激波/边界层干扰问题是可压缩湍流领域中一个基本的流动现象，具有十分重要的工程应用背景。干扰区内的流动非常复杂，其中包括边界层大尺度的分离与再附过程、激波低频振荡现象、大尺度涡结构的生成以及强压力脉动载荷等。本文采用直接数值模拟对来流马赫数Ma = 2.25，33.2° 激波角的入射激波/平板湍流边界层干扰流动的分离特性、再附边界层结构以及摩擦阻力的生成演化机制进行了研究。此外，还对来流马赫数Ma = 2.9，24° 偏转角的压缩-膨胀折角构型中的复杂流动进行了直接数值模拟，探讨了膨胀角高度变化对离心不稳定性以及干扰区的非定常运动特性的影响规律，还对壁面压力脉动场的空间分布结构进行了分析。主要的工作和研究成果如下：

(1) 入射激波与平板湍流边界层干扰流动的分离泡结构具有三维空间分布特征，整体上呈现扁平的单峰造型。通过定量对比三个不同展向站位的分析结果，研究了展向三维结构对分离区非定常运动特性以及瞬态分离微团的几何形态的影响规律。研究发现，分离泡的非定常特性表征为低频的、大尺度的膨胀和收缩运动。壁面压力脉动的功率谱密度结果表明，展向三维结构对非定常运动的特征频率影响较小。通过经验模态分解方法对分离泡面积脉动信号进行了低频重构，从中提取出了膨胀和收缩过程，借此条件统计分析分离泡呼吸运动对瞬态分离微团几何形状的影响。此外，应用本征正交模态分解对流向速度场进行了低阶近似，分析发现分离泡呼吸运动与低阶模态密切相关，前十个低阶模态重构出的流场能够精确复现分离泡低频的、大尺度呼吸运动。再附边界层内湍流运动的统计特性研究发现，边界层外层的相干结构强度明显增加、具有较大的空间尺度。壁面剪切应力脉动与流向速度脉动的时-空相关性分析表明，涡结构倾角在再附边界层的恢复过程中逐渐减小。通过空间两点的振幅调制相关性，对边界层内、外层结构间的尺度干扰效应进行了分析，提出了一种可能的机制来解释激波干扰区下游的强调制作用。

(2) 考虑激波入射角分别为33.2° 和28° 的两种情况，对比研究激波干扰强度对壁面摩擦阻力生成、演化特性的影响规律。研究发现，摩擦阻力在流动产生分离时的恢复速率较为缓慢；而边界层在保持附着的情况下，在一个边界层厚度的流向范围内快速恢复到来流充分发展湍流边界层的水平。本研究将最新提出的平均摩擦阻力分解方法应用在激波干扰区内，定量地考察摩擦阻力在不同激波入射角情况下的演化规律。研究发现，与充分发展湍流边界层不同，再附边界层内湍流相关项的贡献显著增强，而压力梯度引起的空间发展项在很大程度上抑制了摩擦阻力的产生。此外，通过二维经验模态分解技术对湍流脉动实现尺度分解，定量评估不同展向空间尺度的湍流运动对摩擦阻力生成的贡献。结果表明，在分离区和再附区来自外层大尺度结构的贡献占主导地位，而内层小尺度结构的贡献相当有限。然而在弱干扰的情况下，外层大尺度结构的贡献显著减少，导致外层大尺度和内层小尺度运动的贡献相当。

(3) 对压缩-膨胀折角干扰构型中的激波/湍流边界层干扰问题进行了直接数值模拟研究。考虑了膨胀角高度为𝐻/𝛿 = 4.25, 1.22 两种情况。当膨胀角高度较高时，流动规律与传统的压缩折角构型一致；而对于高度较低的情况，激波干扰区受下游膨胀角的存在影响较大，分离区的长度显著减小。壁面压力脉动的功率谱密度结果表明，干扰区非定常运动的低频成分被抑制。极限流线确认了Görtler涡在再附区附近存在。通过对边界层内流线的曲率半径和Görtler 数分布分析发现，离心不稳定性在压缩角附近区域得到保留，但在膨胀角附近区域消失。膨胀折角下游的流动结构相对复杂，可以观察到附加的激波串结构和新的流向涡生成。此外，本研究还在膨胀角上、下游流动之间发现了信号之间的负反馈机制。

