Rotor aerodynamic noise is the main noise source of modern helicopters, which not only affects the physical and mental health and efficiency of the occupants, but also causes the side effects of acoustic vibration coupling and acoustic fatigue, which affects the reliability and safety of equipment. Therefore, studying the mechanism and effect of rotor aerodynamic noise, accurately predicting the intensity and characteristics of aerodynamic noise, and effectively reducing aerodynamic noise are important research topics in the development of helicopters. Beamforming is a sound field visualization technique based on spatial filtering. It is the most commonly used, reliable and effective acoustic testing technology for the modern aircraft acoustic test in the wind tunnel and the outfield, which has become a conventional method for the research of helicopter rotor noise in Europe and America, and is also a research hotspot.

 

Aiming at the shortcomings and limitations of the existing acoustic imaging technology for rotor aerodynamic noise, this dissertation studies the acoustic imaging method for moving sources based on beamforming, and the wind tunnel test technology of rotor aerodynamic noise. Aiming at the environmental limitation of non-uniform jet wind tunnel, a correction algorithm for shear layer is put forward. Aiming at the problem of rotor unsteady source, a beamforming algorithm for unsteady acoustic source is proposed. To achieve superresolution imaging of rotor aerodynamic noise, a super-resolution imaging algorithm for rotating source is proposed based on the combination of moving beamforming algorithm and super-resolution acoustic imaging algorithm. On this basis, the super-resolution imaging of wide-band moving source and the separation algorithm of sources in multiple motion modes are further studied.

 

This work is supported by the 12th Five-Year advance research project and the National Natural Science Foundation of China (No. 11574212). The main research contents of the dissertation are summarized as follows:

 

(1) The theory of the static and moving beamforming are derived. According to the ways of de-Dopplerization, the moving beamforming method is divided into time-domain method and frequency-domain method. In order to verify the effectiveness of the method, this dissertation takes rotating source as the research object, and investigates the applicable range and imaging ability of the moving beamforming method through multiple parameter simulation. In a semi anechoic room, the moving beamforming method is performed on a controllable rotating source testbed by a 56-channel microphone array. The results verify the feasibility and the effectiveness of the method.

 

(2) Further, the moving beamforming method is applied to the aerodynamic noise measurement of a model rotor in the wind tunnel. Aiming at the problem of refraction of the shear layer and the unsteady source, this dissertation puts forward the correction method of shear layer and the beamforming method for unsteady source based on time-varying frequency and source strength compensation. On this basis, the noise measurements of helicopter model rotors are carried out in the acoustic wind tunnel of China Aerodynamics Research and Development Center. A large microphone array with 139 channels and a diameter of 7.5m is used to measure the noise and to locate the noise sources from the model rotors, which verifies the ability of the moving beamforming method for locating high-speed moving sources.

 

(3) To achieve super-resolution acoustic imaging of rotor aerodynamic noise sources, the time-domain de-Dopplerization method is used to construct beamforming output and point spread function, so that the super-resolution imaging algorithm for static source can be extended to the super-resolution imaging method for moving source. In addition, this dissertation presents a deconvolution method of wide-band moving source, to solve the problem of energy leakage in the moving beamforming method. The algorithm considers the whole frequency source at the same time, and approximately linearly expresses the coupling effect between discrete frequencies. By considering the full-band of the sources, the coupling effects between discrete frequencies are approximately linearly expressed. An extended linear system of equations is constructed and solved by deconvolution algorithm, and the spatial energy distributions of the sources on all discrete frequencies are obtained. The effectiveness of the algorithm is verified by numerical simulation and experimental research on rotating sources.

 

(4) To reduce the interference of the other mechanical noise on the rotor aerodynamic noise imaging, this dissertation proposes a hybrid deconvolution method to separate static and moving sources. The algorithm extends the point spread function to the cross-point spread function, allocates static and moving equivalent sources in the corresponding source regions, establishes linear system of equations which are expressed by the linear superposition of cross point spread functions according to the equivalence between the equivalent sources and the actual sources. The spatial energy istributions of static sources and moving sources are obtained by solving the equations. The algorithm can be applied to the separation of the aerodynamic noise of the rotor and the mechanical noise, such as the power system and the transmission system. This dissertation further expands the above method and proposes a source separation algorithm for sources in multiple motion modes. The method can be applied to the acoustic imaging and sound source separation of coaxial rotors, tandem rotors and multiple rotors. The effectiveness of the algorithm is verified by numerical simulations and experiments on combined sound sources with rotating sources and a static source.