The Improvement of Modulating Function Method for Fast Identification of Parameter Faults in Linear Continuous Systems

Witold Byrski and Jędrzej Byrski


Identification, modulating functions, fault detection, linear continuous systems


In this paper essential modification of some identification method for continuous system is presented. This modification enables significant reduction of identification time needed for the on-line calculation of the new value of parameters after their step changes in continuous and linear systems. In the basic identification method the convolution transformation of both sides of the output equation with special modulating functions (MF) is used. This method indicates very good results for continuous, linear, time invariant system CLTIS with constant parameters. The main drawback of the application of this method for the problem of parameters step changes detection is an existence of time delay for the reconstruction of the new value of parameters. Generally this time is the sum of the MF support h and the width T of the moving identification window. The aim of this paper is presentation of the optimal method for the parameter identification after their fault which makes identification time much shorter.

Important Links:

Go Back