As more and more companies are producing highly customized products, the use of sensors in the workplace increases dramatically. They are used in operator information systems, smart tools and so on. Data from these sensors (images, handling time, fastening torque) could be used to detect and resolve anomalies in the assembly system. However, despite the existing technological developments and the clear need for such a product anomaly detection system, this is rarely the case.
In this project, we want to
- Build a methodology on how to set up anomaly detection and monitoring for manual and (semi) automatic assembly.
- Develop the supporting software tools to realize them.
Eventually, this will result in a monitoring system that detects assembly mistakes and geometric product anomalies in the assembly environment, based on models derived from recorded sensor data. Next, a data-driven model and digital twin of the assembly process will be used to perform a root cause analysis on detected anomalies. The combination of data analytics and the digital twin model, as it is a (near) real-time digital representation of the assembly environment, will provide the operator with the key information to determine the necessary countermeasures to resolve the anomalies.
DigitizedAssembly is an ICON-project (Interdisciplinair Coöperatief Onderzoek – Interdisciplinary Cooperative Research). We are looking for companies to detail the project content with us and define use cases for valorisation of the project results.