Model-based fault detection and diagnosis in dynamic systems

Description of impact

A system is considered as faulty when there is a malfunction. This abnormal behaviour may have consequences ranging from economic losses (lower quality, lower production, higher costs) to damages to the people or to the environment. Many differnt techniques have been developed in the recent years which intend to detect (to determine the presence of a fault) and diagnose (to locate, to quantify, to characterize) faults. These techniques may be classified in different ways. For instance, there are techniques that are based on a model of the system and there are techniques that are based in other forms to express the knowledge about the system, like heuristics, statistics, automatic learning, neural networks, etc. The reserach group MICE has participated in an European project and in several Spanish projects in which several tools, mainly based on models, have been developed and applied to industrial processes. One of these tools can be highlighted, SQualtrack, which is based on interval models and whose main characteristic is that it does not generate false alarms. New applications to biological processes are on the way.