wbk Institut für Produktionstechnik

Philipp Gönnheimer, M.Sc.

  • 76131 Karlsruhe
    Kaiserstraße 12

Philipp Gönnheimer, M.Sc.

Forschungs- und Arbeitsgebiete:

  • Steuerungs- und Kommunikationstechnik in Maschinen und Anlagen
  • Baukastenkonzepte für Produktionsanlagen
  • Themenbereich Industrie 4.0



  • Wertstromkinematik – Innovative, wandlungsfähige Produktion der Zukunft
  • I4TP - Deutsch-Chinesische Industrie 4.0 Fabrikautomatisierungsplattform
  • ROBOTOP - Modulare, offene und internetbasierte Plattform für Roboter-Anwendungen in Industrie und Service


[ 1 ] Barton, D.; Gönnheimer, P.; Qu, C. & Fleischer, J. (2018), „Self-describing connected components for live information access within production systems“. 4th International Conference on System-Integrated Intelligence: Intelligent, Flexible and Connected Systems in Products and Production, Hrsg. Denkena, B.; Thoben, K. & Trächtler, A., S. 250-257.
Access to data from components in production systems is potentially an enabler for various data-based approaches. This paper presents a practical approach to transform mechanical components into self-describing cyber-physical systems connected within a local network. The requirements for typical use cases are analysed and a modular cyber-physical connector is proposed. The data is collected by a central OPC UA client and fed into a web-based visualisation, so that it is easily accessible for operators, maintenance staff, and other stakeholders. The approach is illustrated for components with two different levels of complexity.

[ 2 ] Gönnheimer, P.; Kimmig, A.; Mandel, C.; Stürmlinger, T.; Yang, S.; Schade, F.; Ehrmann, C.; Klee, B.; Behrendt, M.; Schlechtendahl, J.; Fischer, M.; Trautmann, K.; Fleischer, J.; Lanza, G.; Ovtcharova, J.; Becker, J. & Albers, A. (2019), „Methodical approach for the development of a platform for the configuration and operation of turnkey production systems“. Procedia CIRP, Hrsg. Putnik, G., S. 880-885.
Shorter product lifecycles lead not only to faster time-to-market for products but also to the need for just as fast available associated production systems. These shorter product lifecycles, as well as the increasing individualization of products, also result in further decreasing production lot sizes. Young companies in China in particular are characterized by a very high speed of innovation but may not have the necessary manufacturing knowledge or capacities to bring their developed products to the market with a scalable production. For this reason, there is a great need to quickly set up and commission turnkey production systems or to reconfigure existing production systems for new production tasks in the shortest possible time. This paper describes the design and architecture of a cloud platform with the aim to support a manufacturer independent design process for turnkey production systems. This process ranges from the product to be manufactured to the operation of the production system. Firstly, the structure and methodology used to link the various objectives are discussed. The system for recording and structuring product and production system data to create reusable modules from components and machines is described. Subsequently, the use of standardized modules is developed to support reconfiguration of the production system during operation. In addition, the digital business models tailored to the production system are proposed to the platform user for commissioning and operation of the plant. A case study is conducted to validate the proposed methodology.

[ 3 ] Barton, D.; Gönnheimer, P.; Schade, F.; Ehrmann, C.; Becker, J. & Fleischer, J. (2019), „Modular smart controller for Industry 4.0 functions in machine tools“. Procedia CIRP, Hrsg. Butala, P.; Govekar, E. & Vrabič, R., S. 1331-1336.
In machine tools, Industry 4.0 functions can increase availability through predictive maintenance, while other functions improve productivity and workpiece quality through process supervision and optimisation. Many of these functions rely on data communication between systems from different suppliers. Requirements regarding latency and computing vary widely depending on the application. Based on an analysis of these requirements, a smart controller for the implementation of Industry 4.0 is designed, using a hypervisor to allow for the integration of soft real-time and best-effort applications.

