Using Supervised Learning Techniques for Diagnosis of Dynamic Systems

Abstract

This paper describes an approach based on supervised learning techniques for the diagnosis of dynamic systems. The methodology can start with real system data or with a model of the dynamic system. In the second case, a set of simulations of the system is required to obtain the necessary data. In both cases, obtained data will be labelled according to the running conditions of the system at the gathering data time. Label indicates the running state of system: correct working or abnormal functioning of any system component. After being labelled, data will be treated to add additional information about the running of system. The final goal is to obtain a set of decision rules by applying a classification tool to the set of labelled and treated data. This way, any observation on the system will be classified according to those decision rules, having a return label indicating the currently running state of system. Returned label will be the diagnostic. This entire learning task is carried out off-line, before the diagnosing.