Virtual Production Intelligence
||Breaking up the disharmony of models used in different design domains by providing semantic connectivity and explorative analysis|
(1) Model map of virtual production systems: Comprehensive and unified virtual production environments by harmonization and generalization of the underlying deterministic, cybernetic, economic and sociological models.
(2) Analysis and evaluation methods: Explorative analysis of simulated and/ or real production processes, accessible via context-adaptive visualization and an integrated knowledge model facilitating the interconnection of different design domains.
The Research Area “Virtual Production Intelligence” (VPI) focusses on the integrated support of collaborative planning processes for production systems and products. The research area focusses the research of processes for information processing in the design domains “factory” and “machine”. These processes provide the integration and interactive analysis of arising, mostly heterogeneous planning information. The demonstrators (flapAssist, memoSlice und VPI platform) that are information systems serve for the validation of the scientific approaches and aim to realize a continuous and consistent information management in terms of the Digital Factory. Central challenges are the data and information integration (e.g. by means of metamodeling), the subsequent evaluation as well as the visualization of planning information (e.g. by means of Virtual Reality (VR)). All scientific and technical work is done within an interdisciplinary team composed of engineers, computer scientists and physicists.
Concerning the design domain “factory” an information model for factory planning has been developed and implemented within the VPI platform using the concept of ontologies. Based on this, a process to define KPI has been established that automatically calculates KPI from semantically annotated planning data and visualizes them in terms of a cockpit. A further KPI for the value stream oriented assessment of the position of process areas in factory layouts has been developed and visualized within the AixCAVE by means of the demonstrator flapAssist. Furthermore, flapAssist has been extended by an annotation system, which allows users to capture comments and decisions during ongoing virtual walkthroughs e.g. via multimodal text input technique in immersive VR systems. All required interaction workflows have been realized by means of an integration concept that is new to VR and have been confirmed in two user studies.
Concerning the design domain “machine”, in the context of modeling and simulation for laser manufacturing processes, a new numerical as well as analytical model for laser drill-ing have been implemented and validated to determine and describe the ablation shape of laser material processing. Furthermore, additional efforts were focused on incorporating a domain-decomposition module that gives the overall structure of the system response in the high-dimensional parameter space (via the Morse-Smale-Complex). Additionally, a clas-sification method has been developed that decomposes the domain space into feasible and non-feasible regions. Concerning the demonstrator memoSlice, new visualization techniques have been integrated whereby analysis options for metamodels were improved as well as the navigation in the multi-dimensional data space were simplified.
Concerning the VPI platform, the data management has been rearranged towards a model-driven approach that is based on the developed ontologies. Furthermore, the interaction concept of the web application has been renewed and a new design was developed. Within both design domains, new functionalities were implemented that provide interactive analy-ses of planning information.
Additionally, the technical integration of the demonstrators memoSlice and flapAssist al-lows the possibility for users to manipulate and test new machine configurations during an ongoing virtual factory walkthrough.
In the context of sensitivity analysis the implementation of variance-decomposition meth-ods has been completed to generate so-called Clique Graphs showing the main sensitivity effects of the parameters on the criteria and the interdependence between those. A new demonstrator based on a Touch-Terminal for increased interactivity was established. At the same time a direct coupling between the metamodel visualization and the direct simulation via the reduced models (mentioned in paragraph “scientific”) was established.
In 2015 our scientific and technical results were presented in two different fairs/congresses each addressing a national and an international audience. The achieved results concerning the modeling and simulation for laser manufacturing processes were presented to an inter-national audience at the LASER World of Photonics fair in Munich 2015. Additionally, the factory planning research topics ware discussed on the congress ‘Excellent Factory Planning 2015’ in Aachen including a visit of the AixCAVE.
Concerning the design domain “factory”, the support of factory planning projects will be enhanced by means of the continuous information management of the demonstrators VPI platform and flapAssist. This is validated with a real planning project. Concerning the design domain “machine”, fundamental extension of metamodeling will be performed which will enable the modeling of spatially-distributed quantities. This requires the investigation and implementation of further techniques in the field of numerics, machine learning and data mining. Thereby, e.g. a tool that is based on the Buckingham theory for reducing the number of physical parameters into a smaller set of dimensionless parameters will be developed. Besides of the developments within the single domains, a comprehensive scenario will be defined that joins all research results. This scenario is based on the planned integration of the existing demonstrators.
For more information, please visit our page of the technical demonstrator.