激波是一种以超声速传播的强扰动波，可以通过气体、液体和固体介质传播。在实验研究中，激波管通常被当作一种理想的激波产生装置。然而由于激波管壁面粘性以及热传导等因素的影响，实际激波管中的气体流动特性在一定程度上会偏离理想激波管理论。由于运动气体状态与静止激波管壁面存在差异，激波在激波管中传播时，会在波后诱导出一非定常边界层，激波速度和激波强度逐渐减小，同时接触面速度逐渐增大，发生明显的激波衰减现象。特别地，对于小尺度以及低压激波管来说，这一激波衰减现象会更为显著。由于激波的减速和接触面的加速，激波管内2区或5区的有效实验时间将大幅减小。同时激波的衰减还将引起流场品质的下降，进一步可能导致激波分叉的产生以及“驱动气体污染”等不利影响。因此，针对激波衰减这一物理问题的机理研究，将有助于对激波管实验的数据解读和平台优化，具有现实指导意义。

平面激波在激波管中的衰减是一类由壁面引起的非定常运动，其中涉及激波间断、激波诱导边界层的产生、边界层与激波的相互作用等复杂物理现象。文献中已有很多研究认识到运动激波在激波管或槽道中传播时会发生衰减，但是激波衰减的详细机理仍不清楚，主要原因有：研究手段缺乏，微观分析困难，计算资源需求大等。本文把二维激波衰减问题分成两个步骤研究。首先，对静止气体在无穷大壁面突变下（一维可压缩Rayleigh问题）的流场发展过程进行数值模拟，分析可压缩气体的运动规律及扰动传播特性。其次，采用直接模拟蒙特卡罗（DSMC）方法模拟二维平面激波的衰减过程，研究激波衰减过程的详细机理，并对影响激波衰减程度的影响因素进行讨论。具体内容如下。

1）分别采用DSMC方法、计算流体动力学（CFD）方法两种手段，针对一维可压缩Rayleigh问题从自由分子流到连续流的全流域流场进行了数值模拟，详细分析了壁面速度及温度间断在平板法向上的发展和传播过程。计算结果发现，在平板起动初期（100个气体初始状态分子平均碰撞时间之前），CFD方法均不能给出平板表面参数以及可压缩气体扰动参量分布的准确估计，而在1000个气体初始状态分子平均碰撞时间之后，相较于AUSMpw格式和Van Leer的FVS格式来说，Roe的FDS格式计算结果与DSMC结果吻合得更好，包括平板热流、剪切应力以及平板上方可压缩气体扰动参量分布。进一步的分析发现，气体流场中通常存在以近声速传播的扰动波以及厚度随时间的0.5次方发展的边界层。具体来说，扰动波的传播速度与壁面的温度、速度间断强度密切相关，其大小与气体声速的量级相当，且逐渐过渡到当地声速；扰动波幅值的大小随着时间的推移会发生明显的衰减。

2）采用DSMC方法对二维平面激波沿激波管轴向传播和衰减的物理过程进行数值模拟，分不同阶段描述了壁面边界层扰动与激波间断结构之间的相互作用，对平面激波衰减过程进行了详细的机理分析，并对影响激波衰减程度的影响因素进行讨论，包括激波管尺寸、马赫数的影响等。计算结果发现，与可压缩Rayleigh问题类似，运动激波的下游处，激波管壁面上会发展出一非定常的边界层，同时流场中诱导出一个低压扰动波，该扰动波受到激波管外形的限制会在激波管高度方向上发生来回多次反射逐渐衰减。激波与边界层干扰区附近，气体法向速度冲向壁面，边界层如同一个“漏”的作用，带走了很大一部分的流场质量。同时激波结构向运动激波下游方向发生明显的弯曲，激波轮廓、激波马赫数衰减曲线均呈现周期振荡的特性，振动频率与下游扰动波在激波管内的反射频率一致。最后，数值模拟结果还表明，激波管尺度越小、初始激波马赫数越大，激波衰减越严重。

