The Ram Accelerator: Review of Experimental Research Activities in the U.S.

Abstract

The ram accelerator, conceived at the University of Washington in 1983, is a scalable hypervelocity launcher, capable, in principle, of accelerating projectiles to velocities greater than 8 km/s. The device operates as an in-bore ramjet in which a subcaliber projectile, shaped like the centerbody of a cylindrical supersonic ramjet, is propelled through a stationary tube filled with a pressurized gaseous propellant mixture of fuel, oxidizer, and diluent. This propellant burns near the base of the moving projectile, generating thrust. The chemical energy density and speed of sound of the propellant can be adjusted, via gas pressure and composition, to control the in-tube Mach number and acceleration history of the projectile. Successful ram accelerator operation has been obtained at gas fill pressures up to 200 bar and projectile velocities up to 2.7 km/s. Scaling has been demonstrated in bore sizes ranging from 25 to 120 mm in research facilities around the world. Potential applications of the ram accelerator include hypervelocity impact studies, hypersonic propulsion research, kinetic energy weapons, and direct launch of acceleration-insensitive payloads to low Earth orbit. This paper presents an overview of the technology of the ram accelerator and its history and state-of-the-art in the United States.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The ram accelerator is a hypervelocity projectile launcher that uses chemical energy to accelerate projectiles to hypersonic speeds [39].¹ Although the ram accelerator launch tube resembles a conventional long-barreled gun, its principle of operation is very different, being closely related to that of supersonic airbreathing ramjet engines. A stationary tube, analogous to the cylindrical outer cowling of a ramjet engine (Fig. 1), is filled with a combustible gaseous mixture, e.g., methane, oxygen, and a diluent such as nitrogen, at pressures of 5–200 bar. Thin diaphragms close off each end of the tube to contain the propellant. No propellant is carried aboard the projectile, which is similar in shape to the centerbody of a ramjet. The projectile has a diameter smaller than the launch tube bore, and is normally fitted with guide fins that provide for centering in the tube. The projectile travels at supersonic speed relative to the propellant gas, which it compresses in the flow area contraction between the nosecone and tube wall. The propellant flow remains supersonic with respect to the projectile as the gas passes through the throat, i.e., point of minimum flow area between the projectile and the surrounding tube wall. Below approximately Mach 4, combustion typically occurs at full tube area behind the projectile and thermally chokes the flow, thereby establishing a normal shock system on the aftbody of the projectile that renders the flow subsonic downstream. This shock system recedes as the projectile Mach number increases. The process of thermal choking replaces the nozzle of a conventional ramjet engine, resulting in a stable combustion process that travels with the projectile in a propulsion cycle referred to as the thermally choked ram accelerator mode.

Fig. 1

Comparison of conventional ramjet engine to ram accelerator

The operational sequence of the ram accelerator (Fig. 2) is initiated by injecting the projectile into the ram accelerator tube at speeds greater than ~700 m/s by means of a conventional powder gun or light gas gun. A lightweight obturator, or piston, in contact with the base of the projectile seals the gun bore during this initial impulse. The acceleration from rest of the projectile/obturator combination compresses residual air in the gun’s launch tube via a series of reflected shock waves [26, 105, 104]. When the projectile punctures the entrance diaphragm, the slug of shock-heated air ignites the propellant near the base of the projectile. A stable combustion zone is thus formed which travels with the projectile, maintaining a wave of high base pressure that propels the projectile forward, in a manner analogous to an ocean wave pushing a surfboard (Fig. 3). The obturator rapidly decelerates following ignition and does not participate in the subsequent acceleration process. To keep the projectile centered in the tube, the projectile is fabricated with fins that span the bore of the tube or the tube is equipped with several internal guide rails that bear on an axisymmetric projectile.

Fig. 2

Operational sequence of ram accelerator. a Gun is loaded with projectile and obturator, and a charge of gunpowder or high pressure gas. Ram accelerator is pressurized with propellant to 5–200 bar. b Gun fires obturator/projectile combination into ram accelerator. c Combustion is initiated and moves with projectile, sustaining high base pressure that accelerates projectile to high velocity

Fig. 3

Pressure distribution in conventional gun and ram accelerator. Projectiles are stabilized with either fins or rails

What distinguishes the ram accelerator from a gun is that its source of energy (the combustible gas mixture) is uniformly distributed throughout the entire length of the accelerator tube, whereas in a gun the energy source is concentrated at the breech as either a charge of gunpowder or high pressure gas. During the ram acceleration process the highest pressure in the tube is always at the projectile’s base (see Fig. 3), rather than at the breech as in a gun, and the bulk of the combustion products moves in a rearward direction. Only a small volume of high pressure gas exits the tube with the projectile. These characteristics of the ram accelerator result in much more uniform acceleration of the projectile, very high velocity capability, and very little muzzle blast and recoil. Furthermore, the acceleration and muzzle velocity of the ram accelerator can be easily tailored to specific needs by adjusting the propellant composition and fill pressure.

