Fraunhofer-Institut für Produktionsanlagen und Konstruktionstechnik IPK

Filters

12
Reset filters

Settings
Now showing 1 - 10 of 12
  • Publication
    Cluster identification of sensor data for predictive maintenance in a Selective Laser Melting machine tool
    ( 2018)
    Uhlmann, E.
    ;
    Pastl Pontes, R.
    ;
    Geisert, C.
    ;
    Hohwieler, E.
    Selective laser melting has become one of the most current new technologies used to produce complex components in comparison to conventional manufacturing technologies. Especially, existing selective laser melting machine tools are not equipped with analytics tools that evaluate sensor data. This paper describes an approach to analyze and visualize offline data from different sources based on machine learning algorithms. Data from three sensors were utilized to identify clusters. They illustrate the normal operation of the machine tool and three faulty conditions. With these results, a condition monitoring system can be implemented that enables those machine tools for predictive maintenance solutions.
  • Publication
    Improvement of Defect Detectability in Machine Tools Using Sensor-based Condition Monitoring Applications
    ( 2018)
    Demilia, G.
    ;
    Gaspari, A.
    ;
    Hohwieler, E.
    ;
    Laghmouchi, A.
    ;
    Uhlmann, E.
    The aim of this paper is to present a reliable methodology for condition monitoring of components of high performance centerless grinding machines. This enables the detection and localization of the defects on the ball screw. The fault detection is realized using a self-implemented classification algorithm and other pattern recognition algorithms. The diagnosis is based on acceleration and AE measurement data performed on an axis test rig using various damaged ball screws at different operating parameters. Moreover, the structure of the pattern recognition process will be introduced, this includes the signal pre-processing, the selection of the most suitable features for this specific application using MATLAB®. Finally, the evaluation of the developed solution will be showed. The actions allowing improving the accuracy of measurement data and the effectiveness of the processing algorithms, based on MATLAB® applications, are described throughout the paper. Satisfactory classification results will be obtained and discussed.
  • Publication
    Intelligent production systems in the era of industrie 4.0 - changing mindsets and business models
    ( 2017)
    Uhlmann, E.
    ;
    Hohwieler, E.
    ;
    Geisert, C.
    Industrie 4.0 has been becoming one of the most challenging topic areas in industrial production engineering within the last decade. The increasing and comprehensive digitization of industrial production processes allows the introduction of innovative data-driven business models using cyber-physical systems (CPS) and Internet of Things (IoT). Efficient and flexible manufacturing of goods assumes that all involved production systems are capable of fulfilling all necessary machining operations in the desired quality. To ensure this, production systems must be able to communicate and interact with machines and humans in a distributed environment, to monitor the wear condition of functionally relevant components, and to self-adapt their behaviour to a given situation. This article gives an overview about the historical development of intelligent production systems in the context of value-adding business models. The focus is on condition monitoring and predictive maintenance in an availability oriented business model. Technical as well as organizational prerequisites for an implementation in the production industry are critically analysed and discussed on the basis of best practice examples. The paper concludes with a summary and an outlook on future research topics that should be addressed.
  • Publication
    Smart wireless sensor network and configuration of algorithms for condition monitoring applications
    ( 2017)
    Uhlmann, E.
    ;
    Laghmouchi, A.
    ;
    Geisert, C.
    ;
    Hohwieler, E.
    Due to high demand on availability of production systems, condition monitoring is increasingly important. In recent years, the technical development have improved for realization of condition monitoring applications as a result of technological progress in fields such as sensor technology, computer performance and communication technology. Especially, the approaches of Industrie 4.0 and the use of the Internet of Things (IoT) technologies offer high potential to implement condition monitoring solutions. The connection of several sensor data of components to the cloud allows the identification of anomalies or defect pattern, this information can be used for predictive maintenance and new data-driven business models in production industry. This paper illustrates a concept of a smart wireless sensor network for condition monitoring application based on simple electronic components such as the single-board computer Raspberry Pi 2 modules and MEMS (Micro-Electro-Mechanical S ystems) vibration sensors and communication standards MQTT (Message Queue Telemetry Transport). The communication architecture used for decentralized data analysis using machine learning algorithms and connection to the cloud is explained. Furthermore, a procedure for rapid configuration of condition monitoring algorithms to classify the current condition of the component is demonstrated.
  • Publication
    Intelligent pattern recognition of SLM machine energy data
    ( 2017)
    Uhlmann, E.
    ;
    Pastl Pontes, R.
    ;
    Laghmouchi, A.
    ;
    Hohwieler, E.
    ;
    Feitscher, R.
    Selective Laser Melting (SLM) is an additive manufacturing process, in which the research has been increasing over the past few years to meet customer-specific requirements. Different parameters from the process and the machine components have been monitored in order to obtain vital information such as productivity of the machine and quality of the manufactured workpiece. The monitoring of parameters related to energy is also realized, but the utilisation of such data is usually performed for determining basic information, for instance, from energy consumption. By applying machine learning algorithms on these data, it is possible to identify not only the steps of the manufacturing process, but also its behaviour patterns. Along with these algorithms, evidences regarding the conditions of components and anomalies can be detected in the acquired data. The results can be used to point out the process errors and component faults and can be adopted to analyse the energy efficiency of the SLM process by comparing energy consumption of one single layer during the manufacturing of different components. Moreover, the state of the manufacturing process and the machine can be determined automatically and applied to predict failures in order to launch appropriate counter measures.
