【作者】 刘加利；

【导师】 张继业；

【作者基本信息】 西南交通大学 ， 载运工具运用工程， 2013， 博士

【摘要】 随着列车运行速度的提高,气动噪声成为高速列车噪声中越来越重要的组成部分。研究高速列车气动噪声的预测方法及控制方法具有重要的意义。针对高速列车气动噪声问题,本文建立高速列车车外及车内气动噪声的计算方法,研究主要气动噪声源部位的远场辐射气动噪声特性。建立高速列车流线型头型的多目标优化设计模型,对高速列车流线型头型进行减阻降噪优化设计。建立低压环境下真空管道高速列车空气动力学计算模型,研究真空管道高速列车的气动阻力和气动噪声源特性。根据高速列车近地面运行的实际情况,分别由FW-H方程和Kirchhoff方法出发,利用半自由空间的Green函数,推导考虑地面效应的高速列车远场声学积分公式,并研究地面声阻抗对高速列车远场气动噪声的影响。由于地面效应的存在,原来的自由声场相当于真实列车声场和镜像列车声场的叠加,且作用于镜像列车上的法向运动速度和力源与真实列车上的相同。考虑到空气介质往往是运动的,由广义Lighthill方程出发,推导考虑介质运动和地面效应的高速列车远场声学积分公式。建立高速列车远场气动噪声计算的气动声学模型方法和混合计算方法,对高速列车车头及受电弓远场气动噪声进行计算分析。气动声学模型方法和混合计算方法得到的远场测点的声压频谱基本相同。高速列车车头远场气动噪声具有宽频特性,主要能量集中在1800Hz～2800Hz。高速列车受电弓远场气动噪声频带较窄,主要能量集中在100～700Hz。高速列车车头和受电弓的远场测点的等效连续A计权声压级近似与列车速度的对数成线性关系。高速列车的头部控制线形状对高速列车远场气动噪声具有较明显的影响,平直纵向剖面线和方形水平剖面线组合下的头型所产生的气动噪声最小,鼓形纵向剖面线和锥形水平剖面线组合下的头型所产生的气动噪声最大。基于计算流体动力学和统计能量分析法建立高速列车车内气动噪声的计算方法。根据统计能量分析法的基本原理,建立高速列车车内气动噪声计算模型,并计算模型中各个子系统的基本参数,即模态密度、内损耗因子和耦合损耗因子。利用大涡模拟方法获得各个车身子系统上的平均脉动压力谱,进而对高速列车车内气动噪声进行计算分析。从线性计权声压级来看,司机室声腔和乘客室声腔的声压具有低频特性。从A计权声压级来看,司机室声腔和乘客室声腔的声压的显著频带范围较宽,司机室声腔的噪声能量主要集中在100Hz～2000Hz,乘客室声腔的噪声能量主要集中在50Hz～2000Hz。司机室声腔和乘客室声腔的线性计权声压级和A计权声压级均与列车速度的对数近似成线性关系。建立高速列车流线型头型的多目标优化设计方法,以气动阻力和偶极子噪声源为优化目标,对高速列车流线型头型进行多目标自动优化设计。利用CATIA软件建立高速列车三维参数化模型,提取出5个优化设计变量,利用自编MATLAB程序和CATScript脚本文件实现高速列车流线型头型的自动变形。采用脚本文件和批处理命令实现高速列车空气动力学计算网格的自动划分及高速列车绕流流场的自动计算,利用多目标遗传算法NSGA-Ⅱ对优化设计变量进行自动更新,实现高速列车流线型头型的多目标自动寻优设计。通过研究优化目标与优化变量之间的相关性,得到影响优化目标的关键优化设计变量,进而得到关键优化设计变量和优化目标之间的非线性关系。通过与原始流线型头型气动性能的对比发现,优化后的流线型头型最大可使高速列车的气动阻力降低4.54%,使高速列车的偶极子噪声源减少4.95dB。为减少多目标优化计算时间,利用径向基神经网络建立高速列车空气动力学计算的近似计算模型,为提高近似计算模型在整个优化设计空间内的近似效果,采用最优拉丁方设计方法获得径向基神经网络的输入和输出。近似计算模型得到的气动阻力的预测值与实际值的误差小于1%,而偶极子噪声源的预测值与实际值的误差小于3dB,近似计算模型具有较好的近似效果,且采用近似计算模型优化计算获得的Parto前沿与采用真实模型优化计算获得的Pareto前沿相差不大。建立低压(0.01atm～0.1atm)环境下真空管道高速列车空气动力学计算的流体模型、数学模型和数值模型,研究管道压力、阻塞比和列车速度对高速列车阻力系数、气动阻力、偶极子噪声源和四极子噪声源的影响,并以明线上运行速度为400km/h的高速列车气动阻力为限值,确定出真空管道高速交通系统的最佳管道压力、阻塞比和列车速度关系。在低压环境下,真空管道中的空气流动可以采用连续介质模型描述。高速列车的气动阻力系数基本上与管道压力和列车速度无关,主要依赖于阻塞比。高速列车的气动阻力与管道压力成线性关系,与列车速度成平方关系,且随着阻塞比的增加而增大。高速列车偶极子噪声源和四极子噪声源均与列车速度的对数近似成线性关系,当列车速度为600km/h时,四极子噪声源较小,偶极子噪声源占据主导地位,随着列车速度的提高,四极子噪声源变得明显,并占据主导地位。为模拟更稀薄环境下的真空管道空气流动特性,建立适用于滑移区和过渡区稀薄气体流动的格子Boltzmann模型,并对二阶速度滑移边界条件进行检验。研究发现,Guo模型、Hisa模型、Zhang模型表现较好。采用Guo模型对滑移区和过渡区的稀薄气体流动进行数值模拟,发现当稀薄参数取1.64时,计算得到的无量纲速度剖面与Karniadakis给出的无量纲速度剖面吻合很好,从而验证了计算模型及计算程序的正确性。更多还原

【Abstract】 With the increasing of the train speed, aeroacoustics of the high-speed train has become an increasingly important part of the high-speed train noise. Study on the computation and control of aeroacoustics of the high-speed train has important significance. For the aeroacoustics issue of the high-speed train, the computational methods of exterior and interior aeroacoustics of the high-speed train were established. The characteristics of far-field radiation aeroacoustics of the main aeroacoustics source parts were analyzed. The multi-objective optimization design model of the streamlined head of the high-speed train was eatablished, and the streamlined head was optimized to reduce the drag force and noise. The aerodynamic computational model of the high-speed train in the vacuum tube in the low-pressure environment was established, the characteristics of aerodynamic drag force and aeroacoustics source of the high-speed train in the vacuum tube were studied.Based on the actual situation of the high-speed train running near the ground, starting by the FW-H equation and Kirchhoff equation respectively, the far-field acoustical integral formula of the high-speed train considering the ground effect was derived respectively using the semi-free-space Green’s function. The effect of ground acoustical impendance on the far-field aeroacoustics of the high-speed train was studied. The former free sound field is equivalent to superposition of sound field of realistic train and sound field of mirror train because of the