Potential applications of the ram accelerator include hypersonic aerodynamic testing [17, 84, 108], scramjet simulation [15], and direct launch to orbit [7, 12, 48, 80, 109, 52, 56], and hypervelocity kinetic energy weapons [66].

The propulsive cycle illustrated in Fig. 1 is the thermally choked ram accelerator mode, which operates with in-tube projectile Mach numbers typically ranging from 2.5 to 4 and at velocities below the Chapman-Jouguet (CJ) detonation speed of the propellant, i.e., at subdetonative velocities [16, 39, 40]. In this mode the thrust is provided by the high projectile base pressure resulting from the normal shock system that is stabilized on the body by thermal choking of the flow at full tube area behind the projectile. The ram accelerator can be modeled analytically using a simple one-dimensional control volume approach [16, 39, 50]. The gasdynamic conservation equations and the ideal gas law are applied to a control volume that contains the projectile (Fig. 4).

Fig. 4

Control volume for one-dimensional, quasi-steady analysis of ram accelerator

Assuming quasi-steady flow, the following expression for the non-dimensional thrust on the projectile can be derived:

$$\frac{F}{{AP_{1} }} = \frac{{\gamma_{1} M_{1} }}{{\gamma_{2} M_{2} }}\left( {1 + \gamma_{2} M_{2}^{2} }\right)\left[ {\left( {\frac{{\gamma_{2} - 1}}{{\gamma_{1} - 1}}} \right) \cdot \frac{{1 + \frac{{\gamma_{1} - 1}}{2}M_{1}^{2} }+Q}{{1 + \frac{{\gamma_{2} - 1}}{2}M_{2}^{2} }}} \right]^{{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-0pt} 2}}} - \left( {1 + \gamma_{1} M_{1}^{2} } \right)$$

(1)

where F is the thrust, P ₁ is the propellant fill pressure, A is the cross-sectional area of the tube bore, M ₁ is the Mach number of the flow entering the control volume (i.e., the projectile Mach number with respect to the undisturbed propellant), M ₂ is the Mach number of the flow exiting the control volume, Q = ∆q/c _p1 T ₁, is the non-dimensional heat release parameter, ∆q is the heat of combustion, c _p1 and T ₁ are the specific heat at constant pressure and the temperature of the undisturbed propellant, respectively, and γ ₁ and γ ₂ are the pre- and post-combustion specific heat ratios. This thrust coefficient equation applies to all ram accelerator propulsive modes operating in a quasi-steady manner, even though no details of the internal flow are considered in its derivation. The quasi-steady assumption is applicable for projectile accelerations up to about 15,000 g [16], and the ideal gas equation of state can be used up to about 25 bar fill pressure. Operation at higher pressures and/or accelerations requires the use of a real gas equation of state and an unsteady analysis, as summarized in a later section of this paper.

If one knows how M ₂ varies with M ₁ in a given propellant, then the projectile thrust can be readily computed for any flight velocity. Thermal choking of the flow behind the projectile (M ₂ = 1) corresponds to an entropy extremum [50]; thus, the details of the process which brings the flow to choking do not affect the end state conditions of the thermally choked ram accelerator mode and do not have to be known to predict the thrust. For propulsive cycles that do not involve thermal choking, such as the transdetonative and superdetonative modes discussed later in this paper, the details of the flow field around the projectile must be considered to accurately predict the exit Mach number, M ₂ [50, 53].

Figure 5 shows a plot of the non-dimensional thrust in the thermally choked mode, as a function of projectile Mach number for several typical values of the heat release parameter, Q. It can be seen that thrust increases with increasing heat release. The model also predicts that the thrust goes through a maximum and decreases with increasing Mach number, reaching zero when the projectile velocity is equal to the CJ detonation speed, V _cj , of the propellant [16, 39]. In order to achieve velocities higher than the V _cj of a particular propellant, the ram accelerator tube can be subdivided into several sections, called stages, each separated from its neighbor by a thin diaphragm and filled with a different propellant, as shown in Fig. 6 [16]. By selecting the sequence of propellants in such a manner that the speed of sound and detonation speed increase toward the exit of the ram accelerator, the projectile Mach number can be kept within limits that maximize thrust and efficiency, resulting in high average acceleration and a higher final velocity than is achievable with a single propellant stage.

