Automated tacholess rotational speed estimation on asynchronous motors for improved vibration analysis

Online vibration monitoring is a widely recognised approach to maximise availability of rotating machines such as induction motor-driven centrifugal pumps. As vibration monitoring relies on the analysis of a well-defined multitude of harmonics of the fundamental frequency (i.e. the motor speed), accurate speed estimation is a prerequisite for continuous vibration monitoring systems without tachometers. While numerous methods have been reported for this estimation, their accuracy for well-balanced induction motor drives running at constant speed is insufficient to accurately establish high-frequency harmonics in the spectrum. This paper presents a novel method that automatically and accurately determines the rotational speed from an individual vibration measurement at constant speed using minimal a priori knowledge of the measured system. The proposed method determines the rotational speed with a root mean square error of 0.10 Hz or a specified maximum error of 0.0165 Hz for 90.1% of cases in the recent NLN-EMP dataset. This dataset contains over 2300 measurements, with a significant variety of healthy and faulty electric motor drive vibration signals, including well-balanced and well-aligned cases, which demonstrates the wide applicability of the proposed method. Finally, a benchmark comparison shows that the proposed method significantly improves upon a comparable method from the literature.