MAT Foundry Group is adopting DISA’s Artificial-Intelligence-powered Monitizer | Prescribe Artificial Intelligence platform to reduce casting scrap at its seven foundries. The iron foundry at MAT Eurac, in Poole, England, will be the test site for the wider deployment across the organization.
MAT has seven foundries pouring more than 500,000 metric tons/year of iron for automotive parts. The Poole foundry casts more than 40,000 metric tons/year, and adopted the Monitizer | Discover program for data collection, data visualization, and real-time monitoring/alerting in 2016
“Our main goals are to reduce scrap and make our process stable,” commented MAT’s Shaun Lindfield, commercial director. “We already have collected and centralized huge amounts of data with Monitizer | Discover that we exploit daily. If the project is a success – and I am sure it will be – we intend to implement Prescribe in other foundries across our global group.”
Eurac Poole will deploy Monitizer | Prescribe for both of its two DISA vertical molding lines. The program uses AI and cloud-based data analysis to optimize the entire molding process. By analyzing live data, it can recommend changes to process parameters that prevent defects occurring.
“Our long partnership with DISA and other foundries’ successful Prescribe deployments give us confidence that this technology is well proven and the correct next step in our digital journey,” Lindfield said. “In fact, this project is very low risk for us. Monitizer | Prescribe uses the same platform as our existing Monitizer | Discover tool and has sensible upfront pricing with ongoing subscription-based payment.”
Driving down scrap will improve Eurac Poole’s profitability and reduce energy consumption.
Prescribe’s automated suggestions also will help fill a potential knowledge gap at the foundry. “We lose a huge amount of expertise when older staff retire,” according to Lindfield. “Upskilling takes training and time. Prescribe will analyze our historical data to learn which inputs work best, and will guide us in adjusting our process to reach the best possible quality. Not letting operators work by trial and error will reduce process variability too.”
The first test runs scheduled for mid-2022.