Authors:
Vincent Nebel
1
;
Pia Goßrau-Lenau
2
;
Harshvardhan Agarwal
1
and
Dirk Werth
1
Affiliations:
1
August-Wilhelm Scheer Institute for Digital Products and Processes gGmbH, Saarbrücken, Germany
;
2
Technology Center Drives, Miele & Cie. KG, Euskirchen, Germany
Keyword(s):
Anomaly Detection, Digital Twin, Prototyping, Machine Learning.
Abstract:
Developing mechanical components, especially complex assemblies like pumps, is a resource and time intensive process. Testing pump prototypes for long-term durability is critical to ensure error-free operation of the final product. Prototypes undergo material and operational tests to determine their expected lifespan, focusing on defects caused by material degradation and water contamination. Long-term tests, lasting months, are necessary to simulate real-world conditions, but limited test bench capacities create bottlenecks, restricting material experimentation. Moreover, monitoring the internal state of pumps during tests is challenging. Undetected defects can worsen or trigger secondary issues, complicating the root cause analysis, which provides valuable information for further product improvements. To address these challenges, a digital twin that integrates geometry and material data, simulations, and sensor measurements was developed. This twin is used as data source for machin
e learning based anomaly detection, allowing tests to stop sooner and preventing further damage when first signs of a defect are detected. A modular serverless architecture is used to host the model inference on the cloud, improving resource usage and scalability as well as reducing operational costs.
(More)