One of a handful of big tech companies that can trace its roots back as far as the 1970s, Seagate Technology, Cupertino, California, has been perched on the cutting edge of data storage and management solutions for decades. Now a $10 billion mass-data storage infrastructure solutions company, Seagate is employing smart manufacturing strategies, including artificial intelligence and machine learning on the plant floor.
In 2017, Seagate implemented AI manufacturing software for the microscopic vision inspection of wafers. Prior, the company had used rules-based machine vision systems to automate the anomaly detection process. They achieved a high accuracy rate, but the company also wrestled with limitations.
The earlier approach required strict parameters for each type of defect, all statically coded. The fixed ranges helped determine the criteria for clearing or stopping a product. However, the evolution of defect appearance or the advent of new types of abnormalities required additional rules that could become increasingly complex and hard to manage as a whole.
By implementing comprehensive digital manufacturing data operations and AI upgrades, the power and scale of image detection have seen dramatic improvements across Seagate’s wafer production facilities in the US and Northern Ireland. Accuracy has gone from 50 percent to more than 90 percent today.
To get there, the company has been extracting value from terabytes of sensor data produced by the high-precision tools the company uses. That data has been normalized and made easier to use by AI systems. As a result, Seagate now has multiple automated fault-detection solutions to help make wafer and tool decisions and a portfolio of AI-augmented detectors to autonomously monitor critical junctures in the manufacturing process with better-orchestrated rules.
Seagate captures the relevant runtime metadata and puts the raw data into context to create useful information in real time—closing the loop between the digital and physical worlds and positively impacting the way products are engineered, manufactured, and serviced.
The digital thread preserves data collected during production runs to know when to enact fast updates to enterprise resource planning and other decision-support systems and generate lessons to improve future automated decision making.
These efficiencies have extended the life of Seagate’s equipment by predicting which assets were on track to fall out of calibration and therefore in need of maintenance. The company realized significant savings on inspection labor, scrap prevention, labor reallocation, and avoided capital outlays for new equipment.
Seagate’s consistent focus on this kind of innovation has improved operating performance, accountability, and increased the effectiveness of enterprise systems through fast and accurate visibility into the entire manufacturing process.
The system of deep-learning algorithms also generates supporting evidence for other factory control systems, meaning Seagate can harness more data for virtual metrology and process control.
High-tech manufacturers like Seagate need to embrace the rapidly evolving opportunities that digital manufacturing, AI, and ML represent. Seagate has not only invested in cutting-edge technology but has also demonstrated how continual transformation can make an organization capable of defining the future of its industry.
Connect With Us
