With the increasing complexity of modern machinery, many rotating system problems require analysis with high
order resolution. Today’s order tracking techniques (e.g. frequency domain method, order domain method, etc.)
only provide decent order values when the order content does not change during the observation interval. So to
get the best out of these algorithms, the ideal measurement sequence would consist of a truly constant RPM
during each observation period, also known as a semi-stationary measurement.
Such an approach has several disadvantages: (i) the total measurement time is far higher than in a run-up measurement; (ii) spurious effects occurring only during the transition from one RPM to the next go unnoticed, giving rise to an unacceptable number of semi-stationary levels; (iii) at high RPMs, even the slightest oscillations in RPM result in large phase shifts.
This paper presents a new Adaptive Amplitude Demodulation (AAD) algorithm which allows separation of order
components in run-up measurements where the aforementioned methods fail. First, an overview is given of some of the existing order tracking techniques. Then, the principles of the new AAD algorithm are outlined.
Finally, the results of a benchmark study are presented in which the new algorithm is compared to existing order tracking techniques for (i) slow and fast run-up measurements and for (ii) orders with slowly and rapidly varying amplitude and phase profile.
