3

Enhancing Dimensional Accuracy and Stability with Digital Integration

Planning committee member Melissa Orme, The Boeing Company, provided an overview of additive manufacturing (AM) at that company. She began by presenting a detailed figure that showed the various pieces of the company’s AM value stream, from concept and design through build and testing and ending with production and service. To illustrate, she offered the example of manufacturing 1,740 aluminum parts for the Wideband Global SATCOM satellite, which were almost all unique and made in small lots. Using a digital twin that eliminated physical iterations and enabled preventive quality control, the company was able to save $4 million per spacecraft.

Orme noted that in AM, the material for a part is being made at the same time that the part is being made. “It is not just about making a shape,” she said.

She also pointed out that in the entire end-to-end value stream, the printing of a component with AM is only a small part. There are many other steps, including, for instance, heat treatment, rough machining, chemical milling, final machining, computed tomography (CT) scanning, inspection, static testing, and fatigue testing, and each of these processes must be stable, repeatable, and reliable. Information about each of those processes is gathered into the digital thread—that is, into the company’s data library, where the data are all standardized in terms of format, contextualized, and made accessible to the data scientists. The data are used in analytics, including artificial intelligence/machine learning (AI/ML), and fed back into the value stream for continuous improvements.

Scaling AM poses a number of challenges, Orme said. Perhaps the biggest is material and mechanical repeatability. When Boeing buys AM printers from a company, for instance, the printers are rarely identical.

Turning to qualification, Orme said that AM is inherently a defect-prone process. Thus, one must design, analyze, and test the parts with that in mind. It will never be a perfect process.

Orme spoke about how to get to the point of stable and repeatable performance in AM parts, which involves simulations and determining how process variables (gas flow, laser parameters, etc.) affect the ultimate characteristics of the part (density, yield strength, high cycle fatigue, etc.).

She closed with five areas of future trends and needs. First, she said, everything needs to be stable and repeatable. Next, design for AM must be integrated with traditional design tools, in turn integrated with simulation for microstructure and distortion prediction accompanying compensation methods. Third, faster, higher-fidelity simulation is needed relating the temperature gradient to the scan path and resulting microstructure. ML improvements for geometric fidelity and material characteristics also offer significant benefits in efficiency. Process control will be crucial, with scan path optimization and feedforward controls. Last, in the area of post-process integration, digital threads should be used with autonomous feedback loops for continuous improvements, future design enhancements, and increased quality.

In a brief question-and-answer session after her presentation, Orme was asked about the challenges of inspecting and certifying parts with complex internal structures. She noted that at this point, they were using AM for small-lot production runs, such as 20 heat exchangers, which enables certification paths based on testing and nondestructive imaging. Much of the inspection of these parts was done with CT or penetrant dye inspection or both, she said, and there was also vibration, acoustic, thermal, static, radio-frequency, and environmental testing performed based on the part, environment, and application. Concerning how they achieve dimensional stability and avoid part distortion, she said that they initially relied on support-structure optimization but have recently pivoted to active compensation based on simulation results, significantly reducing physical iterations, which is an important factor for sustainability and efficiency.