A review on the development of large-scale detonation-driven shock tunnels and the application in hypersonic flow tests

Since the beginning of this century, China has made tremendous progress in the fields of aerospace and deep space exploration. Representative achievements include a space station orbiting the Earth, sampling robots to the Moon, a Mars rover, and hypersonic air-breathing vehicles. When a spacecraft reenters the Earth’s atmosphere or enters the Martian atmosphere, one of the key problems is the hypersonic or high-enthalpy flow around the craft, which may impose extremely high aerodynamic and aerothermal loads on it. Reasonable design of a spacecraft is essential for survival during the high-speed reentry or entry phase, and a failed design may lead to disastrous consequences. In the history of human space exploration, there have been painful cases that demonstrate how important and necessary a rational design is. The space shuttles Challenger and Columbia broke apart in 1986 and 2003 during launch and reentry, respectively, killing all seven astronauts on board in each tragedy. These disasters finally resulted in a permanent halt to the space shuttle program and its replacement by the Dragon space capsule developed by SpaceX to send supplies and crew to the International Space Station (ISS) in recent years. During the design process of a spacecraft, hypersonic or high-enthalpy test facilities are essential requirements for the assessment of its aerodynamic performance.

The term “hypersonic” was originally coined by Tsien¹ specifically to refer to a flow at a speed exceeding five times the local speed of sound. Owing to its promising applications and distinctive fluid physics, hypersonic flow became a hot research field in the “Cold War” era.² Many hypersonic vehicles were developed or put into use, such as the rocket-driven X-15 aircraft, the X-43 airplane propelled by an air-breathing scramjet engine, space shuttles, which are partially reusable rocket-launched vehicles designed to go into orbit around the Earth, and Martian probes. Fundamental research on hypersonic flow has also made great progress in exploring shock–shock interference, boundary layer transition, and aerothermodynamic environments, among other aspects.

Despite more than a half century of research into hypersonic flow, there remain many challenges in designing hypersonic vehicles and accurately assessing their aerodynamic performance. The tragedy of the space shuttle Columbia reminds us that there are still “unknown unknowns” in the research field of hypersonic flight.³ Starship, the largest spacecraft in the world to date, was successfully launched into its scheduled orbit by SpaceX’s Super Heavy booster during its third flight test in March 2024. However, after the spacecraft reentered the atmosphere and encountered a high-temperature hypersonic flow environment, it burned down and failed to land. This implies that “unknown unknowns” in hypersonic reentry flights have not been fully studied and resolved, although two decades have passed after the disintegration of Columbia space shuttle. In a review article, Bertin and Cummings³ pointed out that the inability of wind tunnels to adequately simulate the high-enthalpy flow states of hypersonic vehicles reduced the reliability of test data. Therefore, developing appropriate wind tunnels along with suitable measurement techniques to gain reliable experimental data is the primary approach for hypersonic flow research. The main challenges in developing hypersonic wind tunnels include reproducing extremely high total temperatures, limitations of test flow size, and test duration.⁴ The operation of hypersonic wind tunnels requires very high driving system power, which makes traditional wind tunnels operating in continuous mode unsuitable. For example, the output power of a detonation-driven high-enthalpy tunnel reported by Jiang⁵ is higher even than the total installed power capacity of the Gezhouba hydropower station located in the Yangtze River in China. Owing to this tremendous power requirement, shock tunnels running in pulse mode are the only choice to simulate the high-enthalpy flow states needed for testing of hypersonic vehicles.

In this paper, research progress on detonation-driven shock tunnels, as a type of the high-enthalpy facilities mentioned above, is reviewed along with some application examples. Developments in detonation-driven shock tunnels have already been reviewed by several authors. However, most of these review articles were published before the construction of the large-scale detonation-driven shock tunnels that are the focus of the present work. In addition, benefiting from the advantages of large scale and long test duration, the engineering applications of such facilities have also been significantly expanded.

Many shock tunnels have been put into operation all over the world to study the fundamentals and engineering applications of high-enthalpy flows for hypersonic aircraft or spacecraft in the past decades. Figure 1 shows a schematic of a shock tunnel, which typically consists of six components: a driving shock generator, a driver section, a driven section, a convergent–divergent nozzle, a test section, and a vacuum tank.

As shown in Fig. 1, the driver and driven sections are separated initially by the primary diaphragm (PD). The secondary diaphragm (SD) separates the driven section and the nozzle. Generally, driver gas with a higher pressure p₄ and temperature T₄ fills the driver section, while the required test gas fills the driven section at a lower pressure p₁. The wave dynamic process is depicted in Fig. 2(a), which can explain the working mechanism of a shock tunnel. When the diaphragm PD between the driven and driver sections is broken, a shock wave [the incident shock wave (ISW) as shown in Fig. 2] is generated and propagates rightward in the test gas, followed by a contact surface (CS). Simultaneously, a centered expansion wave (EW, the fan-shaped wave configuration in t–x space) comes into being and propagates leftward into the driver gas. The flow domain is divided into four regions: the initial driver gas ④, the initial test gas ①, the test gas region ② compressed and heated by the ISW, and the uniform region ③ between the EW and CS. When the ISW reaches and is then reflected from the right end of the driven section, it becomes the reflected shock wave (RSW). As it propagates leftward, the RSW further compresses and heats the test gas in region ② into a state with the highest total enthalpy H₀ in gas region ⑤. This state corresponds to the stagnation condition for a shock tunnel working in a reflected mode.

