EVENT RECOGNITION VIA LINEAR STATE AND PARAMETER MODEL

Abstract

A new method based on linearization of the system equations has been proposed to recog- nize malfunctions, which can be associated with system parameters in dynamical systems. The main features of this method are the short computation time, and the ability to filter out those events caused by disturbances. In addition, the employed linear state and pa- rameter model makes possible to compute the solution with a recursive algorithm, which is fast enough for on-line application, consequently, one can avoid false malfunction recog- nition due to operational dynamic transients. To illustrate the method, recognition of malfunctions in the operation of a gravita- tional water tower is presented.