Fig. 5

Non-dimensional thrust as a function of projectile Mach number and non-dimensional heat release parameter, Q = ∆q/c _p /T ₁

Fig. 6

Staging of ram accelerator. Pressure can be different in each stage. a ₃ > a ₂ > a ₁, where a _n is the speed of sound in stage n

It has been observed that acceleration is also possible when the projectile is traveling above the CJ detonation speed of the propellant—this is called the superdetonative velocity regime [70]. The transition from subdetonative to superdetonative operation occurs smoothly, through the transdetonative velocity regime (Fig. 7) [27]. As the projectile approaches V _cj of the propellant in the thermally choked propulsive mode, the combustion begins to move forward relative to the projectile, so that some of it takes place in the space between the projectile and the tube wall [40, 53]. As the projectile continues to accelerate to velocities above about 1.1V _cj , i.e., into the superdetonative regime, the combustion appears to move almost entirely forward of the projectile’s base. It is postulated that during this transition from subdetonative to superdetonative operation the combustion changes from purely subsonic to purely supersonic, and may even stabilize into an oblique detonation wave [40]. During operation in the transdetonative velocity regime, between approximately 0.9V _cj and 1.1V _cj , it is believed that regions of both subsonic and supersonic combustion coexist [27].

Fig. 7

Transdetonative (upper) and superdetonative (lower) propulsion modes

2 Experimental Facilities and Results

The first experimental ram accelerator facility, having a 38-mm tube bore, was completed at the University of Washington (UW) in September 1985, and proof of concept was achieved in June 1986 [38]. Interest in this technology spawned facilities at several other laboratories in the U.S.A. and abroad [11]. They included a 120-mm-bore ram accelerator (the world’s largest) at the U.S. Army Research Laboratory (ARL) at Aberdeen Proving Ground, MD; 90 and 30-mm bore smooth-bore systems, and a 30-mm-caliber railed-tube system, all at the French-German Research Institute (ISL, France); a 25-mm-bore installation at Tohoku University, Japan [94]; and a 15 × 20 mm rectangular-bore facility at Hiroshima University, Japan. In addition, a 37-mm-bore ram accelerator was built and successfully tested at the China Aerodynamics Research and Development Center (CARDC) in Mianyang, China [79]. Projectile masses from 5 g to 5 kg have been launched to velocities up to 2.4 km/s in these facilities. Ram accelerator research at all facilities focused on improving the understanding of the physical principles of ram acceleration, achieving higher velocities, developing robust projectile designs, and studying various near- and long-term applications. Pertinent compendia of results from these research facilities can be found in Takayama and Sasoh [106], and Bauer [2]. Here we focus on the facilities in the U.S.

2.1 University of Washington (UW)

The University of Washington ram accelerator facility has a 16-m-long test section and a 38-mm-bore light gas gun for a pre-launcher (Figs. 8 and 9). The first 4 m of the test section are comprised of thick-walled tubes (203 mm O.D.) for high pressure experimentation, i.e., fill pressures up to 200 bar. The launch tubes of the last 12-m of the test section are limited to a fill pressure of 75 bar. The test section has electromagnetic sensors to track the time-distance history of the projectile to determine its velocity and thrust, and to locate its position relative to the tube-wall Pressure-field measurements made by piezoelectric pressure transducers. Projectiles having three, four, and five fins have been found to have very similar operating characteristics in this facility. The projectile shown in Fig. 10 is a five-fin configuration with a 10° nose cone, 29-mm-diameter throat (point of maximum projectile cross-section), and overall length of 153 mm. Projectiles fabricated from alloys of magnesium, aluminum, and titanium having a mass range of 50–140 g have been used extensively in this experimental program. Ram accelerator operation has been demonstrated at velocities ranging from 0.7 to 2.7 km s⁻¹ and in-tube Mach numbers of 2.5–8.5. Sustained accelerations averaging 54,000 g with 110-g projectiles have been demonstrated with the thermally choked propulsive mode using propellant fill pressures up to 200 bar. While the velocities attained to date can be matched by research-grade powder guns and exceeded by light gas guns, those devices, unlike the ram accelerator, are extremely difficult to scale up to bore diameters greater than 100 mm without degrading their performance.