ground effect. The normal velocities and force sources on the mirror train are the same as that of on the realistic train. Considering that the air medium tends to have movement, the far-field acoustics integral formula of the high-speed train considering the medium movement and the ground effect was derived from the generalized Lighthill equation.The aeroacoustic model method and mixed computational method for the computational of the far-field aeroacoustics of the high-speed train were established. The far-field aeroacoustics of the head and pantograph of the high-speed train was computed and analyzed. The sound pressure spectrum of the far-fied measure point computed by the aeroacoustic model method and mixed computational method is basically the same. The far-field aeroacoustics of the high-spped train head has broadband characteristics and the energy is mainly distributed between1800-2800Hz. The far-field aeroacoustics of the high-spped train pantograph has narrowband characteristics and the energy is mainly distributed between100-700Hz. The equivalent continuous A-weighted sound pressure level of the far-field measure point of the head and pantograph of the high-speed train has a linear relationship with the logarithm of the train speed. The shape of control lines of the train head has obvious effect on the far-field aeroacoustics of the high-speed train. The aerodynamic noise corresponding to the head with the combination of straight longitudinal profile line and the square horizontal section line is the smallest. The aerodynamic noise corresponding to the head with the combination of drum-shaped longitudinal profile line and the cone horizontal section line is the largest.The computational method of interior aeroacoustics of the high-speed train was established based on the computational fliud dynamics and statistical energy analysis. The computational model of interior aeroacoustics of the high-speed train was established based on the theory of statistical energy analysis. The fundamental parameters, i.e., modal density, internal loss factor and coupled loss factor, were determined. The average fluctuating pressure spectrums on the carbody subsystems were obtained using the large eddy simulation, and the interior aeroacoustics of the high-speed train was computed and analyzed. For the liner-weighted sound pressure level, the sound pressure level of the drive’s cab cavity and passenger compartment cavity has low frequency characteristics. For the A-weighted sound pressure level, the sound pressure level of the drive’s cab cavity and passenger compartment cavity has broadband characteristics. The sound energy of the drive’s cab cavity is mainly distributed between100-2000Hz, the sound energy of the passenger compartment cavity is mainly distributed between50-2000Hz.The linear-weighted and A-weighted sound pressure level of the drive’s cab cavity and passenger compartment cavity has a linear relationship with the logarithm of the train speed.The multi-objective optimization design method of the streamlined head of the high-speed train was established. The aerodynamic drag force and dipole noise source were set as optimization objectives, and the multi-objective automatic optimization design of the streamlined head of the high-speed train was carried out. The three-dimensional parametric model of the high-speed train was established using the CATIA software, and five optimization design variables were extracted. The automatically deformation of the streamlined head of the high-speed train was achieved using MATLAB program and CATScript script file. The automatic division of the aerodynamic computational grid of the high-speed train and automatic computation of flow field of the high-speed train were achieved using the script files and batch commands. The