Shock wave is a strong disturbance wave propagating at a supersonic speed, which can propagate through gas, liquid and solid media. Shock waves are usually generated using shock tubes for various experimental studies and are treated as ideal shock waves. However, due to viscous effects and heat conduction of the shock tube wall, the gas flow properties in real shock tube will deviate from the ideal shock tube theory. Since the state of the moving gas is different from the static shock tube wall, an unsteady boundary layer will be induced behind the propagating shock wave, thus the shock wave decelerates gradually and severe shock attenuation occurs in the long time. In particular, this attenuation behavior will be greatly enhanced for smaller scale or low-pressure shock tubes. For the shock tube flow, due to the acceleration of the contact surface and the deceleration of the shock wave, the effective experiment time of zone 2 or 5 in the shock tube will be significantly reduced, and shock attenuation will also greatly decrease the quality of the flow field, which may further lead to shock wave bifurcation, "driving gas contamination" and other adverse effects. Therefore, it is of important to study the physical mechanism of shock wave attenuation, which will in turn help data interpretation and platform optimization of shock tube experiments.

Planar shock wave attenuation is a kind of unsteady motion induced by the static shock tube wall, it involves coupled processes with shock wave discontinuity, unsteady boundary layer, and interaction between shock wave and this induced boundary layer. Although many previous investigations have pointed out that shock wave constantly attenuates when propagating in shock tubes or small channels, the detailed mechanism of this attenuation process has not been clearly discussed yet in the literature, mainly due to lack of research methods, difficulties in microscopic analysis, and large demand for computing resources. This paper divides the study of two-dimensional shock wave attenuation into two steps. First, the flow field evolution of static gas motion above an infinite wall of sudden state changes (that is, one-dimensional compressible Rayleigh problem) is studied through numerical simulations, motion laws and disturbance propagation characteristics of the above compressible gas are investigated. Second, two-dimensional planar shock wave attenuation problem is numerically studied through DSMC method, and detailed mechanism of this attenuation process is investigated, impact factors of shock attenuation degree are discussed as well. The details are as follows.

1）One-dimensional compressible Rayleigh problem is studied through the direct simulation Monte Carlo (DSMC) method and the computational fluid dynamics (CFD) method respectively. Numerical simulations cover the whole flow regime range from free molecular flow to continuous flow regime. Detailed analysis is provided for the development and propagation process of the initial velocity and temperature discontinuity in the normal direction of the wall. Numerical results show that, in the initial stage of wall motion startup (before 10 mean collision times of the undisturbed gas molecules), the CFD method could not give an accurate estimate of both wall surface parameters and disturbance distribution of compressible gas above the wall. In 1000 mean collision times of the undisturbed gas, parameters like heat flux, shear stress on the surface and flow field distribution of the above compressible gas calculated by Roe's FDS scheme are in better agreement with the DSMC results, when compared with the AUSMpw scheme and Van Leer's FVS scheme. Further analysis also suggests that the induced disturbance wave propagates approximately with the velocity magnitude equivalent to the sound velocity, whereas the thickness of the unsteady boundary layer increases nearly proportional to the square root of time. Specifically, this propagation velocity of disturbance wave is closely related to the intensity of initial temperature and velocity discontinuity between the wall and gas, and it will gradually transform to the local sound velocity in the long time; the amplitude of disturbance wave also severely attenuates as time goes by.

2）The DSMC method is employed to study the propagation and attenuation of shock wave in a two-dimensional planar tube. Detailed mechanism analysis is provided for the whole process of shock attenuation, the interaction between wall boundary layer disturbance and shock wave discontinuous structure is described as several different development stages. Impact factors of shock attenuation degree are discussed as well, including shock tube heights and the initial shock Mach numbers. Numerical results show that, similar to the compressible Rayleigh problem, an unsteady boundary layer will develop at the shock tube wall in the downstream of the moving shock wave, and a low-pressure disturbance wave will be induced in the flow field as well, which could be reflected multiple times in the height direction of the shock tube due to the shape limitation and then will be gradually attenuated with time. In the vicinity of the shock wave and boundary layer interaction zone, the normal velocity of the gas rushes to the shock tube wall, and the boundary layer acts as a "leak" taking away a large part of the flow field mass. At the same time, appreciable shock curvature is clearly seen in the downstream direction. Both the shock wave profile and Mach number attenuation curve show periodic oscillation characteristics, and the oscillation frequency is consistent with the reflection frequency of the downstream disturbance wave in the shock tube. Finally, shock wave is found to be more severely attenuated in narrower tube or having larger initial Mach numbers.