There are two critical instants t_D and t_C at which the ISW and the reflected wave of the EW head (RHE) successively reach the right end of the driven section as depicted in Fig. 2(a). The effective test time of a reflected shock tunnel can be generally determined as τ_max = t_C − t_D in a tailored interface operation condition. It is easy to understand that τ_max is roughly proportional to the total length of the driver and driven sections, L_driver + L_driven. On the other hand, the ratio of the diameter to the length of the driven section should be maintained large enough to avoid excessive attenuation of the incident shock wave induced by boundary layer dissipation, for example, D/L_driven > 0.01 was recommended by Mirels.⁶ The main point to be emphasized here is that a large scale is critical to obtain a sufficiently long test time and a sufficiently large test flow for a high-enthalpy shock tunnel.

This is one of the main reasons why the use of a driver gas with a high speed of sound is essential for a high-enthalpy shock tunnel.

Three types of drivers for high enthalpy shock tunnels have been developed and put into operation around the world: free-piston drivers, heated light-gas drivers, and detonation drivers. These driver technologies are briefly introduced here, with a focus on their working principles, advantages, and disadvantages. More information on high-enthalpy shock tunnels can be found in to the work of Lu and Marren.⁷ 

A free-piston driver is a powerful technique to increase the speed of sound of the driver gas for a high-enthalpy shock tunnel. As shown in Fig. 3, the driver gas is compressed and heated by a fast-moving heavy piston. The concept of a free-piston driver was first proposed by Stalker^8–10 and applied to set up a series of high-enthalpy shock tunnels, including T3 at the National University of Australia, and T4 and X3 at Queensland University.¹¹ It has been implemented in several reflected-shock tunnels around the world, including T5 at the California Institute of Technology,¹² HEG¹³ at DLR in Germany, and HIEST at JAXA in Japan.¹⁴ In recent years, two piston-driven shock tunnels have been set up: T6 at the University of Oxford¹⁵ and FD21¹⁶ at the China Academy of Aerospace Aerodynamics. The last of these is the world’s largest free-piston driven shock tunnel in operation by far.

Free-piston drivers have the advantage of generating high-enthalpy test conditions, since the helium is generally used as the driver gas. They can drive the strongest shock waves and provide the highest total pressure capability among the three types of driver technologies. However, these drivers are far more complex mechanically. The need to move a heavy piston of up to a thousand kilograms at high speeds may lead to technical difficulties in launching and recovering. The test time for typical high-enthalpy test conditions is usually a few milliseconds with well-tuned operation. Therefore, only optical measurement techniques or those transducers with high response frequencies can be used in aerodynamic experiments on such high-enthalpy shock tunnels.

In a heated light-gas driver, a high speed of sound of the driver gas is achieved through the use of a light gas such as hydrogen or helium, typically heated to high temperatures by electrical heaters. The configuration of a typical heated light-gas driven shock tunnel is shown schematically in Fig. 4. Helium is the second lightest gas, with a molecular weight of 4 kg/kmol, and was used in early shock tubes and tunnels.¹⁷ Here, its advantages include a high speed of sound and chemical inertness. Hydrogen, with a molecular weight of 2 kg/kmol, is the lightest gas and is thus more suitable for light-gas drivers.

Typical heated light-gas driven shock tunnels are the series of Large Energy National Shock Tunnels (LENS) located at the Calspan–University of Buffalo Research Center (CUBRC) in the USA.^18–23 The driver gases are generally heated externally by electrical resistance heaters to temperatures of up to 800 K²³ to provide high-enthalpy test conditions. The excellent quality of the test flow conditions obtained in the LENS series facilities demonstrates that heated light-gas drivers represent a highly capable approach for high-enthalpy tests.

However, attention should be paid to several problems when hydrogen is used as the driver gas. One problem is that of safety, since the hydrogen is highly combustible and detonable. Another problem is the corrosion of metal structures caused by hydrogen embrittlement at high temperatures. On the other hand, as helium is very rare in the atmosphere and extremely expensive, the operating cost of a shock tunnel driven by helium is extremely high. All of these problems pose technical challenges to the use of heated light-gas-driven shock tunnels, especially with regard to large-scale model tests.⁴ 

