Currently, there is a gap between how data is collected on the shop floor based on resources such as machines, robots, and Autonomous Guided Vehicles (AGVs) and the man-ufacturing orchestration software that sits above these resources and controls their interaction from the point of the creation of single products. Shop-floor resources create data streams that are saved in databases, cleaned, and then re-contextualized, i.e., to connect the data to orders, batches, and single products. New analysis prospects arise when integrating this data and analysis methods with the process-oriented analysis perspective. This paper exploits these prospects based on a real-world case: BPMN models are created for the manufacturing of two real-world products: (1) a low volume, high complexity lower-housing for a gas turbine and (2) a high volume, low complexity, small tolerance valve lifter for a gas turbine. Instead of collecting the data independently based on the participating machines, the data collection of 30+ values is modeled into the BPMN models and enacted by a workflow engine, resulting in execution logs. Conformance checks are conducted and interpreted for the scenarios and it is shown how existing classification and clustering techniques can be used to predict good and bad parts, ex-post and potentially at run-time.

Top