IDS 2017 Industrial Data Science Conference Application of Data Mining for Prospective Assembly Time Determination Dortmund / 05.09.2017, Dr. Olga Erohin and Ralf Kretschmer
Agenda Miele Group and Business Unit Professional Research Project „Pro Mondi“ and time data management Knowledge discovery for prospective assembly time prediction Assembly time prediction in the product development phase Conclusion 2
Scope and goals of the research project Pro Mondi Assembly oriented data model (product and process view) Research project (2012-2015): Cross-domain knowledge „Prospective determination of assembly work content in digital factory (Pro Mondi)“ Identification and Mapping of product representation Detection and description of process-relevant of product-characteristic and process structure process pattern product properties Manufacturing Functional BOM Engineering BOM Manufacturing BOM Manufacturing BOM Product Product planning Process planning Production development Product emergence BOM=Bill of Materials 3 Source: Research project “Pro Mondi”, http://www.pro -mondi.de/
Assembly time is the basis for various processes Work system Human resource Scheduling design planning Cost Benchmarking calculation Time data Production Performance control assessment Value stream Capacity Invest design planning planning 15-70% of production time is assembly time Time-related data are applied in manifold areas Manual assembly is a wide-spread assembly method Time data management (e.g. time studies) is an for multi-variant products essential task field of Industrial Engineering Sources: B. Lotter: Einführung. In: B. Lotter, H.-P. Wiendahl (Hrsg.): Montage in der industriellen Produktion – Ein Handbuch für die Praxis. Springer Verlag 2012. 4 J. Deuse, F. Busch: Zeitwirtschaft in der Montage. In: B. Lotter, H.-P. Wiendahl (Hrsg.): Montage in der industriellen Produktion – Ein Handbuch für die Praxis. Springer Verlag 2012 Miele & Cie. KG.
Two results of the research project Pro Mondi ATP ProWiZei Assembly Knowledge time discovery for MTM-ProKon MTM-Analysis prediction in prospective Product development the product assembly Production planning development time Production phase prediction Time and motion study Self-recording Inquiry Registering by devices Work sampling study Simulation (calculation) Predetermined motion time systems Comparative estimating Time agreement Standard data building blocks 5 Sources: O. Erohin, J. Schallow, J. Deuse, R. Klinkenberg: Application of data mining to predict assembly time in early phases of product emergence. CIE43 Proceedings, 2013. 5 Miele & Cie. KG
KDID in context of time data management ProWiZei Realization 3.2 Create an IT prototype 3.1 Integrate KDID into planning processes and Data Mining 2.3 RapidMiner Preprocessing of raw IT systems data matrix of TDM-data M2 2.4 2.2 Descriptive and Create and apply data explorative analysis of mining models data 2.5 2.1 Build a raw data Visualize and interpret matrix of TDM-data the results M1 Preparation 1.3 Explore relevant IT systems Morphology Determination Pre-Processing Application and TDM-data connections Data management 1.2 Explore the processes of time data management 1.1 Define the goals of KDID: Knowledge Discovery in Industrial Databases knowledge discovery in time Milestones Iterative steps TDM: Time Data Management data management 6 Source: O. Erohin: Wissensgewinnung durch Datenanalyse zur prospektiven Zeitermittlung. Shaker Verlag 2017.
Realization of KDID and some prediction results Relative Data Mining method RapidMiner operator error Linear regression Linear Regression 18,2% Local polynomial Local Polynomial 33,1% regression Regression Regression model tree W-M5P 33,4% Regression tree W-RepTree 29,5% Support vector LibSVM 17,1% regression Support Vector Machine 7 Source: O. Erohin: Wissensgewinnung durch Datenanalyse zur prospektiven Zeitermittlung. Shaker Verlag 2017.
Continuity of time data determination along the product emergence process (PEP) industrial engineer, assembly planner design design phase development phase freeze development process planning designer PEP iteration loop iteration loop iteration loop time prediction time approximate time target time type prospective assembly production-oriented analysis of process time prediction design (assessment of predetermined motion assembly fairness) time systems step software Data class process data product data 8 Source: Personal research work by R. Kretschmer
Mapping of product and process data product data process data (CAD, PDM, …) (industrial engineering, assembly planning, …) product basic assembly time for component A B-building-block(s) V-building-block(s) product-oriented time-data structure component (joining processes) (connection processes) group 1 hierarchical product structure + component A + component B + component C component … group 2 component D component E assignment via ID product data process data … 9 Source: Personal research work by R. Kretschmer
Concept for assembly time prediction data based on the past current data (model use) (model creation based on instances group) components 1-n new component product data identification of k- nearest-neighbours Connection process type(s) + number mapping via ID n n + process data n n n n prediction prediction B-building block(s) V-building block(s) B-building blocks V-building blocks (joining processes) (connection processes) (joining processes) (connection processes) predicted assembly time for a new assembly times 1-n component process steps product data process data 10 Source: Personal research work by R. Kretschmer
Evaluation: Validation and assessment of results Assembly-time prediction compared to system of predetermined time (MTM-TiCon) and other established time-determination systems approximate time according to front panel D ProKondigital assembly time prediction (ATP) for a front panel C changed computer-aided design (k=4) assembly time prediction (ATP) for a new front panel B computer-aided design (k=4) actual time recording according to REFA front panel A 0,0% 5,0% 10,0% 15,0% 20,0% 25,0% 30,0% Deviation to system of predetermined time (relative error) Conclusion: Best results for further development within the component family Widely varying quality of results for complete new design and/or missing component family (=> small data amount for comparable components related to the past) 11 Source: Personal research work by R. Kretschmer
Conclusion Integration of data mining for prospective determination of assembly time leads to essential added value for planning and decision-making … and supports the idea of simultaneous engineering to reduce the product emergence time. Current portfolio of methods for time determination can be successfully extended by new data mining methods. Fundamental factors of success are Integration of specific know-how of the application area (especially at the beginning of knowledge discovery). Overcoming the challenges of “historically evolved” IT infrastructures. 12
Thank you very much for your kind attention! For further information please visit: www.pro-mondi.de Ralf Kretschmer Dr. Olga Erohin Director Segment Professional Director Corporate Development Laundry Technology Lehrte Professional Technology Miele & Cie. KG Miele & Cie. KG Industriestraße 3, 31275 Lehrte Mielestraße 2, 33611 Bielefeld ralf.kretschmer@miele.com olga.erohin@miele.com http://www.miele.com http://www.miele.com 13
Recommend
More recommend
