Research On Time Domain Prediction And Identification Technology Of Aero-acoustic Noise From Rotating Blades

Most of the widely used rotating blades produce aerodynamic noise during work,and the mechanisms of noise generation are very complicated.Further,the differences of blade structure also make the characteristics in noise generation become more complex.Hence,it is necessary to develop an efficient method to predict aero-acoustic sound and to experimentally identify noise sources of rotating blades.To meet the fast prediction demand of various types of blade,we have conducted the in-depth research on extracting noise source from turbulent flows,obtaining modified Green functions for complex geometry,increasing computing speed of large-scale sound field and identifying of rotating sound sources in the paper.The contributions of this thesis are summarized as follows:(1)It is well known that the existing empirical and semi-empirical formulas for aero-acoustic sound evaluation are of low accuracy and poor adaptability.The complete calculations for both flow field and sound field are required to achieve a more accurate prediction,which are usually very difficult and time-consuming.To establish a simplified prediction model for thickness noise and steady-load noise,the orders of magnitudes of acoustic pressure and power radiated by monopole,dipole and quadrupole sources are determined according to the FWH equation and Farassat formulas.Then the vortex sound equation is used to compute the sound source on surfaces.And the total sound field is considered to be produced by the real point sources and image sources inside the solid surface,taking the effects of solid surface into account.Finally,the effectiveness and accuracy of the proposed method are verified by comparing the numerical results obtained and the experimental data.(2)In view of the CFD techniques are time-consuming to obtain high-frequency flow field quantities for aero-acoustics prediction.The discrete vortex method(DVM)couple with vortex sound equation is employed to determine the noise source from a turbulent airflow.In the DVM,spacings between vortex particles are considered according to desired accuracy of induced velocity calaulation on surface,taking account of its curvature.To solve the sound field around an airfoil,one need to relate the existing fundamental solution of circular cylinder to the airfoil shape using conformal mapping.A novel method is then developed with the two-dimensional vortex method used for simulations of separated flows around an airfoil.And the vortex sound equation is adopted to model sound generation by interaction of vortices with the surfaces.The far-field noise characteristics are then analyzed and compared with the experimental data.(3)The splitting method involves some difficulties is practice and cannot meet the requirements of engineering applications due to large amount of computer calculations.To overcome the difficulties encountered,the vortex method(VM)couple with vortex sound equation is employed to determine the source of aero-acoustic noise from the turbulent flow around rotating blades.In the VM,considering that the vortex sheet computation is prohibitively expensive for problems with complex geometries,the adaptive fast multipole boundary element method(BEM)is introduced to determine the strengths of vortex sheets on solid surfaces.A simplified viscous diffusion method is also developed based on the numerical solution of Blasius’ equation.To avoid boundaries crossing redistribution lattices,the overset grid methodology is used to perform high-fidelity vortex particles redistributions.In the acoustic part,the sound field scattered by solid bodies is computed using a time-domain boundary element method combined with the convolution quadrature,by means of which the convolution integral is approximated by a quadrature formula utilizing a Laplace domain fundamental solution.In addition,the fast multipole method(FMM)is applied to improve the computational efficiency.Finally,the accuracy and efficiency of the present method are confirmed by comparing with the experimental data.(4)Obviously,well-designed beamformer array geometries are the possibility of providing high-resolution source maps on source surface.Hence,a sidelobe-minimizing method is presented that efficiently determines the optimal weights given a specified beam width for wideband planar arrays of elements scanned to broadside.Considering that the Doppler shift of frequency produced is dependent on source’s motion relative to the receiver,a passive localization method and a mathematical model are applied to estimate the motion parameters of moving sound source by calculating the wide-band correlative ambiguity function based on the Mellin transformation.Then the time dependent delays are calculated via an advanced time approach,in which the time at receivers are calculated from emission time.Moreover,the Generalized Cross Correlation(GCC)method is used to localize rotating source in time domain.As for frequency domain beamforming,the Green function of moving point source under free space condition is expressed in spherical coordinate.Consequently,the rotating beamforming technique with motion compensation is applied to localize rotating broadband sources.Simulated and experimental results show that the developed method can effectively track sound sources on the surface of rotating blades.This paper systematically studies the numerical methods for fast calculation of aerodynamic noise produced by rotating blades,and the localization and identification of moving sound sources.The research results are conducive to the in-depth understanding of noise generation mechanisms from rotating blades and to the optimization of low noise blades.