A boundary layer transition from laminar to turbulent flow occurs when the Reynolds number exceeds the critical threshold [14], as shown in Figure 4. Accurately predicting flow transition and understanding the underlying flow mechanism are of utmost importance in engineering. Research discovered that the frictional resistance coefficient increases by a factor of 4 times [16] when the supersonic boundary layer with a freestream of Ma = 3 transitions. Following the transition of the hypersonic boundary layer, the turbulent region often experiences a significant increase in both friction resistance and heat flux, often reaching levels that are 3–5 times higher than before [17]. The literatures [18] review the progress of hypersonic boundary layer transition and highlights the limits of the current predominant approaches of predicting transition, which typically rely on transition data or empirical formulas. It also pointed out that flight tests can serve as a viable approach for conducting transition studies in authentic flight situations. Many countries have carried out many transition flight tests, such as the Hypersonic Boundary Layer Transition (HyBOLT) transition control flight test conducted by the United States, and the compression surface transition of the scramjet forebody (LEA) the flight test carried out by France. Aircraft Flight Tests section will introduce these flight tests in detail.

As a necessary experimental facility for simulating real flight, hypersonic wind tunnels play a pivotal role in the study of boundary layer transition. As shown in Figure 5, the quiet wind tunnel [19] built by the National Laboratory of Turbulence and Complex Systems of Peking University can cover supersonic and hypersonic flows in the range of Ma 3.0∼6.5 with the diameter of nozzle exit being 300 mm. At present, the relatively powerful and easy-to-use near-wall measurement technologies mainly include temperature-sensitive paint (TSP), near-wall particle image velocimetry and Rayleigh-scattering visualization. They provide good experimental measurement methods for research transition. In addition, TSP is a non-contact optical temperature measurement technology that can achieve high spatial resolution temperature field measurement.

The hypersonic flow, being a type of high-enthalpy motion, must take into account some unique phenomena, including the failure of the calorically perfect gas assumption, thermochemical non-equilibrium effects, molecular vibration energy excitation, molecular ionization, material ablation, etc. Numerical modeling methods for high enthalpy flow transition and turbulence simulation [20] necessitate the development of many techniques, including real gas models and high-enthalpy boundary layer stability models. Several flow control strategies available for managing high enthalpy boundary layers were proposed, including but not limited to: 1) CO₂ injection, 2) wall blowing/suction, 3) wall porous coatings, and 4) roughness elements. The effectiveness of flow control in delaying the transition of hypersonic boundary layers in four forms of instability, namely, streamwise traveling waves instability, crossflow instability, Görtler instability, and attachment line instability, was highlighted by Liu et al. [21]. The employed techniques encompassed 1) roughness elements and finite amplitude band control, 2) wavy walls, 3) microporous surfaces, 4) localized heating/cooling of walls, 5) heavy gas injection, 6) synthetic jet, 7) blowing/suction, and so forth. The experimental progress in controlling hypersonic boundary layer transition was reviewed by Yang et al. [22]. The flow control techniques were thoroughly summarized and outlined, including roughness elements, cavities, porous walls, wavy walls, jets, wall cooling/heating, plasma, and more. Special emphasis was placed on the feasibility of plasma control to postpone the transition of hypersonic boundary layer.

To date, researchers have achieved proficiency in employing uncomplicated techniques to forecast the transition of boundary layers across fundamental geometries [23, 24]. Additionally, they have devised related algorithms or software [25], which offer valuable support for contemporary flow control analysis and engineering application [26]. However, the transition of hypersonic boundary layers remains a challenging task, and the underlying flow mechanisms are not yet completely comprehended. The aerodynamic and thermal protection design of the next-generation of hypersonic aircraft largely depends on the depth of understanding of transition mechanisms and the ability to control them.

A series of shock waves, which are frequently seen in supersonic inlets and at the heads of aircraft, define shock wave train, a complicated flow phenomenon [27]. Shock/shock interaction, which is intricately linked to shock wave trains, is a significant area of interest and obstacle. Understanding the mechanics of the subject matter and providing an accurate description and prognosis had substantial academic and practical significance. Schematic representations depicting six distinct shock/shock interactions on a blunt leading edge are illustrated in Figure 6.

