Implications from Legacy Device Environments on the Conceptional Design of Machine Learning Models in Manufacturing

Abstract

While new production areas (greenfields) have state-of-the-art technologies for implementing digitalization, existing production areas (brownfields) and devices must first be upgraded with technologies before digitalization can be implemented. The aim of this research work is to use a case study to identify the differences in the implementation of machine learning (ML) projects in brownfields and greenfields. For this purpose, an ML application for the detection of changeover times on milling machines is implemented and analyzed in the brownfield and greenfield scenarios as well as a combined scenario. Particular attention is paid to the selection of sensors and features. It was found that the abundant availability of features in the greenfield scenario poses pitfalls when creating ML projects if the underlying sensors cannot be checked for their suitability. For the changeover detector use case, the best model quality was achieved for the combined scenario, followed by the greenfield scenario.

A Series Production Data Set for Five-Axis CNC Milling

“…In [2], the data set presented here was used to train a Machine Learning (ML), model for detecting changeover periods in production data. These models were then compared to ML models, which were trained with a data set from a DMG 100 U duoBLOCK milling machine.…”

“…20 "Warmup" was derived after the data acquisition from variable No. 3 "ProgramDetail" and added to the data set [2]. Due to the selected data recording concept, the data were either transferred to the database as soon as a new value was assigned to the variables, or every two seconds.…”

“…Table 6 contains a short description of all 23 phases. For more details, please see[2].Basic statistics of the features and labels are listed in Table A1 in Appendix A.…”