Fig. 8

Schematic of University of Washington 38-mm-bore ram accelerator. Test section is 16 m long. High-pressure section installed in 1997 (see Fig. 17 for details)

Fig. 9

View of UW 38-mm ram accelerator facility, ca. 1990. In foreground, from left to right, A.P. Bruckner, C. Knowlen, and the late A. Hertzberg

Fig. 10

Five-finned ram accelerator projectile with obturator. Fin span = 38 mm; nose half-angle = 10°

During the nearly three decades since the first experimental proof-of-concept was demonstrated at the UW [38], much has been learned about the phenomena that govern the ram accelerator, and many strides have been made. Here only some of the salient results are summarized; the interested reader is directed to the references for further details.

Figure 11 shows the pressure distribution on the projectile travelling at 1240 m/s (M = 3.4, subdetonative) past the location of the pressure transducer in a propellant mixture comprised of 2.8CH₄ + 2O₂ + 5.7N₂, at a fill pressure of 25 bar. The time-scaled profile of the projectile is drawn in the figure to illustrate the location of the pressure features with respect to the projectile geometry. The pressure profile of a quasi-one-dimensional flowfield model that accounts for shock losses, viscous pressure drop and finite rate heat addition is shown by the dashed line [49]. It is apparent that the predicted pressure amplitudes in the throat region and the normal shock location do not agree very well with the experiment, however, the predicted net thrust agrees quite well with the experiment, which strongly supports the assumption that the flow is thermally choked behind the projectile. In the subdetonative velocity regime, the observed pressure distribution is characteristic of the thermally choked propulsive mode and moves with the projectile as it accelerates through the tube. The evolution of this pressure distribution as the projectile approaches the propellant CJ speed is described in detail in [1, 59], and more recently in [6].

Fig. 11

Pressure profile on projectile in thermally choked, subdetonative mode

Staging a ram accelerator to attain high velocity when operating in the thermally choked propulsive mode has been demonstrated with many different propellant combinations. Shown in Fig. 12 are the velocity-distance data from a four-stage experiment carried out with an 80-g aluminum-alloy projectile that was accelerated from 1.1 to 2.6 km/s. In all four stages, the projectile velocity was less than 0.8V _cj to maintain high effective thrust. The theoretical predictions agree very well with experiment, again supporting the presumption that the combustion process is thermally choked. The velocity-distance data from a single-stage experiment with a 77-g projectile are also shown in Fig. 12. As the projectile approached the propellant CJ speed, its acceleration increased beyond that predicted for the thermally choked propulsive mode, indicating that the flow had ceased to be thermally choked at full tube area. The projectile accelerated up to ~2.0 km/s, which is approximately 1.2V _cj , and then coasted at nearly constant velocity for the last meter of the test section.

Fig. 12

Velocity-distance profile in four-stage ram accelerator [54]

The experimentally observed variation in the thrust as a function of the velocity ratio V/V _cj , as the propulsive mode makes the transition from subdetonative to superdetonative is shown in Fig. 13 for three different propellants (including data from the single-stage experiment in Fig. 12). The thrust reaches a minimum in the transdetonative velocity regime and then increases in the superdetonative regime, reaching a relative maximum before decreasing in the manner predicted for supersonic combustion ram accelerator operation.

Fig. 13

Dependence of thrust on velocity ratio for various propellants [40]

To date, maximum superdetonative velocities of ~1.5V _cj have been observed at the UW [40, 70] and 1.57 V _cj elsewhere [99]. It has been suggested that a maximum limit of approximately 2V _cj may exist due to energy balance considerations, i.e., the thrust equals drag limit [93], but this putative limit has yet to be confirmed experimentally, and its underlying analysis may be flawed [41]. In other work the maximum velocity of the ram accelerator has also been predicted to be about 2V _cj . For example, in the superdetonative regime maximum velocities in the range of 7–9 km/s for operation in hydrogen-based propellants with CJ speeds of 3–4 km/s have been predicted [111, 112]. Although aerodynamic heating of the projectile at the associated high Mach numbers is expected to be severe, its effects can be minimized through the judicious choice of refractory projectile materials and by other means [7–9]. Another velocity limiting mechanism that has been explored is that of the so-called “doomed propellant fraction,” which refers to the possibility that at sufficiently high Mach numbers the bow shock standing off the finite radius of the projectile’s nose tip may pre-ignite a sufficient fraction of the propellant to cause thermal choking of the flow at or ahead of the projectile throat [34]. This limit has not yet been observed experimentally and, in any case, was predicted to occur at velocities above the thrust equals drag limit.

