INTELL-DIAG will develop an intelligent model-based monitoring, diagnostic and maintenance system for plant, process and product. Models will be partially constructed automatically from the source code for robots, machine tools and PLCs to speed installation, ensure consistency and keep diagnostics models current. INTELL-DIAG will infer heuristic diagnostic information from historic data on quality from a SPC system, and from systematised engineering knowledge.
The models will be held in an distributed CORBA knowledge-base at customer and supplier sites, and used by a multi-model diagnostic system with a "blackboard" architecture allowing:
- remote machine monitoring;
- support in identification of the reasons of quality problems;
- local and remote automatic and manual malfunction diagnosis;
- set-up support;
- repair support;
- on-line technical information support, and;
- remote maintenance.
Three products prototypes: INTELL-DIAG DIAGNOSIS:a knowledge based tool to detect the mal-functioning and problems in the complex manufacturing processes and identify the reasons of the problems. The main features of the tool are:
-Diagnosis based on manually constructed models, derived from expert knowledge of plant operators and records of previous errors discovered while running the plant A&R equipment;
-Updateable knowledge base, taking information from both, interactive sessions with users to maintain heuristic knowledge, and updates of event occurrences during tool operation;
-Run time information. INTELL-DIAG will use run time information (e.g. sensor readings from the PLC) and will therefore decide on the basis of actual data;
-Remote diagnosis. INTELL-DIAG will enable engineers to log in over a telephone or ISDN line and obtain information in-depth (e.g. sensor readings and PLC status) about an existing problem.
INTELL-DIAG MAINTENANCE: a knowledge based tool serving as a maintenance advisor. It is a remotely configurable software tool which is based on the monitoring of machine performance and process parameters (part of the INTELL-DIAG Diagnosis tool) to advise on required maintenance. It is developed following the extraction of diagnostic models and the building of the fault recognition and model-based diagnostic engines. The maintenance adviser is integrated into remote user interfaces, the use of which requires security and authentication software or a private and secure Intranet. The INTELL-DIAG Maintenance Advisor is activated when diagnosis is complete, to advise on required maintenance actions, as well as providing historical data on previous occurrences. It also advises operators on regular maintenance work required, as a preventative measure before errors occur. The maintenance advisor uses an extension to the INTELL-DIAG Knowledgebase to record all maintenance actions taken and information on the components of machines on the production line and what is required to fix/replace them.
INTELL-DIAG Services/Methodology contains a vast amount of knowledge about methods and rules for structuring of models from a new plant, integration rules for set-up of INTELL-DIAG Tool (e.g. machine interfaces, methods for set-up of reasoning methods and their restrictions). The methodology acts as the guideline of the services which can be applied to further plants to accelerate the integration of reasoning tools immensely so that set-up of infrastructure, interfaces, reasoning tools and knowledge repositories won't waste too much time and will improve the acceptance of users by tailored and sophisticated tools and data bases.
Funding SchemeCSC - Cost-sharing contracts
2220 CB Katwijk