The Gunma Industrial Technology Center Implements 24/7 Anomaly Detection System for Product Processing
GITC Reduces Anomalies While Boosting Productivity
“App Designer enables the creation of original dashboards that reflect user requirements in detail. The design of the dashboard can be created without coding, making it accessible even for beginners in programming while offering a low-cost solution.”
Key Outcomes
- MATLAB tools enabled 24/7 anomaly detection, reducing imperfections from 3% to 0.5%
- App Designer made it possible for engineers without prior coding experience to implement the system
- MATLAB tools offered a low-cost solution for reducing defects in product processing
Sensor data is transmitted to ThingSpeak and anomaly notifications are sent through Slack.
Based in Gunma Prefecture, Japan, the Gunma Industrial Technology Center (GITC) is a public research institution that promotes industrial development and advances the regional economy of Gunma. GITC researchers work on technologies used in product processing. One of the main challenges in this field is accounting for unexpected anomalies that can lead to defects in product quality and processing productivity. To improve how these imperfections are monitored, GITC has created a new system that uses MATLAB® tools to process and analyze data for real-time anomaly detection.
GITC identified three areas to target for anomaly monitoring: cutting oil contamination, processing environment temperature, and drill breakage. The sequence for detecting drill anomalies involves accumulating data, performing learning and judgment on a PC, and then sending the results to ThingSpeak™. The team also used Statistics and Machine Learning Toolbox™ to look for patterns in the data that could better predict when anomalies might occur. Finally, they used App Designer to build original dashboards to visualize the data without the need for prior coding experience.
These tools enabled GITC to monitor anomalies in real time and send instant messages to team members, even when off-site. This means that product processing could be monitored 24/7. As a result, GITC was able to improve productivity, save costs, and reduce the number of anomalies from 3% to just 0.5%.
