TU Darmstadt / ULB / TUbiblio

Data mining approach for production order identification in load profiles of machine tools: a change-point and clustering based analysis

Wächter, Andreas ; Ioshchikhes, Borys ; Kolb, Niklas ; Weigold, Matthias (2023)
Data mining approach for production order identification in load profiles of machine tools: a change-point and clustering based analysis.
In: Procedia CIRP, 120
doi: 10.1016/j.procir.2023.09.104
Artikel, Bibliographie

Kurzbeschreibung (Abstract)

With growing awareness of sustainable manufacturing and rising energy costs, the monitoring of energy-related quantities in production is expanded. Measurements of the power consumption of production systems are used to identify energy efficiency potentials or to allocate energy use and associated costs. However, time series of the power consumption of machine tools contain more information. Load profiles of production systems enable drawing conclusions on the specific production order carried out. Time stamps for certain machining operations can be extracted and used to link the machine tool's power consumption to a specific production order. The linkage of this data can assist the order-specific allocation of energy use or associated emissions. This paper presents a data mining approach to gain information on the executed production orders from the load profiles of a machine tool. For this purpose, load profiles of a machine tool are pre-processed and segmented based on change points. Subsequently, the resulting sequences are clustered with respect to their similarity. These clusters contain segments of machining cycles in which the same workpieces were produced. By matching the cluster sizes with the known quantity of the executed orders, associated time series segments are selected and linked to the manufacturing order. The approach is implemented in a Python framework and carried out with data from milling and turning machining centers monitored in an industrial environment. Time stamps for machining cycles can be identified reliably. No training data is required for the recognition of recurring patterns at the load profile of a machine tool, but production orders need to be distinguishable at least by workpiece quantity, which was given with all tested real data.

Typ des Eintrags: Artikel
Erschienen: 2023
Autor(en): Wächter, Andreas ; Ioshchikhes, Borys ; Kolb, Niklas ; Weigold, Matthias
Art des Eintrags: Bibliographie
Titel: Data mining approach for production order identification in load profiles of machine tools: a change-point and clustering based analysis
Sprache: Englisch
Publikationsjahr: 2023
Verlag: Elsevier B.V.
Titel der Zeitschrift, Zeitung oder Schriftenreihe: Procedia CIRP
Jahrgang/Volume einer Zeitschrift: 120
DOI: 10.1016/j.procir.2023.09.104
URL / URN: https://www.sciencedirect.com/science/article/pii/S221282712...
Kurzbeschreibung (Abstract):

With growing awareness of sustainable manufacturing and rising energy costs, the monitoring of energy-related quantities in production is expanded. Measurements of the power consumption of production systems are used to identify energy efficiency potentials or to allocate energy use and associated costs. However, time series of the power consumption of machine tools contain more information. Load profiles of production systems enable drawing conclusions on the specific production order carried out. Time stamps for certain machining operations can be extracted and used to link the machine tool's power consumption to a specific production order. The linkage of this data can assist the order-specific allocation of energy use or associated emissions. This paper presents a data mining approach to gain information on the executed production orders from the load profiles of a machine tool. For this purpose, load profiles of a machine tool are pre-processed and segmented based on change points. Subsequently, the resulting sequences are clustered with respect to their similarity. These clusters contain segments of machining cycles in which the same workpieces were produced. By matching the cluster sizes with the known quantity of the executed orders, associated time series segments are selected and linked to the manufacturing order. The approach is implemented in a Python framework and carried out with data from milling and turning machining centers monitored in an industrial environment. Time stamps for machining cycles can be identified reliably. No training data is required for the recognition of recurring patterns at the load profile of a machine tool, but production orders need to be distinguishable at least by workpiece quantity, which was given with all tested real data.

Freie Schlagworte: data mining, energy analysis, load curve, production system, traceability
Fachbereich(e)/-gebiet(e): 16 Fachbereich Maschinenbau
16 Fachbereich Maschinenbau > Institut für Produktionsmanagement und Werkzeugmaschinen (PTW)
16 Fachbereich Maschinenbau > Institut für Produktionsmanagement und Werkzeugmaschinen (PTW) > ETA Energietechnologien und Anwendungen in der Produktion
Hinterlegungsdatum: 15 Jan 2024 07:10
Letzte Änderung: 06 Aug 2024 07:03
PPN: 514759526
Export:
Suche nach Titel in: TUfind oder in Google
Frage zum Eintrag Frage zum Eintrag

Optionen (nur für Redakteure)
Redaktionelle Details anzeigen Redaktionelle Details anzeigen