No abstract

In this paper, the efficient implementation of a reliable measurement device to measure the angular velocity vibrations of rotating shafts is discussed. The solution is particularly suitable to be fit in measurement embedded systems equipped with field-programmable gate arrays (FPGAs). The proposed measurement method is analyzed from a theoretical point of view, focusing the study on the frequency response of the technique. On the basis of the theoretical results, a reliable and low-complexity design of the measurement device has been proposed. In the adopted solution, the high clock rates of FPGAs are exploited to increase the accuracy of the zero-crossing times estimation, for a given A/D acquisition frequency, using an upsampling approach. The experimental results confirm the validity of the technique. The upsampling method allows an increase of the measurement accuracy and an improvement of about 10 dB for the output signal-to-noise ratio over a wide range of vibration frequencies

This paper describes a selection of Baker Hughes, a GE company (BHGE) activities to support Gas Turbine (GT) design and operation from simple to more elaborate applications of Machine Learning (ML).

The current centrifugal compressor design for the oil & gas market is more and more challenging, and the presence of many competitors is pushing technology towards both a casing size reduction and a rotational speed increase. The first point is leading to an increase in the number of wheels per rotor (to do the same service), and the second point is forcing to cross two or even three rotor modes (hence a higher control of rotor damping is necessary). The two points together are leading to increase the rotor “flexibility ratio” (defined as the ratio between the maximum continuous speed and the first critical speed at infinite support stiffness according to API standard, and finally the rotordynamic stability is very much challenged. The centrifugal compressor's rotordynamic stability is strongly related to the internal seals' dynamic behavior, and for this reason, the authors' company decided several years ago to develop internally a high pressure seal test rig to measure internal seals stiffness and damping. The rig is now in operation, and in a previous paper the authors described its main capabilities, the applied identification procedures, and the preliminary test results captured for a long labyrinth seal (smooth rotor, straight toothed stator) tested up to 200 bar. This paper is intended to show more data for the same long Laby with special focus on some peculiar test as negative preswirl test, single frequency versus multifrequency test, offset versus centered seal test. The negative preswirl test shows a drastic change in the effective damping (from destabilizing to stabilizing) and provides a support in favor of the selection of swirl reversal devices at seals upstream. The multifrequency excitation test approach (based on the concurrent presence of several frequencies not multiples at each other) is compared with a single frequency excitation providing similar results and thus confirming the soundness of the multiple effects linear superimposition assumption. The effect of a static offset (simulating the real position of a rotor inside an annular seal) is also investigated proving that the relevant impact is negligible within the range of eccentricity explored (10% of seal clearance). Moreover, a pocket damper seal (PDS) with the same nominal diameter, clearance, and effective length has been tested (up to 300 bar) and compared with the Laby. As expected, the PDS shows both a higher effective stiffness and damping at the same test conditions, so the promising results already collected in a previous test campaign which was performed on a smaller scale and lower pressure test rig were mostly confirmed with the only exception for the effective damping crossover frequency which was lower than expected.

In this paper, the authors discuss a measurement method, based on the zero-crossing demodulation technique of FM signals, to estimate the angular velocity vibrations of a rotating shaft. The demodulation algorithm is applied without any filtering to the direct voltage output issued by some probes sensing the passages of arbitrarily shaped targets installed on the rotating shaft. The authors discuss a theoretical approach to analyze the measurement problem taking into account the chief nonidealities related to the measurement setup, i.e., they have investigated the effects of both the shaft side vibrations and the irregular shape of the targets. On the basis of the theoretical results, the authors propose a measurement method that can reject the effects of these mentioned nonidealities, exploiting the measurements of two or more probes properly positioned around the shaft