Applying Constraint Databases in the Determination of Potential Minimal Conflicts to Polynomial Model-Based Diagnosis

Abstract

Model-based Diagnosis allows the identification of the parts which fail in a system. The models are based on the knowledge of the system to diagnose, and may be represented by constraints associated to the components. The variables of these constraints can be observable or non-observable, depending on the situation of the sensors. In order to obtain the potential minimal diagnosis in a system, an important issue is related to finding out the potential minimal conflicts in an efficient way. We consider that Constraint Databases represent an excellent option in order to solve this problem in complex systems. In this work we have used a novel logical architecture of Constraint Databases which has allowed obtaining these potential conflicts by means of the corresponding queries. Moreover, we have considered Gröbner Bases as a projection operator to obtain the potential minimal conflicts of a system. The first results obtained on this work, which are shown in a heat exchangers example, have been very promising.