Experiments performed with a variety of propellants have demonstrated the existence of operational limits that are governed by the heat release of combustion, the projectile in-tube Mach number, and the projectile material. Figure 14 shows the operational envelope in terms of the propellant heat release and the projectile Mach number [43, 44]. If the heat release is too small, the driving pressure wave is unable to remain coupled to the projectile and falls behind, resulting in a cessation of thrust, while if the heat release is too high, the driving pressure wave surges ahead of the projectile, causing a sudden deceleration—this is called an “unstart”. Hence, selection of the appropriate propellant composition is crucial to successful operation. The ultimate velocity limits, on the other hand, are believed to be related to projectile structural integrity and to the thrust equals drag limit. As the velocity increases, the pressure and aerodynamic heating increase markedly and are capable of causing structural failure of the projectile. Computations of heat transfer to the nose cone, and to the leading edges and lateral surfaces of the centering fins, performed at the UW [29, 30] and also at ISL [82, 83, 98], have shown that magnesium and aluminum alloys reach their melting points rapidly at these locations, resulting in potentially severe erosion by ablation, and loss of structural strength. Projectiles made of titanium alloy do not suffer these deleterious effects and have been found to attain higher velocities [55].

Fig. 14

Operational envelope of ram accelerator

High spatial resolution pressure measurements of the flow around the projectile have revealed a complex three-dimensional flow structure associated with the centering fins [46]. These observations have been corroborated by high-speed in-bore photography of projectiles through transparent polycarbonate tube sections, as shown in Fig. 15 [58]. Canting of the projectile in the tube has also been frequently detected [47], and is likely due to lateral forces and pitching moments generated by non-uniform pressure distributions around the nose and body of the projectile, coupled with erosion or bending of the projectile’s centering fins. This problem is mitigated through the use of titanium alloy as the projectile material, which is significantly stronger and more heat-resistant than the magnesium and aluminum alloys commonly used in the past [55].

Fig. 15

In-tube photograph of ram accelerator projectile. Propellant: 2.5CH₄ + 2O₂ + 5.6N₂, fill pressure = 6 bar, V = 1600 m/s (M = 4.4)

Studies of the starting dynamics of the ram accelerator were first carried out at the UW [14, 26], and further work on this topic ensued later at other facilities [3, 5, 28, 35, 69, 93, 95–97, 104, 105]. These studies and research on low velocity starting dynamics in propellants with low acoustic speeds carried out at the UW [57, 69, 96], has shown that the starting process is very complicated and highly dependent on initial conditions, such as propellant composition, fill pressure, projectile and obturator mass, entrance diaphragm thickness, projectile velocity, and residual pressure in the launch tube of the pre-accelerator gun.

2.1.1 Operation at High Fill Pressures

Beginning in 1996, the attentions of the UW group turned toward operations at high propellant fill pressures, up to 200 bar [18–25, 51]. The motivation for this work was the desire to achieve higher velocities with shorter tube lengths, and to study real gas behavior and the effects of high accelerations, under which the quasi-steady ideal gas model is no longer applicable. Due to the extremely high pressures of the ram accelerator combustion process that arise when the fill pressure is greater than 25 bar, real gas effects play a significant role by shifting the chemical equilibria and increasing the exhaust pressure. Even though all the thermodynamic parameters are affected by real gas behavior, the dominant effect is the corresponding increase in heat release, Q. Thus, the influence of real gas behavior on the thrust of ram accelerator propulsive modes can be evaluated to the first order by including a corrected value for heat release into the thrust equation already presented (Eq. 1). Figure 16 shows computed results for the variation of thrust coefficient with Mach number in a particular propellant mix at fill pressures of 5 and 20 MPa, assuming quasi-steady flow and the Boltzmann equation of state.

Fig. 16

Predicted thrust coefficient variation with Mach number at various fill pressures for quasi-steady flow using a real gas equation of state

