Estudio comparativo de métodos para la obtención de modelos de sistemas continuos a partir de su identificación por computador

Resum

This paper is devoted to determine the continuous model of a system, from its equivalent discrete model. This is assumed that such a dicrete control system has been obtained usins any one of the classical computer identification methods existins nowadays. The determination of the continuous model of a system becomes necessary, specially when fault detection techniques are used, or the design of control system is done. In despite of this necessity, it does not exist at present a general method to calculate such a continuous model. First, a thorough revision is made of the existing algorithms. They are classified in exact and approximate algorithms. Special stress laid on the condition that the continuous system and the sampling rate should fulfil in order to guarantee the validity of these algorithms. Then, a new algorithm is proposed, original of the authors, that allows to use slower samplig rates and that may be used in the identifiaction of every time invariant linear continuous systems. This algorithm is completely general. In order to perform a comparative study, a concrete example is used, that will allow us to obtain conclusions about the validity of each algorithm.