its own set of challenges, such as the long solve times required to complete the computer simulation calculation. In some cases, this can take much longer than the physical process itself, making the simulation less valuable.

In an effort to dramatically reduce solve times, some simplification of the modeling or calculation approach may be applied. The trade-off for large speed improvements may be a loss in the accuracy of the results. Feedback from DED industry users reveals that a directionally correct simulation result with even 50-percent accuracy is still quite valuable if the simulation time is reasonable.

Another challenge is the availability of material property data, which is a key input for DED process simulation. The accuracy of thermal and mechanical material properties plays a central role in determining the robustness of the end result. Robustly defined material property data is often not available, and assumptions must be made resulting in less accurate simulation results.

Thermal History Prediction is Where It Starts

Process simulation can also effectively predict local thermal history, informing the engineer in two additional ways: 1) predicting time at temperature locally, cooling rates and resulting microstructures, and 2) predicting local thermal history for use in process-control mechanisms. Microstructure and grain prediction enables engineers to evaluate the structure’s fitness for purpose in its intended application. Some micro- and grain-structures are more or less desirable than others, and have different material properties such as yield strength, ductility and elongation, which may meet the specified design / material allowable.

It is possible to control the DED process in a way that achieves a more desirable microstructure along with resulting material properties. Simulation makes this effort feasible. Local thermal history prediction enables engineers to establish what the local surface temperature over time should look like. This information then feeds a process-control system to ensure that the build progresses as planned. Thermal history data also enables engineers to look at regions that may become problematic during the build process. The same data facilitates the determination of ideal pause times (cooling time) between layers or regions of the part.

Realizing the Possibilities of DED

Technology partnerships are actively developing strategies and solutions to overcome many of these challenges. DM3D of Auburn Hills, Mich. is currently working with Hexagon’s Manufacturing Intelligence division of North Kingstown, RI to enable a new era of large-scale DED component production. DM3D is a supplier of laser blown powder DED machines and services with the capability of building some of the largest parts of any AM machine on the market today. Like all metal DED AM machine providers, DM3D has a thorough understanding of the challenges arising from the thermal effects of their process. Hexagon has developed a full suite of simulation software, including a module specific to DED.

“The ability to predict and manage the thermal history, residual stress, and resulting distortion is key to obtaining a usable part. Through the application of Hexagon’s Simufact Welding simulation software, DM3D is able to try out the planned process before kicking off the physical build and make adjustments as necessary, ”said Bhaskar Dutta, President of DM3D. “Primarily, we are looking at the distortion of the structure: location, magnitude, and time when it happens. We also evaluate the thermal history, residual stress profile and loss of Z-height.

“By reducing the amount of extra machining stock added to the part and the number of iterations required, DM3D is able to achieve better parts at a reduced time and cost. We have found Hexagon’s Simufact DED module to be one of the most advanced software products in the marketplace. Simufact has developed a process-oriented graphical user interface (GUI) and a fast solver that significantly reduces modeling effort and computational time, while yielding reasonable results, ”concluded Dutta.

Progressing to the Physical Build with Confidence

A physical build can proceed once a suitable process is determined for the 3D CAD model-to-be-printed, and the try-out is completed via process simulation. The simulation software generates a compensated G-code or geometric representation that, when built, will deform into shape. This build plan is sent to the DED machine and the build commences. At this point, the focus shifts to ensuring the resulting build is progressing as intended and meets the required dimensional specifications. During the build, thermal imaging data is collected in and around the melt pool as well as at other places on the part. This data is compared to simulation data to validate that the thermal history is progressing as expected; this is the single most significant indicator of a successful outcome. Engineers can also inspect the build geometry, typically through optical or tactile measurement systems.

NASA-Nozzle-Scan.jpg
Structured light scan of DED part built by DM3D.

The measurement data and its comparison to the intended part geometry are used to evaluate the printed part for dimensional compliance. However, it is important to note the as-built geometry in DED AM is generally a “near-net shape.” It is not relevant to compare this data to the CAD-nominal geometry, as there is usually a significant amount of excess material added to the build that must be removed during final machining. This additional process is needed as DED does not produce an acceptable surface finish for most applications. To address this issue, engineers must also have an accurate geometric representation of the near-net shape to compare with the measurement data. This step enables engineers to identify nonconformance and remove it from the process as scrap as early as possible, before performing additional value-add operations.

Achieving Tighter Tolerances

Even with effective process simulation, the first build may not be entirely within its dimensional specification due to the highly non-linear, thermal-mechanical behavior inherent in the DED process. In this case, a secondary workflow is required.

Comp-from-scan-workflow.jpg
Enhanced workflow: combined compensation from simulation and scan.

After the first physical build, the part is scanned with a measurement device and a digital representation of the as-built geometry is generated. The advanced metrology software Volume Graphics is used to analyze this data to obtain deformation fields within the part. These deformation fields are used to perform a compensation from scan, morphing the geometry already compensated by the process simulation. In doing so, the engineer can be confident that the tryout process will most likely be limited to no more than one physical build to achieve dimensional control of the DED AM build. However, in the case where tighter tolerances are required, this process can be performed iteratively in an effort to get closer and closer to nominal.

DED AM is a challenging manufacturing process with very high potential. The path to success requires modern solutions to achieve quality control targets, lower production costs, and reduce time-to-first-article. Process simulation, in-process monitoring, and compensation from scan present a powerful combination of engineering tools that can propel metal direct energy deposition to its full potential for large-scale additive manufactured parts.

Leave a Comment