[ 4 ] Gönnheimer, P.; Hillenbrand, J.; Betz-Mors, T.; Bischof, P.; Mohr, L. & Fleischer, J. (2019), „Auto-configuration of a digital twin for machine tools by intelligent crawling“. Production at the leading edge of technology, Hrsg. Wulfsberg, J. P.; Hintze, W. & Behrens, B., S. 543-552.
The digitalisation of production technology is becoming increasingly important today and will play a key role in machine tools in the future. In order to generate as much information as possible about machines and components as well as the product, the number of sensor systems and further devices is con-stantly increasing. A challenge with this increasing number of data sources is the also increasing complexity of the system with regard to the generation of the ma-chine tool’s digital twin that is to be fed from these data sources. For this, a ho-listic approach is necessary which combines all parameters and values with a uniform semantics and links them to the corresponding data sources. For this purpose, this paper uses a uniform information model to describe a machine tool, which is linked to the respective node variable names in the respective OPC namespace of the machine tool. In the event that this linkage plan is incomplete or completely missing, this paper presents a concept for a so-called crawler tool, an OPC client application. The crawler searches the parameters and values in the described OPC namespace of the machine tool, identifies them by domain knowledge-driven plausibility checks and assigns them to the corresponding pa-rameters of the information model of the machine tool.

[ 5 ] Gönnheimer, P.; Kimmig, A.; Ehrmann, C.; Schlechtendahl, J.; Güth, J. & Fleischer, J. (2019), „Concept for the Configuration of Turnkey Production Systems“. Procedia CIRP, Hrsg. Dietrich, F. & Krenkel, N., S. 234-238.
Shorter product lifecycles and increasing individualization of products lead to the necessity for a reoccurring process, which includes the selection and configuration of production systems to provide a system that produces the product. Especially in fast developing countries like China, the offer for this knowledge can hardly supply the demand. In order to solve this, this paper presents a systematic approach in the form of a multi-stage process. In the first stage, a configuration logic maps product requirements with the properties and specifications of production machines together with equipment and matches them using a uniform data information model for both products and production modules. In the second stage, the turnkey production system is set up, commissioned and operated based on the Industrie 4.0 administration shell. The presented approach has been prototypically implemented on an online platform and demonstrated on a real production system using a new product that has been integrated into production.

[ 6 ] Netzer, M.; Gönnheimer, P.; Michelberger, J. & Fleischer, J. (2020), „Skalierbarkeit von KI-Anwendungen in der Produktion“, Fabriksoftware, S. 25. 10.30844/FS20-1_51-54
Bereits heute existieren vereinzelt vielversprechende industrielle Anwendungen der Künstlichen Intelligenz, vor allem in den Bereichen Prozess- und Zustandsüberwachung. Heutige KI-Modelle werden jedoch ausschließlich als Insellösungen für einen Prozess und eine Maschine entwickelt. Durch heterogene Produktionsanlagen existieren kaum prozess- und zustandsübergreifend anwendbare KI-Modelle. Wie gelingt daher eine breite Übertrag- und Skalierbarkeit der Anwendungen in der gesamten Produktion? Dies erfolgt einerseits durch die Vereinheitlichung der Informationsmodelle verschiedener Maschinen durch intelligente Parameteridentifikation (Crawling) sowie in einem zweiten Schritt durch eine Datensegmentierung zum Aufbau strukturierter Datenbasen (Clustering). Auf Grundlage von kontextbasieren Datenbasen, die aus einem einheitlichen Informationsmodell aus unterschiedlichsten Maschinen entstehen, können KI-Ansätze skaliert und auf die gesamte Produktion übertragen werden.