Furthermore, operation at very high propellant-to-projectile density ratios (i.e., at high fill pressure or with a low-mass projectiles) leads to very high accelerations, which also affect the thrust performance. To investigate the effects of projectile acceleration, a _p , on net thrust, a one-dimensional unsteady flow model was developed which accounts for the accumulation of mass and momentum within the finite length of the control volume [25]. The unsteady thrust equation for the thermally choked ram accelerator propulsive mode can be expressed as:

$$\begin{aligned} & AP_{1} M_{1} \frac{{\sqrt {\gamma RT_{1} } }}{\gamma - 1}\left( {1 + \frac{\gamma - 1}{2}M_{1}^{2} + Q} \right) \\ & \quad = \frac{{\frac{7}{2}AP_{1} M_{1} a_{p} L_{CV} }}{{\sqrt {\gamma RT_{1} } }} \\ & \quad \quad + \frac{{\gamma \left( {m_{p} a_{p} + AP_{1} \left( {1 + \gamma M_{1}^{2} } \right) - {{AP_{1} M_{1} a_{p} L_{CV} } \mathord{\left/ {\vphantom {{AP_{1} M_{1} a_{p} L_{CV} } {RT_{1} }}} \right. \kern-0pt} {RT_{1} }}} \right)^{2} }}{{2\left( {\gamma^{2} - 1} \right)\left( {AP_{1} M_{1} \sqrt {\gamma /RT_{1} } - {{AP_{1} a_{p} L_{CV} } \mathord{\left/ {\vphantom {{AP_{1} a_{p} L_{CV} } {M_{1} \sqrt {\gamma R^{3} T_{1}^{3} } }}} \right. \kern-0pt} {M_{1} \sqrt {\gamma R^{3} T_{1}^{3} } }}} \right)}} \\ \end{aligned}$$

(2)

where m _p is the mass of the projectile, L _CV is the control volume length, and T ₁ and R are the propellant static temperature and gas constant entering the control volume, respectively. The other variables are as defined for Eq. 1. For brevity, the steps required to derive this equation are not included here; see the above-cited reference for details and Bauer et al. [4] for refinements that include real gas effects on all thermodynamic parameters and variations in the control volume length with Mach number. Equation [3] is an implicit expression for determining the dependent variable a _p as a function of projectile Mach number M ₁ and real gas heat release value for Q. Based on experiments, it has been found that an appropriate approximate control volume length for illustrating unsteady effects is about twice that of the projectile length, even though it tends to decrease in length as the Mach number increases due to increases in kinetic rates [4].

For the high-pressure experimental studies a 4-m-long section of the ram accelerator was replaced with three thick-walled tubes capable of withstanding a static pressure load of 10,000 bar. Figure 17 shows a schematic of this test section. At the same time a semi-automatic gas fill system was installed that enabled delivery of mixed propellant to fill pressures up to 200 bar. Titanium alloy projectiles at entrance velocities as low as 1200 m/s were successfully started in CH₄/O₂/N₂ propellants at fill pressures of 150 and 200 bar, the latter being the highest operating pressure achieved to date in any ram accelerator. At 200 bar a velocity of 2400 m/s was achieved within the 4-m length of the high pressure section. Due to real gas effects on the acoustic speed of the propellant, the throat-to-bore diameter ratio of the projectiles had to be reduced from the nominal value of 0.76–0.60 in order to enable operation at pressures greater than 150 bar [19, 24].

Fig. 17

Schematic of high pressure section of UW 38-mm-bore ram accelerator

The average acceleration achieved in these high-pressure experiments was much lower than that predicted by the real-gas calculations of the one-dimensional quasi-steady control volume model; an example at 200 bar is shown in Fig. 18. This discrepancy is due to the high accelerations that projectiles undergo at high pressures, which makes the quasi-steady assumption inappropriate for these conditions. The velocity-distance prediction from the unsteady one-dimensional performance model that accounts for the influence of projectile acceleration on the thrust behavior of the ram accelerator [19, 24] is also shown in Fig. 18. The agreement between this latter theory and experiment is very good over a significant velocity range. The experimentally observed velocity begins to diverge upward beyond 1.7 km/s due to the onset of the transdetonative propulsive mode [19, 24].

Fig. 18

Comparison of quasi-steady and unsteady control volume models with experimental data at 200 bar fill pressure, using Boltzmann equation of state [13]

Figure 19 shows how non-dimensional thrust F/p ₁ A is predicted to behave at different pressures for a constant mass projectile, i.e., at different acceleration levels. The curve corresponding to the quasi-steady solution applies strictly for the case of zero acceleration, but is reasonably accurate up to about 15,000 g [16]. As the acceleration increases beyond this value the non-dimensional thrust begins to decrease. It should be noted that the acceleration parameters indicated in Fig. 19 apply only at the Mach number corresponding to peak thrust; at higher or lower Mach numbers the thrust is lower and hence the unsteady effects of projectile acceleration are also decreased. Thus, regardless of the level of peak acceleration, the unsteady model predicts that the thrust of the thermally choked propulsive mode goes to zero at the propellant’s CJ speed. In addition, it is evident that as the fill pressure of the ram accelerator is increased, the thrust increases when the fill pressure is increased but it does not follow the latter proportionally as predicted by the ideal-gas quasi-steady model.

