Model-driven System Architectures for Data Collection in Automated Production Systems

With the advent of the fourth industrial revolution, called Industrie 4.0, the domain of industrial automation transforms rapidly. Due to the ongoing digitization of processes, an ever-increasing amount of data is available from production. Leveraging
this data to adjust machine parameters and production plants is vital for efficient and flexible production. Yet, due to the heterogeneity of systems, the long life-cycles of production plants, as well as the multitude of involved disciplines, data
collection from automated production systems for data analyses requires significant implementation efforts.
This thesis addresses this challenge with an integrated approach for model-driven development of data collection architectures. First, data collection architecture principles and guidelines are derived from state-of-the-art and industrial needs.
Besides, a novel graphical domain-specific language is developed that allows multi-disciplinary experts to model relevant information technology and automation systems, as well as the flow of data, and hence the data collection process. Based on
these formalized models, a subsequent automated generation of the code for data collection architectures is proposed to minimize manual implementation efforts.
The proposed approach was implemented and evaluated on a lab- and semi-industrial-scale. Several case studies and industrial experts confirmed significantly reduced implementation efforts compared to manual implementation, as well as
good applicability of the graphical modeling language. Furthermore, the evaluation contains a unique extrapolation case-study that quantifies the generalized effort savings compared to manual implementation.