Research On Sound Field Prediction Of Underwater Structures Based On Optimal Vibration Monitoring Management

There always exist some low-frequency line spectrums in the radiated sound of underwater vehicles,which are generated by the excitations of internal mechanical equipment inside the vehicles.They attenuate quite slowly with the increase of distance and are difficult to be controlled with conventional methods.Thus,it is important to carry out relevant researches on the sound prediction of underwater structures to provide prior information for active noise control.With the development of sound prediction technology in recent years,to develop a algorithm with high speed is becoming a goal for many scholars,which mainly depends on the calculation efficiency of the algorithm and the design of the structural monitoring system.In particular,the design of the structural vibration monitoring system influences the efficiency of the sound prediction,which has extraordinary research significance.This paper accomplishes the sound prediction of underwater structures with several classical numerical methods and gives out the applicable scope of different algorithms.To break through the limitation of the Nyquist sampling law when designing the vibration monitoring system,it explores the feasibility of introducing the compressed sensing(CS)algorithm into sound prediction to reduce the scale of the system.Based on the equivalent source method(ESM),this paper proposes an improved method called sparse vibration measurement equivalent source method(SVM-ESM).Results of the numerical simulations validate the efficiency of SVM-ESM and indicate that the new method could reduce the scale of vibration monitoring system effectively.In addition,in order to better apply this method for further sound prediction,some factors affecting the effect of this method are also analyzed.Finally,relevant research results are applied to the sound prediction of structures in non-free sound field to validate the algorithms and the influencing factors are also analyzed.The details of the paper are designed as follows:This paper starts with several classical numerical integration sound prediction algorithms for structural sources that are currently widely used.Then,it briefly explains the basic principles of these algorithms,summarizes the advantages and disadvantages of each method through classification and comparison and gives their respective scopes of application.The element radiation superposition method(ERSM)could be applied in some certain conditions because of its high calculation efficiency.However,it needs to satisfy requirement of the far-field approximation.Compared with ERSM,ESM based on monopole equivalent source form is chosen as the main research object since it has better applicability to arbitrary position in the sound field and more complex sound field conditions.Secondly,based on the uniform-design tactic,this paper studies the scale requirement of the structural vibration monitoring system in ESM and ERSM.The results of numerical simulations show that the scale of the monitoring systems is mainly limited by the vibration modes.Thus,the mornitoring system could be managed according to the Nyquist sampling law In addition,in the application of ESM,the configurations of the equivalent source and the vibration monitoring system have significant influences on the prediction effect.Therefore,a few points in the sound field could be regarded as the reference points to match the prediction result for optimizing the equivalent source locations.On this basis,several relevant factors affecting the effect of ESM are analyzed for better applying it for the sound prediction of underwater structures.Subsequently,on the basis of ESM,a set of sparse orthogonal basis derived from the eigenvalue decomposition of the acoustic radiation resistance matrix is used to expand the equivalent source intensity.The process enhances the sparsity of the expand coefficient vector,so the CS theory could be introduced into ESM to reduce the scale of the vibration monitoring system.Compared with ESM,SVM-ESM could break through the limitation of Nyquist sampling law,which means that it could accomplish the sound prediction with a far smaller scale of vibration monitoring system.Several factors which might affect the efficiency of the method are also studied.The results indicate that some conclusions obtained in the study of ESM could also be applied for SVM-ESM.Finally,the sound prediction methods which could be used in non-free sound fields are studied.According to the principle of the image method and ESM,this paper analyzes the composition of the vibration on the surfaces of elastic structures when there are influencing factors such as reflected sound and scattering sound derived from the boundaries.Afterwards,the paper figures out that when the reflected sound has a negligible effect on the vibration,the image method could be used to modify ERSM to make it suitable for the sound prediction of structures in semi-free sound fields in a certain extent.However,it is difficult to be applied for more complicated sound fields and has obvious error near the source.In contrast,ESM is not only applicable for the semi-free field sound field,but can also be combined with either the ray theory or the normal mode method to complete the sound prediction in more complex sound fields,such as waveguides.In particular,according to the impact of reflected sound on the structural vibration,the method of separating the sound field and the fast method of directly ignoring the reflected sound could be used to calculate the equivalent source intensity.In the above two methods,the structural vibration and the source strength are linked with the freefield Green function to improve the computational efficiency and stability.At last,the relevant conclusions about the application of ESM in the non-free sound field could be extended to SVM-ESM,and the feasibility of SVM-ESM used in the non-free field is verified by simulations.Meanwhile,the factors affecting the prediction results are analyzed.The studies offer a new way for the sound prediction of structural sources in complex sound conditions.