Theoretical And Experimental Researches On The Inverse Calculation Of Acoustic Fields In Cabins

Designing the theoretical or numerical model of the sound field in cabins and reversely determining the sound source information or sound field information through the response is the essence of the inverse calculation of sound fields in cabins.The inverse calculation of sound fields in cabins is one key technology for reconstructing the sound fields in cabins.Accurate sound field reconstruction can provide important research value for the prediction and design of cabin noise,evaluation of the sound quality in cabin and sound field reproduction in cabin related to active noise control.In free field,the sound field reconstruction can be easily implemented according to the wave theory of free field.However,for the the internal sound in cabins,the main sound sources include direct sound source body,radiated sound from plates due to the vibration transmitted from cabin power equipment,aerodynamic noise caused by the propeller,and the aerodynamic noise caused by frictions between the fuselage and the impact of the air,etc.Therefore,great challenges are brought by these reflective waves to the actual sound field reconstruction in cabins.The main difficulty is using which method to simulate the original sound field with the same quantization property as the reconstructed sound field by insufficient informations and relate the sound pressure and the property by mathematical and acoustical modeling.In this paper,inverse calculations of sound field are studied in both aspects of sound source reconstruction and sound field reconstruction in cabins.The main researches are listed as follows:1.The sparsity of the spatial distribution of sound source is introduced into the sound source reconstruction problem for non-reflective environments in the research of source inverse calculation.The image source method is introduced into the sound source reconstruction in enclosed cabins to model the reflective waves.Bayesian compressed sensing which can automatically select the regularizedparameter is introduced in the inverse computing of reconstructing the sound source to avoid the tiring work of tuning the regularization parameters.Finally,the sound source reconstructions by the image source model are verified by experiments in an anechoic environment with a reflective plate,a reverberation room and one cabin.2.The sound field reconstruction in cabin is modeled based on the plane wave decomposition,and the reconstruction is realized based on the sparse representation of multi-order plane waves.It can be seen that the higher the frequency is,the smaller the correlation of each plane wave with each other is.The numerical calculation shows that the sparse representation of plane waves can achieve a high performance for the reconstruction.However,the performance of reconstructing the sound field by the sparse representation of plane waves gets decreased as the average reflective coefficient of the boundary of the cabin increases.Finally,performances in different conditions for the reconstruction at different frequencies are studied.The results show that increasing the number of microphones and the number of the plane waves and the sound absorption coefficient of the boundary can improve the reconstruction performance.Increasing the number of the plane waves can broaden the bandwidth range of the accurate reconstruction.Due to the complexity of standing wave form in high modal density of high frequencies,the reconstruction method based on plane wave sparse decomposition can be only applied to the reconstruction in low frequencies.3.The theory and numerical simulation of sound field reconstruction are carried out based on the equivalent sound source method.Next,the sound field reconstruction based on single-layer SESM and multi-layer MESM are proposed respectively using the sparsity of the equivalent sources.The performances of the MESM and SESM are validated through the experiments in rectangular room and cabin structure.Finally,numerical simulations show that SESM and MESM can achieve lower reconstructed errors than those by ESM and with more computing time for MESM.If the number of equivalent sources and layers are changed,SESM and MESM show no significant differences for the reconstructed error.However,smaller reconstruction errors can still be obtained while the number of microphones is decreased.In addition,increasing the reflective coefficient will lead an inferior performance for SESM and MESM.4.The sound field reproduction by loudspeakers in cabins is studied by inverse acoustical calculation.It is assumed that the channel between the reproducing loudspeaker and the receiving position is a linear time invariant system which can be used to is derive the deconvolution theory in frequency domain and time domain for the sound field reproduction.Then the deconvolution method in frequency domain is verified through numerical simulations.Finally,the sound field reproduction by loudspeakers in cabin structure is carried out to verify the accuracy of the deconvolution method in time domain with 3d B for the average sound pressure level difference of each receiving position.