Shock wave/boundary layer interaction is a fundamental phenomenon in compressible turbulence, and has important engineering application background. The flow in the interaction region is very complex, including large-scale separation and reattachment processes of boundary layer, low-frequency oscillations of shock, the generation of large-scale vortical structures, and strong pressure fluctuation loads. In this research, Direct Numerical Simulations (DNS) are conducted to investigate the separation characteristics, reattached boundary layer structures, and evolution characteristics of skin friction for the interaction flows between 33.2° incident shock and flat plate turbulent boundary layer at Mach 2.25. Additionally, DNS is performed for the 24° compression-decompression corner configuration at Mach 2.9 to explore the effects of corner height on the centrifugal instability and the unsteady motion characteristics in the interaction region, and investigate the spatial distributions of pressure fluctuations. The main conclusions are summarized below:

(1) Research on the separation characteristics reveals that the time-averaged spatial distribution of separation bubble exhibits a flattened, three-dimensional peak structure. Through a quantitative comparative analysis of results obtained from three different spanwise locations, the influence of spanwise three-dimensional structure on the unsteady motion characteristics of the separation region and the geometry of separation micro-clusters is investigated. It is found that the unsteady motion of separation bubble is characterized by low-frequency, large-scale dilation and constriction motions. Power spectral density analysis shows that the spanwise three-dimensional structure has little effect on the characteristic frequency of unsteady motion. Empirical mode  decomposition is used to extract the dilation and constriction motion processes from the signal of separation bubble area. Conditionally statistical analysis is then performed to investigate the influence of separation bubble breathing motion on the geometry of separation
micro-clusters. In addition, the proper orthogonal decomposition method is used to perform low-order approximation of the streamwise velocity field. The breathing motion of separation bubble is closely related to low-order modes, and the first ten low-order modes can accurately predict the low-frequency breathing motion. Study on the reattached boundary layer structures indicates that the shock interaction considerably intensified the coherent vortex structures, and the reattached boundary layer exhibited
large-scale structures in the outer region. The space-time correlation between fluctuating wall shear stress and streamwise velocity fluctuation reveal a gradual decrease in the structural inclination angle during the recovery process. The scale interactions are examined using a two-point amplitude modulation correlation, and a possible mechanism is proposed to explain the strong amplitude modulation in the downstream region.

(2) The research for evolution characteristics of mean skin friction considers the separated and attached boundary layers in the interaction region, corresponding to incident shock angles of 33.2° and 28°, respectively. The separated boundary layer shows a gentler and slower mean skin friction recovery rate, compared to the attached case where skin friction completes its recovery within one boundary layer thickness. A new mean skin friction decomposition method is applied in the interaction region to investigate the internal evolution characteristics quantitatively. The contribution associated with Turbulence Kinetic Energy production is significantly amplified, while the spatial growth contribution induced by the pressure gradient largely inhibited skin-friction generation, unlike the upstream turbulent boundary layer. The research also finds that the three decomposition components are distinctly different between the two cases, with outer large-scale structures dominating separation and reattachment regions, while contributions from inner small-scale structures are limited. In contrast, the attached case shows a dramatic reduction in contributions from outer large-scale structures, resulting in the outer large-scale and inner small-scale motions being equally important. Empirical mode decomposition technology is used to focus on the contributions of turbulent motions at different scales.

(3) A DNS is conducted to explore the shock wave/turbulent boundary layer interaction flows in the compression-decompression corner. The shock interactions are considered for two different heights, represented by 𝐻/𝛿 = 4.25, 1.22. A classic shock wave/turbulent boundary layer interaction flow is observed in the higher case. On the other hand, in the lower case, the separation region's size is significantly reduced, and the low-frequency unsteadiness is marginally suppressed in the interaction region, as indicated by the analysis of the mean and fluctuating wall pressure. The flow patterns near the reattachment line exhibited the presence of Görtler vortices. The distribution of curvature radius and Görtler number indicates that a strong centrifuge instability is retained in the compression corner region and reversed in the decompression corner region. The downstream flow of the decompression corner is relatively intricate, and the additional shocklet and new streamwise vortices are observed. Moreover, a negative response mechanism is identified regarding fluctuating wall-pressure signatures between the upstream and downstream of the decompression corner.