A detonation driver uses high-temperature detonation products with a very high average speed of sound as the driver gas to generate the incident shock wave. The configuration of a detonation-driven shock tunnel and its wave diagram are schematically depicted in Fig. 5. The overall structure is similar to that of a typical shock tunnel, except that the driver section is filled with detonable gases, such as a hydrogen/oxygen mixture. During operation, the detonable gases are ignited by a strong energy input to directly initiate a detonation wave front (DWF) running to the left in the driver section followed by Taylor expansion waves (TEW). The detonation pressure is high enough to break the primary diaphragm (PD) and generate the incident shock wave (ISW), which propagates to the right in the driven section, followed successively by a contact surface (CS) and a series of central expansion waves (CEW). Generally, there is a short delay between the start of the incident shock wave (ISW) and the detonation wave front (DWF), as shown in Fig. 5(b). This short delay is caused by dynamic rupture of the primary diaphragm, which is a steel plate with prefabricated grooves. The thickness of the plate and the depth of the grooves are key parameters that need to be optimized to obtain a delay that is not too short and to avoid detonation extinction. The typical wave process in a detonation driver is shown in Fig. 5(b) and the wave structure at t = τ is shown in Fig. 5(c). Such a driving mode is called a backward detonation driver (BDD), because the detonation wave front, which carries the most energy, travels in the opposite direction to the ISW and is not utilized to drive the shock tunnel. If the detonation is initiated from the left end of the driver section, the DWF will travel to the right in the driver section and generate the ISW after breaking through the primary diaphragm. This driver mode is called a forward detonation driver (FDD). In an FDD, the incident shock wave decays continuously owing to the TEW.

The detonation driver was originally proposed and theoretically studied by Bird.²⁴ Experimental investigations of the performance of detonation-driven shock tunnels were conducted by Yu²⁵ and by Coates and Gaydon,²⁶ among others. These studies found that a BDD was not practically applicable owing to the extremely high impact load caused by the reflection of the DWF. This safety problem was eventually solved by Yu et al.^27–29 through the addition of a damping tank to the end of the detonation driver, and a series of detonation-driven shock tunnels were constructed, such as TH2-D^27,30,31 at RWTH Aachen University in Germany, and JF-10, JF-16,^{28,29,32–38} and some others at the Institute of Mechanics, Chinese Academy of Sciences. The HYPULSE hypersonic test facility at NASA’s GASL can also work in a detonation-driven mode.^39,40 Along with HYPULSE, another shock tunnel with a BDD has been put into operation in the USA, at the University of Texas at Arlington (UTA).^7,41,42 Recently, a low-enthalpy shock tube that also employs a detonation driver has been studied at UTA.⁴³ 

After more than half a century of continuous development, detonation-driven shock tunnels have become key members of the family of ground-based high-enthalpy test facilities running in a pulse mode. A series of review articles on the development of detonation shock tunnels have been published in the past two decades.^7,29–31,36 However, in recent years, a number of large-scale high-enthalpy shock tunnels using detonation drivers have been developed in China. These new test facilities, along with their applications in hypersonic testing, are briefly summarized in the following sections.

As indicated by Eq. (1), the total pressure and temperature or the total enthalpy of a shock tunnel depend on the strength of the incident shock wave, i.e., M_ISW. A large shock Mach number is needed to generate high-enthalpy test conditions. Therefore, the driving capability of the driver is proportional to the speed of sound of the driver gas, as expressed in Eqs. (2) and (3). Let us compare the driving performances of the three typical types of drivers used in high enthalpy shock tunnels, namely, free-piston, heated light-gas, and detonation drivers.

Generally, helium or hydrogen is used as the driver gas for a free-piston or heated light-gas driver. In a detonation-driven shock tunnel with a BDD, the effective driver gas is the detonation product mixture located between the tail of the TEW and the head of the CEW, i.e., region ④ in Fig. 5. The driver gas pressures to generate ISW with the required M_ISW are compared in Fig. 6 according to Eq. (3). Obviously, heated hydrogen has the best driving capability, since the lowest driver gas pressure is needed for a given incident shock Mach number. However, its applicability is limited by safety issues, especially for large-scale shock tunnels, along with other problems such as hydrogen embrittlement that the driver tubes may suffer. The driving capabilities of the remaining driver gases are ranked in decreasing order as follows: detonation products of a stoichiometric hydrogen/oxygen mixture (2H₂ + O₂), heated helium at 800 K, and detonation products of a stoichiometric hydrogen/air mixture (2H₂ + O₂ + 3.75N₂). The driving capability of a detonation driver can be easily tuned by varying the volume fraction of nitrogen in the driver gas, i.e., the coefficient α in the mixture 2H₂ + O₂ + αN₂. A value of α > 4 is not recommended, owing to difficulties in direct detonation initiation. Therefore, the driving capability of a detonation driver can be made to cover that of a heated-helium driver for high-enthalpy shock tunnels just by tuning the detonable mixture composition as indicated by Fig. 6. As the operating cost of a driver employing helium for large-scale aerodynamic tests is enormous, a detonation driver is a rational choice for the development of large-scale high-enthalpy shock tunnels.⁴ 

Although detonation drivers have great potential for application in large-scale shock tunnels, a series of technical challenges still need to be investigated and overcome.^44–46 With support from Chinese National Project of Scientific Instrumentation R&D, a detonation-driven high-enthalpy duplication shock tunnel using a BDD as shown in Fig. 5 was developed and commissioned by the end of May 2012.^5,47,48 This shock tunnel has the laboratory series number JF-12 and has been calibrated over the years.⁴⁹ The JF-12 shock tunnel features the largest structural size and the longest test time among hypersonic high-enthalpy shock tunnels in operation over the world by far. Its total length is ∼275 m, and the exit diameter of its nozzle is 2.5 m for relatively larger Mach numbers or 1.5 m for lower Mach numbers. The inner diameter of the driven section is 720 mm, which reduces the dissipative effect induced by the boundary layer. Figure 7 shows a photograph of JF-12 and a schematic indicating its main scales.

