Fabrication and characterization of a piezo-MEMS uniaxial accelerometer as a tool for the monitoring of combustion instability in gas turbine engines

Abstract

This work is focused on the design, simulation, microfabrication, and characterization of an uniaxial piezoelectric MEMS accelerometer. It investigates the capability of the proposed device to monitor the combustion instability induced by the addition of air into jet-A1 fuel in a combustion chamber. The membrane structure of the accelerometer is modeled using the commercial FEM package COMSOL Multiphysics to optimize the geometrical parameters and validated through both analytical and experimental results. Aluminum nitride (AlN) thin film, deposited by sputtering process and characterized for its nanomechanical properties, was chosen as the active piezoelectric material of the accelerometer thanks to its compatibility with the complementary metal oxide semiconductor (CMOS) technology. The accelerometer has been characterized in terms of frequency response, sensitivity, and linearity. All experimental results, which are in good agreement with simulations, show that the functional characteristics of the accelerometer are as follows: resonance frequency of about 3.8 kHz; linear bandwidth in the range 0.3 – 1.2 kHz; dynamic sensitivity of 0.25 mV/g with a linearity of 99.1 %. Experiments for the monitoring of the combustion instability were conducted on a liquid-fueled swirling combustor with a nominal power of 300 kW at two global equivalence ratios ($\Phi$), i.e. 0.36 and 0.18. The microfabricated accelerometer was positioned on the combustion chamber structure near the flame in the combustion zone for the measurement of the chamber vibrations. Its response was compared with the one of a commercial pressure sensor to assess its reliability, demonstrating good correlation between the signals coming from the two devices. Mean, variance and Kurtosis data analysis techniques were employed to elaborate signals coming from both sensors to provide a quantitative indicator of the combustion instability. Given its high sensitivity and wide linear bandwidth, the accelerometer can detect subtle variations in chamber vibrations, which are critical for early identification of combustion instability. By providing real-time, high-resolution monitoring, the device offers a non-invasive method for detecting and diagnosing instability phenomena that are often difficult to assess with traditional pressure sensors. The results demonstrate the potential of the realized piezoelectric accelerometer in understanding combustion instability mechanisms in gas turbines. Moreover, together with being a promising non-invasive tool of diagnosis of this phenomenon, the proposed solution is also expected to be a challenge for combustion instability prediction, contributing to the development of more stable combustors.