Using Description Logics for Consistency-based Diagnosis
Gerd Kamp, Holger Wache · 1996
Using quantitative models of simple mechanisms as an example domain, we show how the basic principles of consistency-based diagnosis can be implemented using description logics with expressive concrete domains. In addition, description logics provide the inference services necessary to organize and validate model libraries, an aspect that is neglected in consistency-based diagnosis until now. 1 Description Logics for Diagnosis Consistency-based diagnosis 1 [7; 10] is a method for the diagnosis of technical devices based on descriptions of the correct and faulty behavior of components. Therefore, for representing, simulating and diagnosing the simple bike drive train in Fig.1 at least the following knowledge must be represented within such a system: 1. The different types of components, e.g. wheels, gearwheels, chains, along with their attributes like forceF, radiusrand torqueM. 2 2. The structure of assemblies, e.g the kinematic structure of the mechanisms using kinematic pairs. 3. Physical laws likeM=rFand constraints imposed on the attributes likeF0andr>0. 4. The normal and faulty behaviors (often called models in consistency-based diagnosis) of components and assemblies, e.g. the propagation of torques from one wheel to another (M1=M2) in a rotational pair. Based on this representation diagnosis is mainly the task of identifying the models of behavior that are consistent with a set of observed parameter values. This is most often done by simulating hypothetized behaviors and comparing the results with the observed values. 1 Also called model-based diagnosis