Fraunhofer-Institut für Produktionsanlagen und Konstruktionstechnik IPK

Filters
Reset filters

Settings
Now showing 1 - 10 of 107
  • Publication
    An Ensemble Learning based Hierarchical Multi-label Classification Approach to Identify Impacts of Engineering Changes
    ( 2020)
    Pan, Y.
    ;
    Stark, R.
    In the process of complex products development, design decisions of products are constantly changed to improve quality or functionality, reduce costs, respond to statutory constraints or implement wishes from customers, with respect to new functionality. The changes on already released design decisions are known as Engineering Changes (EC). Due to interdependency, changes on one component may cause changes on another, and this will spread along the product structure. Therefore, to completely identify the affected components of engineering changes is a major challenge. This paper presents a novel approach to use properties of components as target variables and applying the predicted properties to locate the EC affected components in product structure. We create a hierarchy of the properties and divide the label space into separate communities. A stacked multi-label classifier is trained in each community, the result is obtained by union of assigned labels from different communities. Finally, the predicted labels are adjusted by incorporating the hierarchical relation. Experiments conducted on real-world industrial EC dataset with mixed data types. Results demonstrated that, the ensemble framework in our approach is more efficient and effective than our baseline models and has achieved superior performance on real industrial engineering change data.
  • Publication
    Stufenmodell zur Strukturierung digitaler Assistenzfunktionen: Digital assistierte Montageplanung
    ( 2020)
    Bußwinkel, L.
    ;
    Hauk, J.-C.
    ;
    Schneider, H.
    ;
    Stark, R.
    Die Montageplanung ist durch große manuelle Aufwände sowie eine auf Erfahrungen und Expertenwissen basierende Entscheidungsfindung geprägt. Trotz zahlreicher Forschungsaktivitäten wird nur eine geringe Automatisierung der Montageplanung erreicht. Daher wird für die Montageplanung ein Assistenzstufenmodell erarbeitet, das zukünftig eine systematische Entwicklung digitaler Assistenzsysteme erlaubt, die in der industriellen Praxis einen Mehrwert erzeugen.
  • Publication
    An optimal algorithm for the robotic assembly system design problem: An industrial case study
    ( 2020)
    Hagemann, S.
    ;
    Stark, R.
    The design process of flow-oriented assembly systems is characterized by being both highly complex and time consuming. Especially those design processes categorized into robotic and multi variant encountered in the automotive body-in-white production stages. Unlike established manual and template-based assembly system design models, which are currently applied in industry, the here presented novel approach uses a knowledge-based search algorithm and automatically generates optimal assembly system configurations. The algorithm has been implemented in a software prototype and the results have been validated against different large-size industrial scenarios in the automotive field of body-in-white production.
  • Publication
    Integration of automated structure mechanic analyses into production process simulation
    ( 2019)
    Schmitz, M.B.
    ;
    Stark, R.
    To produce individual variants of a given product, each variant has to be assessed with respect to their mechanical integrity. This is usually done in the course of product design by assessing all required variants. However, during the lifetime of a product several changes of the existing variants occur or new variants will be created. In this paper a method is described how to integrate and automate the assessment of the mechanical integrity of the product variants directly into the production process by exploiting modern computer methods and automated set-up of state-of-the-art CAE methods. We show this on a demonstrator of a smart factory, that produces disks which can be customized to a large extend. The disks undergo a milling process, which alters the structural properties of the disks. Generally, it cannot be guaranteed upfront that a variant that can be produced is also strong enough to withstand the loads in the application. Therefore, the integrated FEM analyses are computing and assessing the load cases deduced from the use cases of the disks and the loads that occur during milling and handling in the factory. The assessment of the results is done automatically and if positive the production process is started.
  • Publication
    Lücken und Herausforderungen bei der praktischen Umsetzung des Model-Based Systems Engineering
    ( 2019)
    Fresemann, C.
    ;
    Stark, R.
    ;
    Sauer, C.
    ;
    Schleich, B.
    ;
    Wartzack, S.
  • Publication
    Configuration Equilibrium Model of Product Variant Design Driven by Customer Requirements
    ( 2019)
    Yang, Q.
    ;
    Bian, X.
    ;
    Stark, R.
    ;
    Fresemann, C.
    ;
    Song, F.