A test time of over 100 ms can be obtained for a hypersonic test flow with nominal Mach numbers between 5 and 9. This makes it possible to conduct hypersonic tests with large-scale models. The test time is long enough for flowfield establishment or the sensor response in aerodynamic force measurements with large test models. Moreover, the test conditions feature perfect repeatability, as depicted in Fig. 8, which is critically important for facilities running in the pulse mode. Hypersonic test conditions with required Mach numbers, Reynolds numbers, and total temperatures can be tuned easily by changing the composition of the detonable mixture or the initial filling pressures, and through the use of nozzles with different expansion ratios. Table I gives the typical test conditions obtained with the JF-12 shock tunnel. In this table, p₅, T₅, and H₀₅ are the pressure, temperature and enthalpy in the reservoir, i.e., the region ⑤ bounded by the RSW and the right end of the driven section, as depicted in Fig. 5. These parameters correspond to the stagnation conditions of the shock tunnel and are also called the total pressure, total temperature, and total enthalpy, respectively. The parameters with subscript ∞ represent the flow conditions in the test section, including the Mach number M_∞, temperature T_∞, pressure p_∞, Reynolds number Re_∞, and dynamic pressure Q_∞. In summary, therefore, the large-scale high-enthalpy JF-12 shock tunnel has a series of advantages, including long test time, low operating cost, good repeatability, and large-scale test flow field.⁴ 

The JF-12 shock tunnel works with a backward detonation driver (BDD). Test condition calibration and applications in large-scale hypersonic tests indicate that its driving capability is remarkable. However, as depicted in Fig. 5(c), its effective driver condition is the steady region between the TEW tail and CEW head, with a relatively lower pressure P₄ than the peak pressure of the DWF, P_CJ. In addition, the fluid in this region is static, unlike the detonation products at the DWF, which move at the local speed of sound relative to the DWF. This implies that a large proportion of the detonation energy is wasted. Even worse, this energy carried by the DWF results in huge loads of force and heat at the end of the BDD. A forward detonation driver (FDD) is a better choice than a BDD to produce hypervelocity high-enthalpy test flows where strong driving capability is necessary to generate strong incident shock waves.

The geometrical configuration of a shock tunnel driven by an FDD is the same as that shown in Fig. 5(a) for one driven by a BDD. However, the ignitor is moved to the left end of the detonation tube when the shock tunnel works in an FDD-driven mode. Therefore, the DWF propagates to the right, followed by the right-running TEW and drives the ISW in the driven section after rupture of the main diaphragm. The typical wave processes and transient structures for a shock tunnel with an FDD are shown in Fig. 9. In this driving mode, the TEW runs to the right and may catch up with the ISW in the driven section, resulting in attenuation of the latter. Therefore, the pressure after the incident shock wave (region ②) cannot remain steady, owing to the effect of the TEW in an FDD-driven mode, as shown in Fig. 9(c).

As the gas at the DWF has the highest pressure and temperature, an FDD possesses strong driving capability. However, the effective driver pressure p₄, along with other parameters, keeps decreasing owing to the interaction of the right-running TEW and left-running CEW. This is different from the behavior in a BDD shock tunnel. In a BDD shock tunnel as shown in Fig. 5, both the TEW and CEW travel synchronously in the same direction at the same speed, i.e., the local speed of sound at the TEW tail. Therefore, the driver conditions of a BDD remain steady, which is one of the advantages of BDDs. A cavity ring has been proposed and appended at the connection between the driver and driven sections to weaken the effect of the TEW. Application of this modification to FDD shock tunnels, including JF-10 and JF-16,^{29,32,34–36} have indicated that such a technique does work.

With the experience gained from JF-10,³⁴ JF-12,^7,36,47,48 and JF-16,^35,37,38 a hypervelocity high-enthalpy shock tunnel (laboratory series number JF-22) was built and commissioned in May of 2023 with support from the National Natural Science Foundation of China for Instrument R&D. JF-22 is located next to JF-12 in the same laboratory on the outskirts of Beijing, as can be seen in Fig. 10. The total length of JF-22 is ∼180 m, which is shorter than JF-12 owing to restrictions of laboratory space. The exit diameter of the nozzle is 2.5 m or 1.8 m while the inner diameter of the test section is 3.5 m. The lengths of the driver and driven sections are both 42 m. JF-22 is by far the largest facility in the world that can simulate hypervelocity flight conditions up to Earth-orbital speeds.