A study was conducted to provide a comprehensive understanding of internal structures, oscillatory behaviors, and active/passive controls of shock wave trains [29]. The control methods in this study encompass three main approaches: 1) boundary layer suction, 2) bump control, 3) vortex generator. Via utilizing intelligent sensing technology, real-time information in the flow field is reconstructed. By employing adaptive adjustment of the suction air or vortex generator switch, the working efficiency is enhanced throughout a broad spectrum of inflow conditions. Luo et al. [30] provided detailed information on various techniques for controlling shock waves at the leading edge. These techniques include reverse jet flow, laser energy deposition, and plasma synthetic jets. For controlling oblique shock waves in the inlet and side wings, the main methods include plasma discharge, compression corner control, and magnetohydrodynamic methods. For controlling shock/shock interaction, the main methods include reverse supersonic jets, laser-based energy deposition methods, and plasma discharge.

As shown in Figure 7, in the SAV21 supersonic cascade [27], the shock wave train induced by high back pressure exhibits highly three-dimensional structure, facilitating the passage of the leading edge of the shock wave train down the throat. This phenomenon leads to a decrease in the mass-capturing coefficient and the occurrence of stall. When designed surface or end-wall suction slots were adopted, the maximum back pressure could be improved by 20%. Zhang et al. [31] investigated the asymmetry of oblique shock wave trains at Ma = 2.7. The experimental findings demonstrated that the oblique shock wave trains exhibited flow separation regions upon travelling ramps, characterized by a rapid increase in their motion velocity. Additionally, the direction of asymmetric separation deflection may undergo a change. Based on the interaction characteristics between the oblique shock wave train and upstream shocks, the slope control in the pipeline is employed to generate asymmetric upstream flow conditions by manipulating the deflection direction of the oblique shock wave trains.

The presence of shock waves/boundary layer interactions is prevalent in the flow patterns both within and outside transonic, supersonic, and hypersonic aircraft. Since first observed by Ferri [32] in 1939, it has quickly become a hot topic in the research of supersonic flow control. Simplified models are proposed to investigate the underlying physical mechanism of shock wave/boundary layer interactions, as shown in Figure 8, owing to the intricate nature. Swept compression ramps are commonly found in the inlets or surfaces of supersonic/hypersonic aircraft. Similarly, the utilization of dual-fin design is frequently observed in the upper sections of sidewall compressions. Surface protrusions on aircraft surfaces are frequently fabricated using cylinders, half cones, and fins as their primary shapes.

What we need to regulate is the most important aspect of flow control. Dolling [34] considered that the flow control of shock wave/boundary layer interactions was to reduce peak heat flux, decrease the fluctuating pressure loads, diminish the scale of separated flow, and move the frequency of fluctuation outside of the critical range. There are three strategies for controlling shock wave/boundary layer interactions summarized [35]: 1) increasing energy in the bottom layer of the boundary layer to enhance its resistance to adverse pressure gradients; 2) reducing pressure differences before and after shock waves in the near-wall region; 3) injecting energy to increase wall temperature, which raises the speed of sound and lowers the intensity of the shock wave. Specifically, methods such as bleeding and transpiration, perforations and porous media, MART technology (Mesoflaps for Aeroelastic Recirculating Transpiration, MART), streamwise slots, secondary flow circulation, and wall bump were employed. Zhong et al. [36] provided an overview of the evolution of shock/boundary layer interactions and its impacts on the flow process. Furthermore, Shi et al. [37] elaborated on methods for controlling shock wave/turbulent boundary layer interactions, including micro-vortex generators, plasma control, electromagnetic coupling effects, and other techniques.

The interaction between shocks and boundary layers is an inherent physical phenomenon that is commonly observed in high-speed aircraft. This phenomenon can manifest in several areas, including inlets, flow corners, and wings. Nevertheless, the majority of existing flow control studies primarily concentrate on supersonic flow characterized by Mach numbers spanning from 1 to 4, and there is still little research on hypersonic flows. Consequently, there is a pressing requirement for further exploration of hypersonic flow in forthcoming research endeavors. Furthermore, it is crucial to consider the significance of low-frequency oscillations [38] in the flow control of shock/boundary layer interactions.