  • Publication
    Zustandsüberwachung in der Cloud
    ( 2016)
    Uhlmann, E.
    ;
    Laghmouchi, A.
    ;
    Hohwieler, E.
    ;
    Geisert, C.
    Aufgrund der hohen Verfügbarkeitsanforderungen an Produktionsmaschinen wächst das Interesse an zustandsbasierter Instandhaltung. Der Einsatz von Zustandsüberwachungssystemen (Condition Monitoring-Systemen) zur Steigerung der Verfügbarkeit von Maschinen und zur Reduktion der Instandhaltungskosten spielt dabei eine entscheidende Rolle und hat in den letzten Jahren zugenommen. Da am Markt verfügbare und auf Industriesensoren basierende Lösungen meist anwendungsspezifisch, kostenintensiv und in der Inbetriebnahme aufwändig sind, wurde am Fraunhofer IPK ein Konzept für die Zustandsüberwachung in der Cloud entwickelt, das mithilfe von Einplatinen-Computern und MEMS-Beschleunigungssensoren (Mikro-Elektro-Mechanisches-System) als Sensorknoten eine preisgünstige und einfach zu handhabende Alternative darstellt.
  • Publication
    Intelligentes Elektroniksystem für Condition Monitoring in Industrie 4.0
    ( 2016)
    Uhlmann, E.
    ;
    Laghmouchi, A.
    ;
    Ehrenpfordt, R.
    ;
    Hohwieler, E.
    ;
    Geisert, C.
    Im Rahmen dieses Beitrags werden die geplanten Arbeiten des Forschungsprojekts ""AMELI4.0"" vorgestellt. Der Schwerpunkt dieses Projekts liegt in der Entwicklung und Umsetzung hochintegrierter, vernetzter, energieautarker MEMS-Multisensorsysteme (Mikro-Elektro-Mechanische Systeme - MEMS) mit intelligenter Echtzeit-Datenverarbeitung auf Sensorebene bei hoher Daten- und Systemsicherheit. Das Multisensorsystem integriert mehrere MEMS-Sensoren zur Erfassung von Körperschall und akustischer Schall in Kombination mit der energieeffizienten Signalvorverarbeitung auf Sensorebene (Smart Data statt Big Data) bei hoher Systemrobustheit in einem modularen Hardware- und Plattformdesign. Des Weiteren werden die adressierten Anwendungsfälle und der Forschungsschwerpunkt des Fraunhofer IPK zum Thema Datenanalyse und Datenmanagement vorgestellt.
  • Publication
    Condition monitoring in the cloud
    ( 2015)
    Uhlmann, E.
    ;
    Laghmouchi, A.
    ;
    Hohwieler, E.
    ;
    Geisert, C.
    Due to the very high demands on availability and efficiency of production systems and industrial systems, condition-based maintenance is becoming increasingly important. The use of condition monitoring approaches to increase the machine availability and reduce the maintenance costs, as well as to enhance the process quality, has increased over the last years. The installation of industrial sensors for condition monitoring reasons is complex and cost-intensive. Moreover, the condition monitoring systems available on the market are application specific and expensive. The aim of this paper is to present the concept of a wireless sensor network using Micro-Electro-Mechanical Systems - MEMS sensors and Raspberry Pi 2 for data acquisition and signal processing and classification. Moreover, its use for condition monitoring applications and the selected and implemented algorithm will be introduced. This concept realized by Fraunhofer Institute for Production Systems and Design Technology IPK, can be used to detect faults in wear-susceptible rotating components in production systems. It can be easily adapted to different specific applications because of decentralized data preprocessing on the sensor nodes and pool of data and services in the cloud. A concrete example for an industrial application of this concept will be represented. This will include the visualization of results which were achieved. Finally, the evaluation and testing of this concept including. implemented algorithms on an axis test rig at different operation parameters will be illustrated.
  • Publication
    Selbstorganisierende Produktion
    ( 2013)
    Uhlmann, E.
    ;
    Hohwieler, E.
    ;
    Kraft, M.
    Zukünftig sollen in der selbstorganisierenden Produktion mit verteilter Intelligenz die Objekte in der Produktion zusätzlich mit eigener Intelligenz ausgestattet werden und so Aufgaben der Koordination und Steuerung von Produktionsabläufen übernehmen können. Diese produktgesteuerte Fertigung sieht statt der bisherigen zentralen Planung und Steuerung ein Multiagentensystem mit der Möglichkeit zu Auktionen durch Verhandlungsmechanismen als Mittel zur Selbstorganisation vor. Der Beitrag gibt einen Überblick über die im Projekt "Selbstorganisierende Produktion - SOPRO" verfolgten Ansätze bei denen Softwareagenten zwischen Aufträgen und Ressourcen zur Festlegung der Bearbeitungsreihenfolge verhandeln.
  • Publication
    Monitoring of slowly progressing deterioration of computer numerical control machine axes
    ( 2008)
    Uhlmann, E.
    ;
    Geisert, C.
    ;
    Hohwieler, E.
    The feed axes of computer numerical control (CNC) grinding machine tools are among the most mechanically stressed components of machine tools owing to the high process forces and rough manufacturing environment which they encounter. The resulting wear and tear depends strongly on the product range and the manner of machine operation. To counteract a functional deficiency of these central machine units, the current usual approach is preventive maintenance. The manual inspection of feed axes is complex and time consuming. A complicating matter is that the deterioration normally progresses very slowly and depends on the position of the stress along the axis. Existing approaches to automated estimation of the 'health status' of feed axes do not take this factor into account. This paper presents a procedure that addresses this gap. During simple test routines, the drive current, axis position, and feed rate are recorded. With the help of additional machine data, characteristic values are computed directly at the computer of the human-machine interface (HMI). The results are then transferred to and stored on a database server at the machine manufacturer. This approach enables the service technician to trace the progression of the axes' 'health status' over a long time. This approach makes it possible to detect trends in the characteristic values at an early point in time. This leads to a better planning of necessary maintenance actions adapted to the remaining lifetime of the wearing component.