JF-22 can run in a backward-detonation-driven shock-reflection tunnel (BDD-SRT) mode, a forward-detonation-driven shock-reflection tunnel (FDD-SRT) mode, and a forward-detonation-driven shock-expansion tunnel (FDD-SET) mode. There are three convergent–divergent nozzles for the SRT mode and a divergent nozzle for the shock-expansion tunnel (SET) mode. The reservoir pressures p₅ for the BDD-SRT and FDD-SRT modes are shown in Figs. 11 and 12, respectively, which were obtained during calibration. A numerical comparison is presented in these figures. The BDD-SRT running mode has a longer test time, around 40 ms. The FDD-SRT mode results in a higher reservoir pressure, 20 MPa, around twice that obtained in the FDD-SRT mode under the unique initial filling condition in the driver section. This confirms that an FDD has a stronger driving capability than a BDD. Without doubt, the test time of the FDD mode is shorter than that of the BDD mode. The test conditions of JF-22 available so far are listed in Table II. It should be noted that the highest velocity shown in the table, 10.1 km/s, is achieved when the shock tunnel works in an SET mode, which is not discussed in the present review. More information on the SET can be found in the literature.^{21,22,35,37,38}

With the large-scale hypersonic test flow conditions provided by the JF-12 shock tunnel, a series of high-enthalpy tests have been conducted on large-scale models of hypersonic cruising vehicles. As the test gas can be pure air at high total temperatures, detonation-driven shock tunnels are suitable for combustion-related experiments for new-concept engines working at high Mach numbers. The test time is sufficiently long to allow the use of strain gauge balances for aerodynamic force measurement under high-enthalpy flow conditions in the JF-12 shock tunnel.^50–52 Such measurements encounter great technical challenges for hypersonic high-enthalpy test facilities running in the pulse mode. The large-scale test models that can accommodated in the test section of JF-12 make it realistic to conduct special measurements for high-temperature hypersonic flow. Some of the hypersonic tests conducted with the JF-12 shock tunnel are summarized briefly in this section.

The original intention when developing the JF-12 shock tunnel was to conduct fundamental research on supersonic combustion for hypersonic airbreathing propulsion, taking advantage of its long test time and high-temperature test flows of pure air. However, further experiments have focused on fundamental studies of hypersonic flows and on aerodynamic testing of hypersonic vehicles.

China’s first Mars probe, Tianwen-1, was launched on July 23, 2020 and successfully achieved a soft landing on the planet’s surface on May 15, 2021. During the design phase of Tianwen-1, a large-scale test model of the entry vehicle was tested in the JF-12 shock tunnel. Although JF-12 was initially designed for hypersonic test flows of air, its large-scale nozzle is also applicable for reproducing the hypersonic carbon dioxide flow conditions encountered during Mars entry.^53,54 The maximum Mach number of the Mars entry test flows of carbon dioxide that can be simulated by JF-12 is 7.3, which is lower than the maximum Mach number of 9 initially designed for air flows. A convergent–divergent nozzle needs a larger expansion ratio for carbon dioxide than air, owing to their different specific heat ratios. One of the primary challenges in simulating hypersonic Mars entry conditions is the nozzle design, which requires an extremely large ratio of exit to throat area.

A 1:4 scale test model of Tianwen-1 installed in the test section of JF-12 is shown in Fig. 13. This is the largest model by far tested in hypersonic Mars entry conditions, with a maximum diameter D_max = 850 mm. The schlieren images shown in Figs. 14(a) and 14(b) were obtained in hypersonic test flows of M_∞ = 6.2 and 7.3, respectively, generated by JF-12. The flow structures appear unsteady under these test conditions. The average relative shock detachment distance D_sdd/D_max of the tests were 0.063 and 0.039, respectively. This investigation revealed that the driver capability of BDD is still overpowered for carbon dioxide flows of moderate enthalpies. In addition, the hypersonic nozzle flow of carbon dioxide with its lower specific heat ratio than air makes the situation more challenging. The flow structures for a 1:8 scale model of Tianwen-1 in test flows of air are relatively steady, as can be seen in Fig. 15. The relative shock detachment distance is around 0.07 for these cases. From a comparison of the shock detachment distances in Figs. 14 and 15, one may roughly conclude that the carbon dioxide flow undergoes more compression than air. Strong luminescence can be seen in Fig. 15(b) for the case of higher total temperature, indicating that the air molecules in the shock layer have been excited to high-energy states.

In the test shown in Fig. 15, coaxial thermocouples were installed along three model generatrixes to measure the heat flux distribution along the model surface. Figures 16(a) and 16(b) show the results obtained under different conditions of total temperature. Obvious differences can be seen between the two tests, especially near the stagnation region, where the local segment of the bow shock wave is strong.

Shock–shock interaction (SSI) is one of the fundamental phenomena associated with hypersonic aerodynamics and can induce abnormal aerodynamic heating problems in hypersonic vehicles. SSI has been extensively investigated in both experiments and numerical simulations in the past half-century. Diverse overall interaction configurations arise, depending on the strengths and geometric parameters of the intersecting shock waves. Edney⁵⁵ defined six types of oblique/bow shock–shock interactions, namely, types I–V, on the basis of his tunnel tests as well as theoretical analysis. Experimental and theoretical studies have revealed that the most severe pressure and heating loads occur at the point where a jet impinges on the surface of a blunt body in a type IV shock–shock interaction.^55–57 

As the core flows of hypersonic and high-enthalpy test facilities are limited, small-scale test models are generally used in tests of SSI. In addition, the peak pressure and heating loads induced by jet impingement in type VI SSI occur within a very tiny region. The peak load point may be missed owing to installation of too few sensors during experimental measurements using small-scale test models. This problem of insufficient spatial resolution is often encountered in numerical validation and verification in the literature. The JF-12 high-enthalpy shock tunnel can provide large-scale test flows to accommodate large models for hypersonic tests. In addition, the maximum total temperature of the test facility reaches around 3700 K, which is high enough to emulate real hypersonic flight conditions.

