Integrated system for knowledge discovery from heterogeneous systems (Ideal - ist Partner Search)
FOF 08 - ICT - enabled modelling, simulation, analytics and forecasting technologies
Deadline Date: 04-02-2015 17:00:00 (Brussels local time)
Recent trends in solution, design and implementation of process control systems can be structurally characterised by hierarchical arrangement of hardware and software tools. The basic prerequisite for the hierarchical control system to perform safe and reliable control functions is an adequate information support (data collection and processing, check data validity, data transfer between individual levels of control, etc.). This project proposal is aimed at solving the above issues.
The project is to result in proposing the process of knowledge discovery from heterogeneous levels of control. The control systems utilised at horizontal and vertical levels of control are assumed to be inhomogeneous. The proposal definition is supposed to represent a holistic approach to solving issues related to collecting and processing a large amount of data for the purposes of complex system control.
Based on the research and development in the area of intelligent and knowledge-oriented technologies, the project proposal intends to encourage using the methods of knowledge-based economy in production enterprises.
Data base information model for hierarchical systems of enterprise control
The purpose is to create a holistic proposal for a unified data model integrating data from all levels of control. The project proposal envisages the design of a universal methodology for identifying the appropriate data with respect to the key performance indicators resulting in defining an open standard for objects and their relations.
Processing a large amount of data by other than the traditional techniques (e.g. big data) is presumed. Data will be collected from all hierarchical levels of control and transferred to a common data base through a communication subsystem.
The aim is to apply the process mining techniques for mapping the business processes. For each process, these techniques enable the deviations from the defined rules to be identified based on the available data.
Comparing the data mining methods according to relevant metrics.
The aim is to propose a methodology for selecting the appropriate metrics possible to assess the specific methods for the purposes of the implementation of the knowledge discovery process not only individually, but also collectively. The objective of comparing the accuracy of the data mining methods is the analysis and calculation of errors according to a defined methodology. The result is to determine the suitability of the analysed methods for specific sets of problems.
Formulation of conceptual proposal for knowledge discovery process in control processes.
The aim is to create a conceptual proposal for knowledge discovery in hierarchical control systems. The proposal can be formulated as a holistic approach to solving the problems related to processing of a large amount of data for the purposes of complex system control.
Knowledge discovery for the support of the control in heterogeneous systems.
Applying the process of knowledge discovery using the selected methods, the knowledge of the given process is to be acquired. The aim is to forecast the behaviour of manufacturing systems and processes with respect to the defined Key Performance Indicators (KPI). The discovered knowledge of the analysed process is to be generalised using production rules.
Knowledge verification utilising simulation models (Virtual commissioning)
The purpose is to design and debug the interaction and mutual correlation between virtual proposal and real manufacturing technologies in various industrial areas. A creation of virtual model based on real actually existing technology is presumed. The purpose is to establish the procedures and methodologies for virtual commissioning supporting the Factory Acceptance Testing (FAT) implementation and automatic generation of relevant documentation.
The discovered knowledge of system and process control is to be verified on the proposed models. The models will provide information on the prediction of the given system and process behaviour as well.
Proof of concept