The Future Foundry is a Smart One: AI in the Metal Castings Industry
June 10, 2022
By Susan Bear, Vice President, Engineering, Program Management and Quality
Jurassic World: Dominion premieres this month, bringing a billion-dollar movie franchise to a close. Looking back to the first Jurassic Park film reveals more than just a summer blockbuster that stood the test of time.
An intriguing scene between Dr. Ian Malcolm (Jeff Goldblum) and Dr. Ellie Sattler (Laura Dern) where the pair discuss chaos theory has direct correlations to the foundry environment.
Photo credit: Jurassic Park movie clips: http://j.mp/1nXDPTF
In the scene, Ian asks Ellie to predict which direction a drop of water will fall after placing it on her hand. After two tests yield differing results, he says:
“It changed. Why? Because tiny variations — the orientation of the hairs on your hands, the amount of blood distended in your vessels, imperfections in the skin — that never repeat vastly affect the outcome. That’s unpredictability.”
Reducing Defects and Scrap Through Artificial Intelligence
Beyond an entertaining plot point, the scene perfectly describes what it’s like to produce metal castings in a foundry environment where miniscule variations in factors like temperature and humidity can alter the outcome of finished products.
Such variations result in scrap, which foundries like Grede are continually seeking to minimize. Analyzing data after part production is helpful but nowhere near as beneficial as using Internet of Things (IoT) technology to catch variations and account for them before production.
Live monitoring creates the opportunity to identify and eliminate defects after the first mold, significantly reducing scrap.
To borrow another movie analogy, foundry engineers are able to go from chasing ghosts to busting them.
Grede Integrates Cutting Edge IoT Technology in Two Ways
At Grede, we’re turning our plants into smart foundries with the integration of IoT technology across our footprint. By mid-2023, we expect two artificial intelligence-driven innovations to be driving efficiency across our plants.
First, Grede is using IoT and big data technology to improve quality control outcomes. We are implementing it into DISA systems, which control all the machines for molding across our corporation.
Powered by artificial intelligence, the system monitors everything done on our production lines. Data are being collected and leveraged on everything from melt to finishing, including:
- Mold count
- Pour temperature
Literally hundreds of elements comprising the foundry process are monitored, analyzed and then used to alert team members when systems begin to reach upper or lower control limits.
After as little as one month of data collection at high-volume plants, and several months at lower-volume plants focusing on fewer, more customized parts, the DISA system will have enough input to move from issuing alerts to proactively fixing issues before they happen, without ever needing to alert operators.
A Second Application of AI – Flow Analysis
An additional area where Grede is using artificial intelligence to enable innovation and efficiency is with flow analysis. IoT-enabled Magma software informs a number of critical decisions, including:
- Production feasibility
- Pour temperature
- Gating speed and feed direction
Proper analysis ensures the avoidance of issues that threaten production, such as:
- Slow pouring temperatures
- Accelerated metal cooling
- Turbulence resulting in metals flowing too fast or sand breaking up
Historically, this analysis has yielded insight that results in more efficient production and reduction of scrap — but it relies on human experience. With AI-powered Magma, the system is informed by the collective human experience. That information is then captured and applied for continual improvement.
How AI Helps Foundries Address Labor Challenges
The metal casting industry is not unique in the labor challenges it faces — from the high number of experienced workers retiring to turnover rates among new entrants.
As a generation of laborers with 40 years of experience leaves the industry, a rich knowledge base of how to identify defects and eliminate them is literally walking out the door.
When training can’t keep pace with turnover and retirements, AI-driven smart systems are the future for foundries.
Just like people, the system continues to learn. As it learns, it improves and reacts faster. And it enables scaling at levels not possible in the current and predicted labor environment.
Where Foundries Will be in Five Years
Applying IoT tech to systems as Grede is currently doing is the next step for the metal casting industry overall. Coupled with 3D-printing of sand cores, the entire manufacturing process can be improved.
Five years from now, automation will be running foundry production.
Instead of manual tweaking, machines will automatically adjust earlier in the process.
In a world with ever-increasing unpredictability, AI provides much-needed consistency.
While AI might not be able to keep an engineered T-Rex in a fictitious park, it can keep American foundries at the forefront of metal casting engineering, design and manufacturing — helping deliver desired outcomes to clients and their customers.