The phenomenon of sonic boom is exclusive to supersonic flight. As illustrated in Figure 9, when an aircraft operates at supersonic speeds, its components and the gas emitted will cause significant disruptions in the surrounding air, leading to the formation of shock/expansion wave systems. The interaction and propagation of these wave systems towards the ground result in the formation of two primary shock waves, known as leading and trailing. As these two shock waves sweep across the ground, an observer perceives two explosive-like sounds, which are referred to as a “sonic boom” [40].

Zhang et al. [40] conducted a review on the sonic booms, covering the generation mechanisms, prediction methods and suppression techniques. The flow controls involved in mitigating sonic booms include: 1) quite spike, 2) staged aft body, 3) energy injection, 4) membrane vibration. Sonic boom refers to the acoustic phenomena that occurs when pressure waves originating from a supersonic source propagate through the atmosphere [41], and the flow control will be an important means for suppressing sonic booms in future supersonic civil aircraft. There are several potential flow control systems that could be considered for future application. These tactics encompass thermal energy injection, electromagnetic focusing, acoustic impedance, and vectored thrust engine technology. It was found that the use of a noise-reducing nose cone on an aircraft [42] leads to a notable decrease in the magnitude of the sonic boom, but with a slight increase in the drag coefficient. The critical length of the quiet cone varies with the flying altitude and the Mach number. Therefore, the use of multi-stage quiet cones with each level reaching its critical length was suggested. Ye et al. [43, 44]proposed an active control method of suppressing sonic booms. The proposed technique entails the establishment of an aperture in close proximity to the leading edge of the lower surface of an airfoil to facilitate suction. Simultaneously, an ejecting flow is introduced at the trailing edge of the aforementioned surface, while ensuring an equitable distribution of suction and ejection. The proposed methodology has the potential to substantially reduce the sonic boom level and flow drag experienced by supersonic aircraft. When a mass flow rate of 7.5 kg/s was applied to a NACA0008 airfoil, the absolute value of the maximum negative overpressure decreased by 56.77% and the drag coefficient was reduced by 10.96%.

As shown in Figure 26, the numerical methods employed for simulating high-speed flow primarily consist of direct numerical simulation (DNS), Reynolds-averaged Navier-Stokes simulation (RANS), and large eddy simulation (LES). The DNS directly solves the discretized Navier-Stokes equations, yielding very comprehensive information about the flow field. This includes time series data of three-dimensional flow fields, spatial structures of turbulence, as well as time-averaged values and fluctuations of specific flow variables. To address the upwind issue in the convective term of the Navier-Stokes equations, various approaches have been developed. These include flux vector splitting techniques such as Steger-Warming, Lax-Friedrich, Van Leer, as well as flux difference splitting schemes like Roe and HLLC. Shock-capturing techniques with high accuracy include the WENO, NND, GVC and WCNS. For high-enthalpy hypersonic flows, due to chemical non-equilibrium effects, there exists a weak coupling relationship between the vibrational energy of gas molecules and pressure. The shock sensors proposed by Jameson [100] or Ducro [101] are no longer applicable. Passiatore et al. [102] enhanced the shock sensor by replacing pressure variables with vibrational temperature, resulting in the effective detection of shock waves and contact discontinuities. Pirozzoli [103] identified several concerns about discontinuous regions. Overcoming Gibbs phenomena and integrating artificial dissipation continue to be difficult for DNS.

The advantage of the RANS method lies in its allowance for larger grid scales, which reduces computational costs significantly. However, it is necessary to develop models for the Reynolds stress terms. There are various prominent turbulence models, such as the k-ε model, the k-ω model, the SST model, and the S-A model. Knight and Degrez [104] performed a comprehensive evaluation of turbulence models for analyzing SWBLIs using RANS. The findings revealed that each turbulence model has a specific range in which it is suitable, and simulating flows with distinct characteristics could lead to substantial inaccuracies. Fu and Wang [105] summarized the RANS model for high-speed flows and proposed a RANS model with wider applicability and better robustness. The fundamental idea behind the LES approach is that small- and large-scale vortices are separated by the filter function, with the small-scale vortices being modelled and the large-scale vortices being solved directly [106–109]. This method’s computing cost lies in between the other two numerical approaches since it may be thought of as a hybrid of the DNS and RANS methods.

