Abstract
This paper investigates the adaptive order tracking techniques, based on the RLS, the extended RLS, and the Kalman recursive algorithms, which all could be realized as real-time applications in the diagnosis of defects in rotating machinery. The numerical implementations of the considered methods through simulations on a representative noisy synthetic signal are performed. The results illustrate a possible degradation in the tracking performance of the RLS algorithm, and the effectiveness of the extended RLS, as well as Kalman based algorithms, for order tracking and distinguishing. Also, one example of practical implementation of the considered adaptive order-tracking methods is given. The analysis of a real technical system – a twisting machine for the polypropylene yarn production, justifies the implementation of the extended RLS algorithm, which is not used so far as order tracking tool for industrial purposes.