    In view of the dynamic change of customer requirements (CRs) during the process of product use, in this paper we propose a Bayesian Nash equilibrium configuration model for product variant design driven by CRs. By analyzing CRs, the complete variant requirements of the products can be obtained. Combined with modularization and parameterization variant design methods, a parametric variant instance is proposed. Since cost and delivery time are affected by the product variant design, firms and customers are established as two decision-making bodies, and Bayesian Nash theory is introduced to the product configuration. The theory takes the product cost and customer satisfaction as the payoff function of the game, and based on the threshold value search of the customer satisfaction it determines the strategy set of the two parties. The Nash equilibrium solution equation is established and solved by a simulated annealing algorithm. The optimal product configuration scheme satisfying the interests of both sides of the game is obtained. Finally, the automatic guided vehicle (AGV) is taken as an example to illustrate the effectiveness and practicability of the method.
  • Publication
    Study on the feasibility of modelling notations for integrated Product-Service Systems Engineering
    ( 2019)
    Halstenberg, F.A.
    ;
    Stark, R.
    Developing products and services concurrently bears great potential as well as challenges. Systems Engineering (SE) is an interdisciplinary approach, which aims to enable the realization of successful systems. Model-based systems engineering (MBSE) is a SE approach focusing on the utilization of models for information exchange instead of documents. Methods and tools for MBSE are advanced and effective. Nevertheless, the field predominantly focuses on technical systems rather than on services and business models. In the context of the increasing importance of the Service Economy and rising innovative business models, an extension of the MBSE methods and tools to a service- and business model-centric view constitutes a research gap. Sufficient methods for requirements definition in PSS design have been developed. In concept design, the outline or rough concept of the product is defined. This step is currently not supported with useful methods. In order to address this research gap, the authors have proposed Integrated Product Service Architectures (IPSSAs) in previous research, which depict physical product architectures and service architectures in one integrated model. In previous studies, Systems Modelling Language (SysML) was identified as suitable modelling notations for IPSSAs, but several drawbacks of the notation were noted. Within this paper, a study is presented, which identifies suitable combinations of modelling notations and diagrams for IPSSA modelling. The combination of several diagrams from SysML, Business Process Modelling Notation (BPMN) and Data Flow Architecture is used on the practical example of ELSA (Engineering Living Systems of Autonomy), a service robot for guest and plant care. In order to address the lack of process-based service modelling in SysML, BPMN is added to common SysML modelling. In order address the data flow dimension, a third modelling notation, the Data Flow Architecture is proposed.
  • Publication
    Hybrid artificial intelligence system for the design of highly-automated production systems
    ( 2019)
    Hagemann, S.
    ;
    Sünnetcioglu, A.
    ;
    Stark, R.
    The automated design of production systems is a young field of research which has not been widely explored by industry nor research in recent decades. Currently, the effort spent in production system design is increasing significantly in automotive industry due to the number of product variants and product complexity. Intelligent methods can support engineers in repetitive tasks and give them more opportunity to focus on work which requires their core competencies. This paper presents a novel artificial intelligence methodology that automatically generates initial production system configurations based on real industrial scenarios in the automotive field of body-in-white production. The hybrid methodology reacts flexibly against data sets of different content and has been implemented in a software prototype.
  • Publication
    Neural Network Hyperparameter Optimization for the Assisted Selection of Assembly Equipment
    ( 2019)
    Hagemann, S.
    ;
    Sünnetcioglu, A.
    ;
    Fahse, T.
    ;
    Stark, R.
    The design of assembly systems has been mainly a manual task including activities such as gathering and analyzing product data, deriving the production process and assigning suitable manufacturing resources. Especially in the early phases of assembly system design in automotive industry, the complexity reaches a substantial level, caused by the increasing number of product variants and the decreased time to market. In order to mitigate the arising challenges, researchers are continuously developing novel methods to support the design of assembly systems. This paper presents an artificial intelligence system for assisting production engineers in the selection of suitable equipment for highly automated assembly systems.
  • Publication
    Approaching Knowledge Dynamics Across the Product Development Process with Methods of Social Research
    ( 2019)
    Wang, W.M.
    ;
    Mörike, F.
    ;
    Hergesell, J.
    ;
    Baur, N.
    ;
    Feufel, M.
    ;
    Stark, R.
    Knowledge is a crucial factor in state-of-the-art product development. It is often provided by stakeholders from divers disciplinary and individual backgrounds and has to be integrated to create competitive products. Still, it is not fully understood, how knowledge is generated, transformed, transferred and integrated in complex product development processes. To investigate the dynamic interrelations between involved stakeholders, applied knowledge types and related artefacts, researchers at the TU Berlin conducted and evaluated a student experiment to study basic phenomena of development projects. In relation to research methods and instruments applied in this experiment, various improvement opportunities were identified. In this paper, the experimental setting and its results are critically analysed from a social science perspective in order to generate improved research design. Based on the results of this analysis, a first set of methods and instruments from social sciences are identified that can be applied in further experiments. The goal is to develop a methodological toolbox that can be used to approach research on knowledge dynamics in product development.