Supersonic and hypersonic flow control often involves dealing with complex geometric boundaries, such as grooved surfaces, wavy walls and porous media. Studying such problems should not directly employ real geometric boundaries for computation. Instead, it’s preferable to model and handle these boundary conditions equivalently. Wang et al. [110] utilized an aperture-type acoustic metasurface model for equivalent treatment, thereby avoiding the need to solve complex geometric boundaries and simplifying the computation process. Zeng et al. [111] employed a channel flow DNS program combined with the immersed boundary method to model and simulate flow over non-smooth blade surfaces in real-world scenarios, considering parameters such as arithmetic mean height, roughness, and effective slope. Ghosh et al. [112–114] employed the immersed boundary method (IBM) and a hybrid RANS/LES approach to solve for flow phenomena related to shock-boundary layer interactions.

With the rapid development of parallel computing technology, CFD methods have become increasingly imperative and gradually served as a mainstream approach in flow control research. Utilizing CFD methods for flow control in drag reduction and thermal protection applications enables the calculation of physical quantities that may be difficult or impossible to measure experimentally, and provides rapid access to flow information, which demonstrates strong flexibility. However, this method also has its limitations. For instance, the accuracy of CFD methods relies on the accuracy of the computational models and the applicability of the assumptions made, hence it is important to consider beforehand whether the selected model is suitable for the intended application.

Wind tunnel experiments are a critical tool for investigating high-speed flow control. By placing scale or subscale models within a wind tunnel, equipped with various sensors and flow visualization instruments, these experiments can simulate the real flow situations. Here are some famous hypersonic wind tunnels illustrated in Figure 27. Liu et al. [119] conducted a review of experiments on the receptivity stage and the linear growth stage of hypersonic boundary layer transition. The Supersonic Nano-tracer Planar Laser Scattering (NPLS) approach was described by Yi et al. [120] Based on this principle, further techniques for detecting density fields, Reynolds stress, and aerodynamic optics were also developed. A survey of flow control techniques for reducing heat and drag in supersonic and hypersonic flows [121] was conducted, including forward-facing cavities, counter-flowing jets, aerospikes, and energy deposition, as well as the advancements in experimental studies pertaining to their combination. Berry et al. [122] carried out experiments on the effect of multiple porous models of the Hyper-X on boundary layer transition. The findings confirmed that the configurations with serrated grooves or a single row of large holes can effectively force a boundary layer transition. Borg and Schneider [123] placed a 20% scale model of the X-51A fore-body in a quiet wind tunnel and equipped it with a trip wire. They discovered reducing the noise level of freestream could increase the transition Reynolds number by 2.4 times, significantly affecting the transition caused by roughness. As demonstrated in Figure 28, Zhu et al. [124] investigated the impact of the jet on the aerodynamic performance of the airfoil at various angles of attack in a Ma = 5.0 hypersonic wind tunnel. The findings indicated that manipulating the jet has a positive impact on the lift and pitch torque coefficients, and this effect becomes more pronounced as the jet outlet flow rises at each angle of attack. Xia et al. [125] studied the effects of Ramp-VG array (ramp vortex generator array, RVGA) on the supersonic mixing layer. The results confirm that RVGA can improve the energy distribution in the supersonic mixing layer, achieving the performance of increasing flow velocity and delaying transition.

In the past few years, supersonic and hypersonic wind tunnels have achieved significant progress, with the advent of high-enthalpy wind tunnels, shock wind tunnels and gun wind tunnels, making a further enhancement in the modeling capabilities of wind tunnel experiments. Nevertheless, there is still potential for further progress in wind tunnel tests. There is a need to enhance the manufacturing capabilities and precision of models, as well as to systematically address the effects caused by scale factors. However, it is imperative to optimize the efficiency of wind tunnels in order to simulate a wide range of flow environments.