automatic update of optimization design variables were achieved using the multi-objective genetic algorithm NSGA-Ⅱ, then the automatic optimization design of the streamlined head of the high-speed train was achieved. From the research on the correlation between optimization objectives and optimization variables, the key optimization design variables which have effects on the optimization objectives could be obtained. Then the non-linear relationship of the key optimization design variables and optimization objectives was obtained. By contrasting with aerodynamic performances of the original streamlined head, the maximum reduction of the aerodynamic drag force of the optimized streamlined head is4.54%, and the maximum reduction of the dipole source of the optimized streamlined head is4.95dB. In order to reduce the computational time of the multi-objective optimization, the approximate computational model of aerodynamic of the high-speed train was established using the radial basic function neural network. In order to get a good approximate in the whole optimization design space, the inputs and outputs of the radial basic function neural network were obtained using the optimal Latin hypercube design. The error between the predicted values computed by the approximate computational model and actual values of the aerodynamic drag foece is less than1%. The error between the predicted values computed by the approximate computational model and actual values of the dipole noise source is less than3dB. The approximate computational model has a good approximate. The Pareto frontier computed by the approximate computational model is basically the same as that of computed by the actual model.The fluid model, mathematical model and numerical model of aerodynamic computation of the high-speed train in the vacuum tube in the low-pressure (0.01atm-0.1atm) environment were established. The effects of tube pressure, blockage ratio and train speed on the drag coefficient, aerodynamic drag force, dipole noise source and quadrupole noise source were studied. Setting the aerodynamic drag foece of a400km/h train in the open field under the standard pressure condition as a limit value, the best relationship between tube pressures, blockage ratios and train speeds for the vacuum tube high-speed transportation system was determined. In the low-pressure environment, the air flow in the evacuated tube can be described using a continuous model. The aerodynamic drag coefficient basically has nothing to do with the tube pressure or the train speed, which mainly depends on the blockage ratio. The aerodynamic drag force of the high-speed train is linear with the tube pressure, and square with the train speed, and also increases with the increase of blockage ratio. The dipole noise source and the quadrupole noise source of the high-speed train are almost linear with the logarithm of the train speed. When the train speed is600km/h, the quadrupole noise source is small and the dipole noise source has a dominate status. With the increase of train speed, the quadrupole noise source becomes apparent and has a dominate status. In order to simulate the air flow characteristics in the vacuum tube in the more rarefied environment, the lattice Boltzmann model for rarefied gas flow in the slip and transitional regime was established, and the second order velocity slip boundary conditions was studied. Guo model, Hisa model and Zhang model have a good perform. The rarefied gas flows in the slip and transitional regime were simulated using Guo model. When the rarefaction parameter is equal to1.64, the computed dimensionless velocity profile is good agreement with the dimensionless velocity profile given by Karniadakis, which verify the correctness of the computational model and computational program.更多还原