May 18, 2021
EAST SYRACUSE, N.Y.—A new anomaly detection system from INFICON uses machine learning (ML) to empower process and equipment engineers with easy-to-use tools that are said to reduce product risk and rapidly resolve production issues. The SmartFDC Machine Learning Anomaly Detection System does so while automatically providing fault detection and classification (FDC) coverage, INFICON said in a release.
According to INFICON, the automated ML analysis results are easy to interpret, reducing the time needed to detect inline issues and diagnose excursions identified by the system. Besides reducing fault detection implementation time, SmartFDC is reported to maximize detection capabilities and provide powerful analytical tools to engineers of all skill levels.
The SmartFDC ML algorithm automatically learns the processes that run on the equipment with no additional setup or configuration, meaning that every tool and chamber is monitored from the moment the system is enabled. While the system fine tunes the learning, engineers are provided insight into the processes and equipment that enhance their ability to identify and troubleshoot issues, the company said.
INFICON (https://ims.inficon.com) provides manufacturing software and hardware for the semiconductor and related industries. The company offers instrumentation, critical sensor technologies, and smart manufacturing, Industry 4.0 software that is said to enhance the productivity and quality of tools, processes, and complete factories.
INFICON’s modular SmartFDC is also said to augment traditional FDC systems by merging machine learning and engineering experience into a single intuitive platform. Its architecture enables factories to augment existing third party FDC systems, including the INFICON FabGuard® FDC System, with ML capabilities that can target specific process areas or broadly cover the entire facility. This enables factories to easily ramp implementation and add capabilities based upon specific needs, without requiring significant changes to existing data collection or excursion detection infrastructure, the company said.