Increasing productivity for a cost-competitive US workforce
Much of today’s manufacturing processes rely heavily on human labor, especially in the final assembly process for many goods. This reliance is a driving factor in the offshoring of production, where companies can often hire and pay foreign-based workers less than US-based employees.
A key approach to revitalize domestic manufacturing, bring domestic jobs back, realize clean energy goals, and increase productivity is automation that helps improve worker conditions and productivity while also providing improved process and part quality.
The design of advanced controllers for manufacturing requires the confluence of expertise in automation and controls, toolpathing, data-science and machine learning, computing, and material science that is a strength of MDF.
Control Architecture
Researchers at MDF have developed algorithms that ensure rapid, collision-free builds of components across a variety of 3D-printing technologies. During production, in-situ data is collected across a variety of sensing modalities that is used to create a digital twin that can be used in process planning and post-production data exploration.
This architecture is being deployed in multiple systems at MDF including large-scale metal additive manufacturing and omnidirectional composites printing for rapid manufacture of large geometries.
Scientific Challenges
Scientific Challenges
Despite significant, ongoing advancements in computing power and sensing technologies, automating anything beyond highly repetitive, constrained tasks is still very difficult, and even relatively minor unexpected changes in the environment are difficult to overcome. Integrating and programming automation from one job to the next is also time consuming and costly, pricing many small- and medium-sized enterprises out of consideration.
Recent advances in machine learning and artificial intelligence have begun to eliminate the environmental awareness problem. Once this gap is closed, the remaining challenge is using that enhanced awareness to inform control decisions through data-driven control design and machine-learning-based algorithms. Using this approach, teams at MDF are developing truly flexible control systems that can learn to adapt to new and novel environments.
Highlights
Hip for Hydro
A team at MDF used MedUSA to print a 900lb, stainless steel “PM-HIP can” for a hydro impeller in just 46 hours. The project demonstrates an alternative pathway to creating an additively manufactured, large-scale, near-net-shape component.
100lb Per-Hour Build
A team at MDF demonstrated the ability to print 100lb per hour on MedUSA, a system that uses three deposition heads working in harmony to create complex metal shapes. Increasing the productivity of large-scale metal printing is critical to lowering the cost of creating large metal components for things like power generation, pipes, valves, and vessels.
Get in touch
Joshua Vaughan
Group Leader
Manufacturing Robotics & Controls