Fig. 19

Projectile acceleration effects on thermally choked ram accelerator operation at high pressures, compared to ideal quasi-steady case

2.1.2 Baffled-Tube Ram Accelerator

An alternative approach to increasing acceleration by the use of high fill pressure is to develop a means to achieve ram accelerator operation in very energetic propellants. It has been empirically determined that the heat release of CH₄/O₂/N₂ propellant mixtures must be reduced to ~1/3 of the maximum available without diluent to effect stable ram accelerator operation, which limits the peak thrust to a value lower than the theoretical maximum. Using a propellant with a greater heat release results in the undesirable situation in which the driving combustion wave surges past the projectile, causing a diffuser “unstart” [43, 44]. In addition, the unstart phenomenon also limits the lowest Mach number at which the ram accelerator process can be initiated (e.g., Mach 2.5 for the nominal ratio of projectile throat diameter to tube bore diameter of 0.76), which puts a much bigger onus on the muzzle velocity capability of the pre-launcher for applications that require massive projectiles. Thus, the challenge for generating high ram accelerator thrust at low fill pressure is to devise a means to allow reactive propellant to be ingested by the projectile throat while keeping the combustion-driven compression waves from propagating forward through the throat, i.e., some form of “one-way valve” is needed.

A novel concept of using baffles on the wall of the tube, shown in Fig. 20, was proposed by Higgins [42] to enable ram accelerator operation in the most energetic of propellants. Baffles, or annular rings, attached to and/or machined into the tube wall act to isolate the combustion process behind the projectile from the intake of unburned propellant past the conical nose of the projectile. This isolating effect allows more energetic mixtures to be used without the risk of the combustion driving a shock wave upstream of the projectile throat and causing an unstart. Since the baffles act to contain the combustion behind the projectile, the tube-to-projectile-throat area ratio can be increased, allowing successful starting of the ram accelerator at as low as Mach 2 without unstart. The use of more energetic propellant, a greater tube area, and operation at lower Mach number all act to increase the thrust on the projectile without having to increase the propellant fill pressure. In addition, the projectile now rides on the baffles, eliminating the need for fins to center it in the tube.

Fig. 20

Baffled tube ram accelerator flow field schematic. Baffles prevent combustion pulsations from being driven ahead of projectile

The baffles have a hole bored through their centers that is just large enough to allow the passage of the projectile. The spacing of the baffles is such that the cylindrical mid-body of the projectile completely spans at least two baffles at any time. This forms a sequential series of propellant chambers down the bore of the tube, as shown in Fig. 20. The propellant is initially ignited behind the projectile and the combustion process raises the pressure at its base and in the annular chamber around the mid-body. The baffles act as one-way valves whereby propellant can be ingested by the supersonic diffuser of the projectile, yet the combustion-driven pressure wave system cannot be pushed upstream. Consequently, the propellants can be formulated to be as energetic as possible to maximize acceleration. The ultimate velocity limitation of this concept occurs when the strength of the precursor shock wave, generated by the leading edge of the projectile shoulder as it just enters a chamber, is sufficient to directly initiate a detonation wave that can travel upstream through the next chamber before the projectile shoulder seals against the next baffle. Thus, the thickness of the baffles, their spacing, and the volumes of the expansion chambers all play a significant role in the application of this concept.

Preliminary experiments in a 1-m-long, 38-mm-bore baffled tube, with a propellant fill pressure of 20 bar, have proven that axisymmetric projectiles can be accelerated in propellants having twice the maximum heat release able to be used for ram accelerators operating in smooth bore tubes [45]. The test section, projectile and representative velocity-distance data from these experiments are shown in Fig. 21. The corresponding non-dimensional thrust was twice that ever generated with propellants at this pressure. The peak operating Mach number of the baffled-tube ram accelerator has not yet been experimentally determined due to the short length of the test section. Experiments are planned for a 4-m-long baffled tube in which more of its thrust characteristics and maximum operational Mach number can be explored.

