Project Leader:	Professor Dr-Ing Horst Rulfs
Research Assistant:	Dipl-Ing J López Cortés
Duration:	01.01.1991-31.03.1994

 

Overall aim of this project was the application of knowledge-based systems for the monitoring and the model-based operative diagnose of ships. Knowledge-based systems enable processing with informatics of available knowlegde on a system also when it is represented in general through results of the most different types. The main focus of the work at TUHH was to secure the knowledge from complex technical systems in ships, structure it to make it applicable to model-based diagnostic expert systems and test it. The analytical decomposition of the ship's technical systems and their subsequent synthesis with the aid of classical system theory forms a particularly strong strategy for developing models for widely differing applications.

Key elements of the method used are the technical system, its components, the main flow diagram, the transformator, the topology, the interface and linkages. Elementary technical components, having a clearly defined function, are considered from the viewpoint of diagnosis as flaw carriers throughout a technical system and form the starting point for modelling. Such components interact with the environment through their main and peripheral interfaces. The grouping of compnents to new component assemblies makes possible to build hierarchical models with a unlimited degree of formal abstraction. In contrast to the main interfaces the peripheral interfaces crisscross the topology and represent interconnections between the interacting systems. On the basis of these structural models it is possible to describe the behaviour and function of components at various operating conditions. The models developed in this form make possible to handle the given complexity of such systems and to perform  complete simulations, to yield model-based system diagnoses.

Because of the substantial structural similarity of such technical systems, by achieving close contact to the praxis, generality and flexibility and by validating the knowledge bases thus developed not only the further application and interoperability of the databank are assured but also their technical and scientific value are improved. This results from the easier applicability of such models also for other systems of the same ship or for modified systems of other ships.