The BRIDGE product is a new generation AI tool-kit which will support the building of industrial-sized fault diagnosis systems. Unlike most products currently being used for supporting such processes, BRIDGE is based on case-based reasoning, which has the simple advantage over many current approaches in that a user does not need to have full knowledge of the process he/she wishes to apply fault diagnosis to.
The BRIDGE product will consist of a PC-Windows based development environment, allowing the engineers to specify the fault applications in detail. Thereby a new, simplified declaration for faults, failures, errors, symptoms and actions is introduced, from which a run-time diagnostic system can be generated automatically. The run-time diagnostic system will allow for operation in real-time, i.e. with fast, predictable and guaranteed response times, due to a novel approach to real-time expert systems. As the generated diagnostic system will be delivered in C++, implementation on almost all platforms becomes feasible.
The PC-based development environment allows for:
- rapid prototyping of the case-base during system development, and refinement during testing and operational life,
- modular extension for new technical systems and newly detected faults,
- guaranteed consistency in the exchange of information between diagnostic systems, and
- identification of necessary modifications to failure detection systems or additional tests.
During operation, when new, previously unknown faults occur, the system can be extended both on-line with temporal addition of the newly uncovered fault cases and off-line, after verification and confirmation by experts in the field, in a new version of the run-time system generated from the fault declarations. In this way, an increasing coverage factor of the diagnostic system results, due to direct incorporation of operational experience.
The overall efficiency of the technical application increases because of:
- knowledge on experienced and anticipated problems is made directly accessible throughout companies,
- strong reduction in efforts for development and maintenance of diagnosis facilities,
- unambiguous exchange of information on problems,
- reduction in raw material consumption,
- existing control and manufacturing systems are enhanced, allowing appropriate reactions to occurring problems to be taken in time.
Therefore, the impact on the industry is potentially very large.
Technical details and Innovations
Publications