The test model consists of a wedge to generate an incident oblique shock wave and a cylinder to generate a bow shock wave, as shown in Fig. 17. The key geometrical parameters are a wedge angle of 15°, a cylinder diameter of 240 mm, and wedge and cylinder spans of 600 mm. The streamwise and vertical positions of the wedge can be altered using adjusting-rails mounted on the base and wedge strut, respectively. This test model is large enough to allow the installation of sensors such as coaxial thermocouples and piezoresistive pressure sensors that are sufficiently robust to survive the extremely harsh test environment in the JF-12 tunnel.

Through the adjusting-rails mounted on the wedge strut as shown in Fig. 17, the vertical position of the wedge can be changed to obtain different interference points between the oblique and bow shock waves. Figure 18 shows flow structures from four runs with different interference conditions. Run 1 gave a detached bow shock wave without any interaction with the oblique shock wave, while runs 2, 3, and 4 corresponded to type IV SSI in which the shock interference point moved down successively. The supersonic jet is clearly visible in each frame. However, the locations of impingement on the model surface are different in each case.

Thermocouples of different sizes, i.e., Ø0.12, Ø0.7, and Ø1.4 mm, were used in the measurement to accurately capture the peak heat flux induced by jet impingement. The nondimensional heat flux distribution for run 3 is shown in Fig. 19. In this case, the supersonic jet hits the surface perpendicularly, which represents the worst aerodynamic heating state. The maximum Q/Q_s is ∼10.4, where Q_s is the heat flux at the stagnation point of the cylinder obtained in run 1 without shock interference. The temperature signals obtained by the thermocouples in the vicinity of the impingement point are also shown in Fig. 19. Oscillations in the profile indicate that the jet impingement point is not fixed, but oscillates along the cylinder surface.

Hypersonic high-temperature boundary layer transition is a fundamental phenomenon and has crucial effects on the aerodynamic and aerothermodynamic performance of hypersonic vehicles. General high-enthalpy shock tunnels with limited test time and test flow size cannot provide the spatial and temporal lengths for a boundary layer to evolve into a full developed state. The JF-12 shock tunnel, however, is a suitable test facility for such research. In recent years, experimental investigations on high-temperature hypersonic boundary layer flows over large-scale test models have been conducted using the JF-12 tunnel.^58–60 

Figure 20 shows the large-scale conical test model used in the JF-12 tunnel to investigate hypersonic boundary layer transition. This test model is the largest used in any such test to date, with a length of 3 m and a half-cone angle of 7°. Two bands of transition-forced trips as shown in Fig. 20 were installed, and these can also be replaced with conical rings for smooth wall conditions. Six hundred coaxial thermocouples were positioned on the model surface to record the heat flux distribution.

The surface heat fluxes from four tests under different conditions are shown in Fig. 21. In each panel, different colored symbols represent measurements along different generatrixes. For the case shown in Fig. 21(a), with lower total temperature and higher unit Reynolds number, the boundary layer undergoes a natural transition, as indicated by the rise in heat flux. For the cases shown in Figs. 21(b)–21(d), where T₀ is higher while Re is lower, transition occurs only when both 1# and 2# trip rings are installed simultaneously. The transition processes along different generatrixes are obviously asymmetric. Such experiments indicate that creating appropriate conditions for tests of high-temperature hypersonic boundary layer transition still faces great challenges.

Hypersonic air-breathing propulsion has been attracting attention from the academic community since the beginning of this century. The oblique detonation engine (ODE), a propulsion concept proposed more than 80 years ago, has reentered the field of view of researchers worldwide. Theoretically, the ODE features a simple configuration, with a short combustion chamber, high thermal efficiency, and applicability to high-Mach-number flight. However, reported research on ODE has been limited to numerical analysis and very fundamental experiments on the initiation of oblique detonation waves. One of the main challenges faced by experimental studies of prototype ODEs is the lack of high-temperature large-scale ground-based test capability. The JF-12 shock tunnel, however, can provide the conditions required for ODE tests. A series of tests on a prototype ODE have been successfully conducted in the JF-12 tunnel.^61–65 

Figure 22 shows a prototype ODE test model consisting primarily of an inlet wedge, a detonation chamber, and a truncated nozzle. The wedge is 1.6 m long with an angle of 25° relative to the freestream flow direction. Fuel is injected from the supply struts located at the leading edge of the inlet and mixed with the air test flow of JF-12 at high Mach numbers. An experimental schlieren image and temperature contours from a numerical simulation are depicted in Fig. 23(a) for an ODE fueled with hydrogen. A detonation wave remains straight and steady, and this is followed by a rapid combustion layer, indicated by the bright zone in the chamber. For the liquid fuel case shown in Fig. 23(b), the detonation wave is curved, forming a bow detonation wave (BDW) structure, which is different from the ODW in the case of a hydrogen-fueled ODE. A forced-initiation trip (FIT), a half-cylinder-shaped bar, is added to the inlet surface and placed shortly downstream of the leading edge. This simple structure leads to the initiation of a BDW. A further test without the FIT resulted in failure of detonation initiation.⁶⁵ These experiments demonstrate that the high-temperature test flows in the JF-12 shock tunnel meet the requirements for testing of ODEs, involving complex and time-consuming processes, including fuel injection, droplet breakup and evaporation, mixing, and detonation combustion. To the best of our knowledge, this is the first time anywhere in the world that ground-based ODE tests have been conducted with a prototype ODE model on an engineering scale.