Although numerous studies have explored flow control in high-speed fields, practical engineering applications in aircraft is still limited. This discrepancy highlights a gap between theoretical advancements and their implementation in practical flying scenarios [126]. This gap can be attributed to several factors, including the complexity of integrating flow control technologies into existing aircraft designs, the high cost of development and testing, and the challenges associated with scaling laboratory results to full-scale applications.

Choosing suitable flow control techniques for an aircraft’s transition in-flight is challenging because of the variations in flow conditions between flight and being on the ground. Without actual flight data as a reference, it becomes impossible to make accurate decisions. Consequently, numerous countries are currently engaged in active hypersonic boundary layer transition flight tests in order to get further flow data that corresponds to these tests. As shown in Figures 30A, B, the United States, in conjunction with Australia, launched the HIFiRE-1 [127] and HIFiRE-5b [128] vehicles in 2010 and 2016, respectively. HIFiRE-1 is a blunt cone with zero angle of attack, a half-cone angle of 7°, and a head radius of 2.5 mm, flying at Mach numbers ranging from approximately 3.04–5.79. HIFiRE-5b is an elliptical cone with a short axis half-angle of 7°, an aspect ratio of 2:1, and a head radius of 2.5 mm. It can be clearly seen from Figure 29 that there is a significant discrepancy between the real transition Reynolds number of HIFiRE-5b and the wind tunnel experimental results. As shown in Figures 30C, D, two vehicles, the MF-1 [130] and HyTRV [131], were launched in China in 2015 and 2019, respectively, and are used to collect transition data for blunt cones with a small angle of attack and lifting bodies with a variable angle of attack at hypersonic speeds. These data provided information on pressure, heat flux, and unstable wave amplitudes under different flow situations.

In some flight tests, attempts have been made to employ active flow control designs. A collaborative effort involving the United States, France, and other countries was conducted in the HyShot program, aimed at exploring the basic performance of scramjet engines at high Mach numbers. In 2002, the HyShot II [132] experiment was conducted, utilizing a fixed-geometry two-dimensional inlet with suction devices in the inlet to ensure its initiation. Active flow control could also enhance rudder authority by reducing airflow separation during large rudder deflections and significant sideslip angles. In 2015, Boeing and NASA [133] tested a pneumatic sweeping-jet-based active flow control system on the modified Boeing 757 eco Demonstrator. The flight test data suggested that the flow control actuation might be able to provide an approximately 14% increase in side force at the maximum tested rudder deflection and at critical sideslip angles.

In addition, some flight experiments achieved passive control by altering the flow profile through the installation of aerospikes or quiet spikes. For example, aerospikes were utilized on the HIFiRE 7 aircraft [134]. Gulfstream and NASA [135] installed multi-stage quiet spikes in front of the F-15B nose, conducting multiple flight tests in the range of from Mach 0.8 to 1.8. It was found that the quiet spikes successfully attenuated the strong shock waves at the nose into a series of weaker shocks, effectively reducing the sonic boom. The X-43A aircraft used rough strips [136] at the front end of the air inlet to induce forced transition of the boundary layer and lessen the sensitivity to flow separation. Saric et al. [137] conducted a study on the infrared thermal imaging results of the swept wings of the F-15B aircraft at Ma = 1.85. They discovered that the inclusion of periodic discrete roughness elements can increase the laminar flow area to 70%∼80% and reduce drag by 20%∼25%.

Flight experiments remain the most genuine and dependable step for pushing flow control to practice application. Given the exorbitant expenses associated with flight experiments, it is imperative to explore various methods to mitigate these expenditures. However, it is crucial to extract the maximum amount of relevant information from the minimal flight data that is accessible. Furthermore, it is imperative to enhance the sophistication and dependability of measurement instruments and sensors, while consistently investigating the potential for combining flying experiments with other research methodologies. As a result, while research continues to advance the theoretical understanding of flow control, translating these innovations into practical solutions for aircraft requires overcoming significant technical and economic hurdles. Addressing these challenges could potentially lead to enhanced performance, fuel efficiency, and overall safety in future aerospace vehicles.