Experimental monitoring of vibrations and the problem of amplitude quantification

Abstract

Experimental monitoring of blade vibration in turbomachinery is typically based on blade-mounted strain gauges. Their signals are used to derive vibration amplitudes which are compared to previously determined modal scope limits, including a safety factor. According to industrial guidelines, this factor is chosen conservatively to ensure safe operation of the machine.For the experimental campaign with the open test case fan ECL5, which is representative for modern lightweight UHBR architectures, it is planned to conduct measurements close to the stability limit. These investigations require a close approach to the limit and hence demand for accurate quantification of vibration amplitudes to ensure secure operation without exhaustive safety margins. It is required that the surveillance is possible in real time and not only in post-processing.Historically, short-time Fourier transformations of vibration sensors are used, but the complex nature of coupled phenomena near the stability limit has an influence on the amplitude accuracy, depending on evaluation parameters. This was demonstrated in a previous study using fast response wall pressure transducers. The present study investigates the influence on blade vibration data of a modern composite material transonic fan. Different methods are compared, sensitivity to evaluation parameters is analyzed and guidelines are given for fast and robust surveillance of critical vibration modes.