As described in the preceding sections, the detonation driver technique has been applied to develop high-enthalpy shock tunnels. These large-scale high-speed test facilities have been used for hypersonic experiments for both fundamental and engineering studies. However, there is still room for performance improvement of detonation-driven high-enthalpy shock tunnels. In other words, both backward and forward detonation drivers (BDD and FDD) have their own shortcomings or deficiencies when running separately.

From the operating principle illustrated in Fig. 5, it can be seen that most of the energy carried by the detonation wave front (DWF) in a shock tunnel with a BDD is wasted, since the it propagates in the opposite direction to the test flow. The gas pressure P_CJ after the DWF is much higher than the pressure P₄ in region ④, which is the effective driver gas in a BDD shock tunnel. Even worse, because of the reflection of the DWF from the left end of the driver section, there may be unacceptable pressure [around (3γ − 1)P_CJ/(γ − 1)] and heat loads on the facility. In a shock tunnel with an FDD, the driver gas continuously loses pressure due to Taylor expansion waves (TEW), as illustrated in Fig. 9(c). In addition, the high-temperature gas in the stable state at the tail of the TEW, i.e., region ⑦ in Fig. 9, is wasted. In short, either BDD or FDD alone can only utilize a portion of the detonation energy, while wasting the remainder.

A straightforward idea is to simultaneously utilize both the gases associated with the high-energy DWF and the stationary end of TEW as the driver gases in a shock tunnel. Such a detonation driving mode is called a bidirectional detonation driver (BiDD), as depicted in Fig. 24. A transient pressure profile along the BiDD tubes is also shown in Fig. 24(a), and this allows the advantages of the BiDD to be explained. The detonation products in the stationary condition (region ④) with moderate pressure and temperature are suitable for driving an incident shock wave (ISW) of medium strength moving to the right in the driven Sec. I (T1) as a BDD mode. In the opposite direction, the detonation products shortly after the DWF in the Chapman–Jouguet condition with the highest pressure, temperature, and velocity (highest total enthalpy) are suitable for generating a strong incident shock wave moving to the left and running in an FDD mode. This incident shock wave is a PSW in the driven Sec. II (T3) but it becomes an SSW in the acceleration section (T4). Consequently, both medium-enthalpy and high-enthalpy hypersonic test airflows can be achieved in a single test with a shock tunnel running in the BiDD mode. It should be noted that the left half of the configuration shown in Fig. 24(a) is a shock-expansion tunnel (SET), which is generally used to simulate hypervelocity flows at orbital speeds.^{21,22,35,37,38} A reflected shock tunnel (SRT, such as the right half of the configuration) mode can also be an option if the divergent nozzle (nozzle 2#) is replaced with a convergent–divergent nozzle such as nozzle 1#. A BiDD provides a new approach for the development of next-generation detonation-driven high-enthalpy shock tunnels. Further details of the BiDD mode can be found in a recent report of a numerical study.⁶⁶ Owing to the fact that a shock tunnel operating in the BiDD mode requires more laboratory space and construction costs than one running in the BDD or FDD mode, no engineering implementations of this approach have been reported so far.

Despite the successful development of detonation driving techniques, challenges still remain for the design and operation of detonation-driven high-enthalpy hypersonic tunnels. The main problems that need further research are material or structural failures caused by surface corrosion or hydrogen embrittlement or by thermal and force loads under extremely high-temperature and high-pressure conditions, precise calibration of test flows, and measurement techniques in harsh testing environments.

Both the JF-12 and JF-22 high-enthalpy shock tunnels use nickel–chromium–molybdenum-based alloys, namely, PCrNi1Mo and PCrNi3MoVA, respectively, to ensure structural strength of tubes and connectors for the driver and driven sections, which have to operate in extreme environments with high mechanical and thermal loads, as well as strong corrosive influences. Here, the chromium, nickel, and molybdenum elements in the alloys together provide excellent corrosion resistance. Chromium forms a stable oxidation film on the surface of the alloy, effectively preventing the penetration of oxygen and water, and thereby protecting the internal metal from corrosion. Nickel enhances the overall corrosion resistance and stress corrosion cracking (SCC) resistance of the alloy. Molybdenum further enhances the corrosion resistance of the alloy in reducing environments such as detonable gaseous mixtures containing hydrogen. In particular, the chromium, molybdenum, and vanadium elements in the alloys can significantly improve resistance to hydrogen embrittlement at high temperatures. Although these alloys work well for the harsh environment of detonation-driven shock tunnels, corrosion cannot be avoided completely. The tubes for the driver and driven sections of the JF-12 shock tunnel need to be replaced or repaired after a certain period of operation. Corrosion can also cause contamination of the test flows to a certain extent. Therefore, new alloys with better corrosion resistance are still required in the future.

