From a mathematical viewpoint, aeroengine fan noise in a nacelle is superposition of a series of azimuthal and radial modes. Measurements of these acoustic modes are of great importance for experimental investigations of fan noise. In this thesis, a novel mode detection method based on compressive sensing is proposed for aeroengine fan noise testing.

Mode detection, also known as modal decomposition, is a commonly-used testing method that enables researchers to: (i) identify the noise source and reveal the noise generation mechanism; (ii) understand the characteristics of noise radiation or scattering; (iii) evaluate certain noise reduction concepts; and (iv) provide a database for validation of analytical models or numerical codes. Conventional methods for both azimuthal mode a...[ Read more ]

From a mathematical viewpoint, aeroengine fan noise in a nacelle is superposition of a series of azimuthal and radial modes. Measurements of these acoustic modes are of great importance for experimental investigations of fan noise. In this thesis, a novel mode detection method based on compressive sensing is proposed for aeroengine fan noise testing.

Mode detection, also known as modal decomposition, is a commonly-used testing method that enables researchers to: (i) identify the noise source and reveal the noise generation mechanism; (ii) understand the characteristics of noise radiation or scattering; (iii) evaluate certain noise reduction concepts; and (iv) provide a database for validation of analytical models or numerical codes. Conventional methods for both azimuthal mode and radial mode detection are based on the Shannon-Nyquist sampling theorem, thus require a large amount of sensors for high order modes detection in practical tests, leading to a complicated and costly testing system. But at certain blade passing frequencies, there exists only several primary modes in the aeroengine duct. In other words, the acoustic modes are sparse, which allows the adoption of the compressive sensing. This method is able to surpass the limitation imposed by the Nyquist criterion, and reconstructs the mode amplitudes with the number of sensors much fewer than those required by the conventional methods.

In this thesis, the compressive sensing theory is incorporated into the mode detection. To validate the proposed method in the presence of mean flow, a duct acoustic test rig is designed for the anechoic wind tunnel, in which the fan noise is emulated by a mode synthesizer. First, the azimuthal mode detection method based on the compressive sensing is developed. The attention is primarily focused on the examination of the associated reconstruction accuracy and the probability of success with severe background noise interference. The compressive sensing method is then extended to radial mode detection. A couple of new radial mode analysis strategies are proposed for in-duct tonal sound fields. The first method uses a compressive-sensing-based azimuthal mode detection and an inverse-method-based radial mode decomposition. The second method employs the compressive sensing directly from an in-duct sound propagation model. The analysis shows that a reasonably good accuracy can be achieved with a significantly reduced number of sensors by using the proposed compressive-sensing-based methods. In addition, the second approach is more advantageous in terms of the performance and flexibility in arrangement of the sensors, thus is experimentally demonstrated.

Finally, a potential application of the compressive-sensing-based mode detection method is proposed for aeroengine fan health monitoring. To enable such a concept, the array design strategy and optimization method are studied, by maximizing the incoherence of the so-called sensing matrix. Then, the idea of the fan noise monitoring is conceptually demonstrated in wind tunnel tests, by taking account of possible accidental scenarios with foreign body intrusions. Under such circumstance, remarkable changes appear in the azimuthal mode spectrum from the fan noise. We demonstrate that the fan noise variation can be successfully detected by the compressive sensing method just with six sensors. In this way, the foreign body intrusion can be diagnosed through the combination of the compressive sensing and mode detection. Overall, the work in this thesis clearly shows the potential capability of the proposed new fan mode detection method for aeroengine aeroacoustics tests.

Keywords Aeroengine fan noise; mode detection; compressive sensing; health monitoring