A Microservice-Based Architecture for Flexible Data Acquisition at the Edge in the Context of Hairpin Stator Production

Abstract

Data-driven technologies such as machine learning promise great potentials for electric drives production. However, present information systems do not allow flexible and reliable accumulation of high-dimensional data such as images and time series, so feature-rich data are often discarded in practice. The advent of edge computing and microservice-based software architectures in recent years enables low latencies, high data integrity, and flexibility towards different communication protocols, thus offering new possibilities to accumulate data in a manufacturing environment. Hence, this paper proposes a microservice architecture deployed on the edge that serves as a communication layer between data sources and downstream data analytics capabilities. The architecture's flexibility is demonstrated and validated on different process steps along the process chain of hairpin stator production. In conclusion, the developed microservice architecture is suitable for reliable data acquisition at the edge and can enable subsequent machine learning analyses across the process chain.