In the hypersonic research field, many high-enthalpy shock tunnels of different scales and with various driving methods have been successfully constructed and put into engineering application. The construction of these shock tunnels has coincided with the boom in hypersonic research during the first two decades of this century. However, the developments of the techniques required for precise calibration of test flows has posed a great challenge. Any intrusive diagnostic technique may cause serious interference with the test flow induced by the strong bow shock waves ahead of the probes. Descriptions of the test flow calibration methods used with existing high-enthalpy shock tunnels can be found in the literature.^36,67–69 A “flight time” method is generally utilized to calculated the speed or Mach number of incident shock waves through a series of ionization probes or thermal transducers mounted along the inner tube wall of the driven section. Then, from the incident shock Mach number along with the stagnation pressure measured at the reservoir by pressure transducers, the stagnation conditions of shock tunnels can be determined. In the test section, the test flow properties are generally evaluated using a calibration rake as shown, for example, in Fig. 25. A series of probes with pressure and/or thermal transducers are mounted on the rake to obtain the spatial distributions of flow parameters such as pressure and heat flux. The results of these measurements are then used for indirect calculations of the test flow conditions, such as the flow velocity or Mach number, static temperature and pressure, and dynamic pressure. Quantities associated with flow structures, such as the shock-detachment distance, oblique shock angle, and surface data, along with standard models based on simple geometries, are also used for hypersonic flow calibration. For the diagnosis and calibration of thermochemically nonequilibrium flows, optical diagnostic techniques such as absorption or radiation spectroscopy are promising because of their nonintrusive nature in contrast to conventional measurement techniques.

Computational fluid dynamics (CFD)-assisted test flow diagnostic methods have been studied and applied to the development of high-enthalpy shock tunnels.^70–74 However, owing to the extremely complex flow phenomena involved, including but not limited to strong unsteady discontinuities, shock wave–boundary layer interaction, high-temperature phenomena such as chemical reactions, and shock wave–diaphragm interaction, significant challenges still need to be addressed. The development of precise calibration techniques for hypersonic test flows is a promising and challenging topic for future research.

Reducing the total temperature is a practical method for simulating high Reynolds numbers in conventional hypersonic wind tunnels according to Eq. (4a), but it fails to simulate high-temperature-related behavior. On the contrary, some high-enthalpy facilities focus on the simulation of total enthalpy and total pressure, abandoning the pursuit of test flow scale or Reynolds number.

A sufficiently long test duration is a critical requirement for the establishment of the flow around tested models or for an adequate response of measurement instruments.

From the above observations, it is obvious that P₀, T₀, H₀, and D₀ are the critical parameters for high-enthalpy shock tunnels, as well as the flow Mach number M_∞ or velocity V_∞. The performances of some of the high-enthalpy shock tunnels in operation around the world are briefly summarized in Table III. Several high-enthalpy facilities in the construction or calibration phases are not included in the table. More details of the comparative capabilities of existing hypersonic facilities, including those working in the shock-expansion tunnel (SET) mode, can be found in a review article by Gu and Olivier.⁷⁵ In the future, to systematically improve the performance of high-enthalpy testing facilities, a large number of technical challenges will need to be addressed.

With the boom of deep space exploration in China since the beginning of this century, the demand for experimental research on high-temperature hypersonic flow has continued to increase. High-enthalpy hypersonic shock tunnels are suitable test facilities to conduct such fundamental research featuring the coupling of complex aerodynamics and thermochemical kinetics. Detonation drivers are among of the few methods capable of generating high-enthalpy hypersonic test flows. The advantages of detonation-driven shock tunnels include scalability, long effective test duration, and low operating costs.

In this paper, the primary configurations and operating mechanisms of three types of high-enthalpy hypersonic shock tunnels using free-piston, heated light-gas, and detonation drivers, respectively, have been briefly summarized. In the subsequent discussion, the structure and working principle of the detonation-driven shock tunnel have been emphasized. The development of large-scale high-enthalpy hypersonic shock tunnels driven by a backward detonation driver (JF-12) and a forward detonation driver (JF-22) at the Institute of Mechanics, Chinese Academy of Sciences has been described in detail. A series of applications to hypersonic flow tests with large-scale test models has demonstrated the success and advantages of this kind of high-enthalpy hypersonic shock tunnels.

Only a few examples of applications using the JF-12 hypersonic shock tunnel have been reviewed in this paper, while applications of the JF-22 hypervelocity high-enthalpy shock tunnel are still at the starting line. Both of these shock tunnels, however, will play an increasingly important role in the field of deep space exploration and research.

This study was co-supported by the National Natural Science Foundation of China (Grant Nos. 12172365, 12072353, and 12132017).

The authors have no conflicts to disclose.

Hu Zongmin: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Funding acquisition (equal); Investigation (equal); Methodology (equal); Project administration (equal); Resources (equal); Supervision (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal). Wang Wenhao: Data curation (equal); Investigation (equal); Methodology (equal); Writing – original draft (equal). Zhang Zijian: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Writing – original draft (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.