A new benchmark dataset for machine learning applications in discrete manufacturing: CiP-DMD

The development of machine learning applications in manufacturing depends primarily on the availability of meaningful and reliable. In this paper, we present the Center for industrial Productivity - Discrete Manufacturing Dataset (CiP-DMD), the first open-source discrete manufacturing dataset of a multi-step machining process for which detailed documentation is available. The dataset encompasses a comprehensive description of the individual machining processes, the quality controls and the traceability concept used to match the process and quality data corresponding to the production of 847 pneumatic cylinders. With this new benchmark dataset, we aim to contribute to the development of new approaches and foment the permeation of machine learning in manufacturing.