[ 7 ] Gönnheimer, P.; Netzer, M.; Mohr, L.; von Hörsten, G. & Fleischer, J. (2020), „Erhöhung der Skalierbarkeit von KI-Anwendungen in Produktionsanlagen durch intelligente Parameteridentifikation und Datensegmentierung“, ZWF Zeitschrift für wirtschaftlichen Fabrikbetrieb, S. 517-519. 10.3139/104.112318 [30.11.-1].
Die Digitalisierung von Maschinen und Anlagen im Kontext von Anwendungen der K?nstlichen Intelligenz (KI) gewinnt heute zunehmend an Bedeutung. Heutige Anwendungen im Bereich KI werden jedoch größtenteils als prozess- oder maschinenspezifische Insellösung entwickelt, deren Übertragbarkeit durch die Heterogenität von Produktionsanlagen stark eingeschränkt ist. Ziele aktueller Forschungsarbeiten sind daher eine intelligente Parameteridentifikation zur Findung von Datenquellen und Vereinheitlichung von Informationsmodellen verschiedener Maschinen sowie eine darauf aufbauende Datensegmentierung strukturierter Datenbasen.

[ 8 ] Gönnheimer, P.; Puchta, A. & Fleischer, J. (2020), „Automated Identifcation of Parameters in Control Systems of Machine Tools“. Production at the leading edge of technology, Hrsg. Behrens, B.; Brosius, A.; Hintze, W.; Ihlenfeldt, S. & Wulfsberg, J. P., Springer, Berlin, Heidelberg, S. 568-577.
Especially in the context of Artifcial Intelligence (AI) applications and increasing Overall Equipment Effectiveness (OEE) requirements, the use of data in production is gaining in importance. Applications in the feld of process or condition monitoring use, for example, machine component parameters such as motor currents, travel speeds and position information. However, as the data is usually only accessible in the machine control systems in non-standard structures and semantics, while having a large number of potential variables, the identifcation and use of these parameters and data sources represents a signifcant challenge. This paper therefore presents an approach to automatically identify and assign machine parameters on the basis of time series data. For the identifcation, feature- and deep learning-based classifcation approaches are used and compared. Classifcation results show a general usability of the approaches for the identifcation of machine parameters.

[ 9 ] Mühlbeier, E.; Gönnheimer, P.; Hausmann, L. & Fleischer, J. (2020), „Value Stream Kinematics“. Production at the leading edge of technology, Hrsg. Behrens, B.; Brosius, A.; Hintze, W.; Ihlenfeldt, S. & Wulfsberg, J. P., Springer, Berlin, Heidelberg, S. 409-418.
The trend towards individualized products and the increasing demand for a greater variety of variants create new challenges for existing production environments and require a re-thinking of production. Established manufacturing systems that provide the desired flexibility are associated with significant productivity restrictions and are therefore unable to compete economically with production from rigid production lines. They are therefore often limited to serving niche markets. Consequently, an approach is needed that combines high productivity with high flexibility. For this purpose, this paper presents a new approach to manufacturing with an equally high productivity and flexibility, so-called value stream kinematics. The basic idea of value stream kinematics is to combine the advantages of specialized machines with the versatility of industrial robots. The vision behind this is to be able to realize entire value streams with uniform robot-like kinematics and no need for special machines.

[ 10 ] Qu, J.; Barton, D.; Gönnheimer, P.; Pinsker, F.; Kufer, D. & Fleischer, J. (2020), „Self-Aware LiDAR Sensors in Autonomous Systems using a Convolutional Neural Network“. Intelligent, Flexible and Connected Systems in Products and Production, Hrsg. Thoben, K.; Dekena, B.; Lang, W. & Trächtler, A., Elsevier, S. 50-55.
Autonomous systems, as found in autonomous driving and highly automated production systems, require an increased reliability in order to achieve their high economic potential. Self-aware sensors are a key component in highly reliable autonomous systems. In this paper we highlight a proof of concept (PoC) of a deep learning method that enables a LiDAR (Light detection and ranging) sensor to detect functional impairment. More specifically, a deep convolutional neural network (CNN) is developed and trained with labelled LiDAR data in the form of point clouds to classify the degree of impairment of its functionality. The results are statistically significant and can be regarded as a general classifier for objects within LiDAR data, applied to selected cases of sensor impairment. In detecting impairment and evaluating the correctness of the captured data, the sensor gains a basic form of self-awareness. The presented methods and insights pave the way for improved safety of autonomous systems by the means of more sophisticated ?self-aware? neural networks.