High-speed Train Interior Full-spectrum Noise Prediction And Sound Quality Optimization Under Multi-physical-field Coupling Excitations

Statistical acoustic energy flow (SAEF) method and sound quality objective-evaluation modified algorithm were proposed, with the combination of which the interior full-spectrum noise, loudness, sharpness and roughness could be accurately predicted. In order to validate the effectiveness of the method and the reliability of the predicted results, four aspects of work were required:First of all, accurate multi-physical-field coupling excitations were extracted. When a train made uniform linear motion in open spaces, the excitation sources inducing interior vi-bro-acoustic response were mainly composed of sencondary suspension forces, aerodynamic noise, wheel/railway noise and equipment noise. These sources were formed by different physi-cal effects, of which the sencondary suspension force, aerodynamic noise and wheel/railway noise were mechanical force, fluid-borne noise and structurally radiated noise, respectively. They could be predicted by rigid multi-body dynamics (MBD), large eddy simulation-indirect boundary element analysis (LES-IBEA) and fast multipole boundary element analysis (FMBEA). And the results in 50～4000 Hz were directly or indirectly validated by experiment. The equipment noise could not be accurately simulated due to complex noise sources. Thus, it was obtained by experiment, which was found to be an approximate reverberant sound field.Second, the interior full-spectrum noise was predicted under multi-physical-field coupling excitations. Finite element analysis-boundary element analysis (FEA-BEA), hybrid finite ele-ment analysis-statistical energy analysis (FEA-SEA), and statistical energy analysis (SEA) were employed to calculate the low frequency, medium frequency, and high frequency interior noise, respectively. Deviations between the simulated and experimental sound pressure levels were generally controlled under 3 dB in each frequency band of interest, which fulfilled engineering permission. Note that the devision principle of the low, medium and high frequency band was proposed, which could be used to select appropriate method for interior acoustic analysis of high-speed trains. Besides, the FE model of the fully trimmed carbody was constructed, includ-ing body-in-white, interior trim parts and traction system. The accuracy of the model was vali-dated by modal experiment of the fully trimmed carbody. The simulated and experimental global modal shapes were highly consistent with each other, and the modal frequency deviations were under 10%. In addition, the medium and high frequency acoustic models were constructed on the basis of the FE fully trimmed carbody, in order to enhance the interior noise prediction ac-curacy in the whole frequency range of interest.Third, for avoiding the disadvantages of above acoustic methods (such as limitation of the analytical frequency range), statistical acoustic energy flow (SAEF) method was proposed and applied in interior full-spectrum noise prediction. SAEF method only concerned the process of the acoustic energy of the excitation sources transmitted into the interior acoustic cavities through the sound insulation attributes of the carbody panels, which fouced on the acoustic en-ergy flow from exterior to interior of the carbody. The SAEF model consisted of 4 parts, includ-ing the SEA subsystems of the exterior acoustic cavities, body-in-white structure, interior acous-tic cavities, and surface or line connections between subsystems. Among them, the exterior acoustic cavity subsystems were used to exert the multi-physical-field coupling acoustic excita-tions as the source of the acoustic energy flow; the surface connections between the body-in-white subsystems and the interior acoustic cavity subsystems were defined with the sould insulation attributes of the panels as the acoustic energy transfer paths; the interior acous-tic cavity subsystems were used to extract the interior acoustic response. Thus, the interior full-spectrum noise under multi-physical-field coupling excitations were calculated and validat-ed by on-line experiment.At last, the modified algorithm of core parameters of sound quality objective evaluation (including loudness, sharpness and roughness) were proposed, considering the acoustic transfer functions of external ear, middle ear and inner ear to enhance the algorithm integrity and accu-racy and apply in high-speed train interior sound quality prediction. The weighting coefficints of the sound quality mathematical model were analyzed in detail, which resulted in the proposal of optimization procedure to effectively reduce the loudness, sharpness and roughness. The interior noise and sound quality optimization was conducted by strategy of active optimization supple-mented by passive optimization. The sound pressure level of the interior center was reduced by 1.2 dB(A), while the loudness, sharpness and roughness were decreased by 11.2%,6.8% and 4.9%, respectively. The optimization could be perceived by auditory system.