Fine-Tuning the Factory: Simulation App Helps Optimize an Additive Manufacturing Facility

To fulfill its promise of rapid, precise, and customizable production, additive manufacturing (AM) demands more than just a retooling of factory equipment; it also calls for new approaches to factory operation and management. This is why Britain’s Manufacturing Technology Centre (MTC) has enhanced its in-house metal powder bed fusion AM facility with a simulation model and app to help factory staff make informed decisions about its operation.

"The model helps predict how heat and humidity inside a powder bed fusion factory may affect product quality and worker safety," says Adam Holloway, a technology manager within the MTC's modeling team. The simulation app, meanwhile, shows the potential for pairing a full-scale AM factory with a so-called "digital twin" of itself. "When combined with data feeds from our facility, the app helps us integrate predictive modeling into day-to-day decision-making."

How Heat and Humidity Affect Metal Powder Bed Fusion

Maintaining careful control of heat and humidity is essential at the MTC factory. "The metal powder used for the powder bed fusion process is highly sensitive to external conditions," says Holloway. "It can begin to oxidize and pick up ambient moisture even while it sits in storage. Exposure to heat and moisture will change how it flows, how it melts, how it picks up an electric charge, and how it solidifies," he says. "All of these factors can affect the resulting quality of the parts you’re producing."

Careless handling of powdered metal can threaten the health and safety of workers as well. "The metal powder used for AM processes is flammable and toxic, and as it dries out, it becomes even more flammable," Holloway says. "We need to continuously measure and manage humidity levels as well as how loose powder propagates throughout the facility."

To maintain proper atmospheric conditions, a manufacturer could augment its factory's ventilation with a full climate control system, but that could be prohibitively expensive. Using multiphysics simulation for careful process management could provide a cost-effective alternative.

Responsive Process Management with Multiphysics Modeling

"We created a model of our facility using the computational fluid dynamics (CFD) capabilities of the COMSOL® software. Our model (Figure 1) uses the finite element method to solve partial differential equations describing heat transfer and fluid flow across the air domain in our facility," says Holloway. "This enabled us to study how environmental conditions would be affected by multiple variables, from the weather outside to the way machines were positioned inside the shop. A model that accounts for those variables helps factory staff adjust ventilation and production schedules to optimize conditions," he explains.

A Simulation App that Empowers Factory Staff

The team made its model more accessible by building a simulation app of it with the Application Builder in COMSOL Multiphysics® (Figure 2). "We’re trying to present the findings of some very complex calculations in a simple-to-understand way," Holloway explains. "By creating an app from our model, we can empower staff to run predictive simulations on laptops during their daily shifts."

Figure 2. A simulation app of the powder bed fusion facility, showing the machines it contains and the locations of the air vents. In this case, some doors (highlighted in pink) have been left open. Image courtesy of COMSOL.

The app user can define relevant boundary conditions for the beginning of a factory shift and then make ongoing adjustments. Over the course of a shift, heat and humidity levels will inevitably fluctuate. Perhaps factory staff should alter the production schedule to maintain part quality, or maybe they just need to open doors and windows to improve ventilation. Users can change settings in the app to test the possible effects of actions like these. For example, Figure 3 presents isothermal surface plots that show the effect that opening the AM machines’ build chambers has on air temperature.

Figure 3. Isothermal surface plots show changes in temperature at 30 seconds (left) and 60 seconds (right) after opening the build chambers of every AM machine in the facility.

A Step Toward a "Factory-Level Digital Twin"

The app currently requires workers to manually input relevant data. Looking ahead, the team envisions something more integral and, therefore, more powerful: a "digital twin" for its AM facility. "To make our factory environment model a digital twin, we'd first provide it with ongoing live data from the actual factory," Holloway explains. "Once our factory model was running in the background, it could adjust its forecasts in response to its data feeds and suggest specific actions based on those forecasts."

"We want to integrate our predictive model into a feedback loop that includes the actual factory and its staff. The goal is to have a holistic system that responds to current factory conditions, uses simulation to make predictions about future conditions, and seamlessly makes self-optimizing adjustments based on those predictions," Holloway says.

Simulation at Work on the Factory Floor

The simulation app has already proven its worth. "Our manufacturing partners may already see how modeling can help with planning an AM facility, but not really understand how it can help with operation," Holloway says. "We're showing the value of enabling a line worker to open up the app, enter in a few readings or import sensor data, and then quickly get a meaningful forecast of how a batch of powder will behave that day."

Beyond its practical insights for manufacturers, the overall project may offer a broader lesson as well: By pairing its production line with a dynamic simulation model, the app project has made the entire operation safer, more productive, and more efficient. The team has achieved this by deploying the model where it can do the most good — into the hands of the people working on the factory floor.

This article has been abridged. The full story of MTC's use of COMSOL Multiphysics can be found here.

COMSOL and COMSOL Multiphysics are registered trademarks of COMSOL AB.