Currently, European small and medium-sized enterprises (SMEs) are experiencing increasing pressure to provide high quality goods with customised features while at the same time remain cost effective and competitive in the global market. In the future, manufacturing systems need to be able to cope with constantly changing market requirements. Consequently, there is a need to develop the research foundations for a new generation of manufacturing systems composed of intelligent autonomous entities which are able to reconfigure themselves and to adapt their performance as a result of product and environmental changes. The research described in this thesis addresses the issue by developing three distinctive elements of an adaptation framework for next-generation manufacturing systems. The first element is a capability-based data model for the representation of manufacturing resources to enable self-awareness. The model captures the resources’ life cycle and performance indicators to provide information about the resources’ condition. The second element is a multi-agent architecture for plug and produce and the reconfiguration of manufacturing systems. The resource data model is utilised by the agent society, which is able to instantiate a model to represent a physical resource in the virtual agent society. The shift to the virtual environment enables a communication infrastructure for heterogeneous resources and the application of the digital twin concept. The agent architecture applies negotiation techniques to establish a plan for system adaptation. The third element is a methodology for automated experience-based manufacturing system adaptation. The adaptation methodology is based on previous runtime experience instances to generate adaptation knowledge. The information generated is applied to the current context and part of the agent negotiation which is dynamically executed in case of a disturbance. Collectively, these three elements significantly increase the flexibility and reconfigurability of a manufacturing system reducing the time required for integration and maintenance of complex systems on demand, improving their effectiveness. The developed framework is implemented and evaluated experimentally on a physical, industrial standard demonstrator and using a virtual simulation model. The experimental results confirm a significant step towards new solutions for the deployment of self-adaptable intelligent manufacturing systems.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:706394 |
| Date | January 2017 |
| Creators | Antzoulatos, Nikolas |
| Publisher | University of Nottingham |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://eprints.nottingham.ac.uk/39583/ |
Page generated in 0.0016 seconds