Search

Für produzierende Unternehmen sind Produktionsdaten das neue Gold. Insbesondere für die Produktionsplanung und -steuerung sind die Qualität und die Aktualität der erfassten Daten entscheidend. Der aktuelle Stand der Technik beim Schürfen des Goldes ist jedoch noch immer geprägt von manuellen Tätigkeiten. Wie effizient diese Praxis bei der Erfassung der Daten in den Unternehmen ist und welche Auswirkung diese auf die „Reinheit“ der Daten hat, zeigt dieser Beitrag.

Production systems have to meet quality requirements despite increasing product individuality, varying batch sizes and a scarcity of resources. The transfer of experience-based knowledge in a flexible and self-optimizing production and process planning offers the potential to meet these challenges. Biological systems solve conceptually similar challenges pertaining to the transfer of knowledge, flexibility of individual reactions and adaptation over time. Thus, in the context of digital transformation, mechanisms derived from biology are interpreted and applied to the knowledge domain of production technology. To be able to exploit the potential of bio-inspired production systems, genetic and intelligent properties of technical components and machines were identified and brought together under the concept of “Gentelligence”. Expanding upon this concept with the new idea of process-DNA and biologically inspired optimization algorithms facilitates a more flexible, learning and self-optimizing production, which is shown in three different applications. By using the new concept of gentelligent process planning it is possible to determine machine-specific process parameters in turning processes in order to ensure appropriate roughness within the requirements. Furthermore, the combination of the concept with a material removal simulation allows the determination of the resulting process force in tool grinding for subsequent unknown workpiece geometries. As a result of using the process-DNA, a workpiece-independent knowledge transfer and thus process adaptation for shape error compensation becomes possible. Gentelligent production scheduling enables a process-parallel, holistically optimized machine allocation, and as a result, a significantly reduced lead time.

Flexible Job Shop Scheduling is one of the most difficult optimization problems known. In addition, modern production planning and control strategies require continuous and process-parallel optimization of machine allocation and processing sequences. Therefore, this paper presents a new method for process parallel Flexible Job Shop Scheduling using the concept of quantum computing based optimization. A scientific benchmark and the application to a realistic use-case demonstrates the good performance and practicability of this new approach. A managerial insight shows how the approach for process parallel flexible job shop scheduling can be integrated in existing production planning and control IT-infrastructure.

Higher customer requirements and increasing competitive pressure lead to a growing number of variants and smaller batch sizes in manufacturing. As a result, companies have to deal with challenges of small series manufacturing. A unidimensional set-up optimization cannot provide a satisfying solution. At the same time, production system-specific control algorithms are usually complex and require constant adaptation as customer requirements change continuously. To address these issues, this paper presents a hybrid genetic algorithm which allows modular sequence optimization in production scheduling. Hybrid elements are used to reach a high solution quality within a short runtime.

Sales revenues of enterprises are often subject to seasonal fluctuation. This leads to high or low utilized resources and this in turn to revenue losses. Hence, the enterprises invest a high effort to improve long and short-term resource utilization. In this context, disregarding future capacity utilization within the process of quotation leads to short-timed capacity adjustments for instance, additional work hours across seasons. This paper presents an approach which focuses on dependencies between costs and capacity by linking cost pricing and production scheduling. A first evaluation at an MTO supplier shows that order delays can be reduced by up to 95% and total costs by 21% compared to using the most appropriate priority rule.

Unforeseen machine tool failures due to technical issues can cause downtimes leading to delays during production. To reduce delays, rescheduling of the production is, in most cases, necessary. However, warranting such a change requires reliable knowledge about the duration of the failure. This article presents a method to provide this knowledge by estimating the duration of a machine tool failure based on previous failure durations. Using the cross-industry standard process for data mining (CRISP-DM) and statistical methods, the embedded model for failure classification and duration is continuously improved. The method is thoroughly tested using multiple distributions, parameters and a practical use case. The results show high potential for predicting the duration of machine tool failures, which consequently could lead to improved quality of rescheduling.

Production systems have to meet quality requirements despite increasing product individuality, varying batch sizes and a scarcity of resources. The transfer of experience-based knowledge in a flexible and self-optimizing production and process planning offers the potential to meet these challenges. Biological systems solve conceptually similar challenges pertaining to the transfer of knowledge, flexibility of individual reactions and adaptation over time. Thus, in the context of digital transformation, mechanisms derived from biology are interpreted and applied to the knowledge domain of production technology. To be able to exploit the potential of bio-inspired production systems, genetic and intelligent properties of technical components and machines were identified and brought together under the concept of “Gentelligence”. Expanding upon this concept with the new idea of process-DNA and biologically inspired optimization algorithms facilitates a more flexible, learning and self-optimizing production, which is shown in three different applications. By using the new concept of gentelligent process planning it is possible to determine machine-specific process parameters in turning processes in order to ensure appropriate roughness within the requirements. Furthermore, the combination of the concept with a material removal simulation allows the determination of the resulting process force in tool grinding for subsequent unknown workpiece geometries. As a result of using the process-DNA, a workpiece-independent knowledge transfer and thus process adaptation for shape error compensation becomes possible. Gentelligent production scheduling enables a process-parallel, holistically optimized machine allocation, and as a result, a significantly reduced lead time. © 2020 The Authors

Bestehende Ansatze zur Integration der Produktions- und Instandhaltungsplanung bleiben auf die Koordination weniger Teildisziplinen der Produktions-/Instandhaltungsplanung begrenzt. Zudem wird die Komplexitat realer Produktionssysteme unzureichend bzw. unter stark restriktiven Annahmen abgebildet, so dass wesentliche Erkenntnisse hinsichtlich der Interdependenzen der Produktions- und Instandhaltungsplanung als auch einer Koordination auf den einfachsten Fall eingeschrankt bleiben und kaum auf industrielle Anwendungen ubertragbar sind. Daher wurde am Institut fur Fertigungstechnik und Werkzeugmaschinen ein dynamischer Planungsansatz zur Koordination der Produktions- und Instandhaltungsplanung mittels Simulationstechnik entwickelt und umgesetzt. Hierdurch werden die Defizite bestehender Ansatze aufgehoben, die Komplexitat, Dynamik und Stochastik realer Produktionssysteme berucksichtigt und eine hohe Planungsqualitat gewahrleistet. Grundlage fur den Transfer dieser Methodik in die Praxis ist deren Erweiterung um eine einfache und schnelle Modellerstellung und -pflege durch selbstparametrierende bzw. anlernende Simulationsmodelle (z. B. mit Echtzeitdaten aus den datenverarbeitenden Systemen der Produktion).

It is often unknown whether a new approach for production control, e. g. control strategies, leads to improvement in a real production system. Due to concerns about poor performance, it is practically impossible to test such approaches in real production. In addition, it is doubtful that testing in a single production system allows a general statement about potentials and limitations. There is a lack of a standardized method for generating validation model sets with which new approaches can be tested. To overcome this problem, this paper presents a tool for the automatic generation of simulation model sets. These model sets represent different types of production systems and thus enable the standardized and structured validation of new production planning and control methods.