Fig. 21

One-meter-long baffled tube test section (upper left), axisymmetric projectile (upper right), and experimental velocity-distance data (left). Projectiles accelerated through more than 0.50 m of tube length

2.2 U.S. Army Research Laboratory (ARL)

At the U.S. Army Research Laboratory a ram accelerator program was pursued from 1991 to 1997. The main thrust of the work was to demonstrate operation at a larger scale (120-mm-bore) (see Fig. 22), increase its velocity capability [60–62, 64, 67, 68], and develop computational fluid dynamics (CFD) codes for improved predictive capabilities [67, 85–91]. Projectiles of 5-kg-mass were accelerated to velocities up to 2 km/s in a two-stage configuration. The experimental research also included flow visualization of the thermally choked propulsive mode in sacrificial 1.83-m-long transparent acrylic tubes [63]. The results were recorded using high-speed cinematography. Figure 23 shows frames from two different runs. In Fig. 23a the projectile had already accelerated through a 9.4-m-long steel accelerator section with the same propellant mixture and pressure. It is evident that the combustion zone enveloped the aftbody of the projectile well behind the throat. Figure 23b shows a projectile as it entered the accelerator (injected at ~1200 m/s) and started (in this case the accelerator consisted only of an acrylic tube). The frame was taken near the end of the transparent tube, where the combustion stabilized; note the “discarded” obturator at the right.

Fig. 22

120-mm-bore ram accelerator at ARL (low pressure version). The initial launcher was a M256 120-mm tank gun, and the ram accelerator section consisted of three such gun barrels, each 4.7 m long. Another gun barrel (perforated) was used as a vent section between the launcher and the ram accelerator

Fig. 23

Frames from high speed films of 120-mm projectile accelerating through transparent acrylic tubes [63]. a Projectile cruising near end of transparent tube. Propellant is 3CH₄ + 2O₂ + 10N₂ at 51 bar, and projectile velocity is 1,480 m/s (Mach 4.1). b Projectile just after entering ram accelerator (note obturator at right). Propellant same as above but at 20 bar; projectile velocity is 1,300 m/s (Mach 3.6)

In 1997 the 120-mm-bore ARL facility was upgraded with a new test section, consisting of a single, constant diameter (324-mm-O.D.), thick-walled tube having a length of 4.57 m, which was designed to operate at fill pressures up to 100 bar [62–64]. This facility was initially not equipped with a vent tube between the gun and the ram accelerator test section. Starting the ram accelerator without venting at high pressure was found to be more sensitive to initial conditions than with venting, and was not successful even though a previous test without venting (in the lower pressure rated tubes) was successful. Reinstalling the vent tube allowed a partially successful test at 102 bar fill pressure to be conducted in the new facility [65]. The ram accelerator program at ARL was discontinued shortly after its principal investigator (Kruczynski) left for private industry (see below).

2.3 UTRON, Inc

Low velocity ram accelerator starting was investigated by Kruczynski at UTRON, Inc., Manassas, VA, in collaboration with the UW group [69]. Starting at entrance velocities as low as 760 m/s using various combinations of CO₂ diluent levels and obturator configurations in stoichiometric CH₄/O₂ mixtures was successfully demonstrated. However, the combustion could not be stabilized beyond 2.5 m of travel or 1000 m/s under these conditions. To determine the potential upper operating limit of CO₂-diluted propellants a second CH₄/O₂/CO₂ fueled stage was added and the projectile, entering with established ram combustion, was able to accelerate an additional 3.5 m down bore and up to 1150 m/s. At this velocity the projectile can easily make a transition into propellants having higher sound speeds for continued acceleration. A mechanical pyrotechnic ignitor (Fig. 24), based on one originally developed at the University of Washington [37], was successfully tested and showed the capability to ignite previously un-ignitable propellant mixtures at low entrance velocities [69]. In this case, starting of the ram accelerator was achieved without the presence of a normal shock on the aftbody of the projectile. Such techniques offer the ability to extend the starting regime of ram accelerators further.

Fig. 24

Diaphragm-impact-initiated onboard igniter developed at UTRON, Inc. note arrangement of primer, firing pin, and primer driver to move forward on impact with the diaphragm [69]

3 Computational Modeling

Beginning shortly after the ram accelerator was first conceived, various researchers began to engage in computational modeling of the ram accelerator, both in the U.S. and elsewhere, in support of the experimental efforts. In the U.S. such work was carried out by Brackett and Bogdanoff [10], Yungster [110], Yungster et al. [112], Kruczynski and Nusca [67], Yungster and Bruckner [111], Soetrisno and Imlay [101], Burnham [26], Hinkey et al. [46], Hinkey et al. [47], Li and Kailasanath [72–74], Li et al. [75–77], Nusca [85–89, 92] and Soetrisno et al. [102, 103]. In other countries computational work has been done by Choi et al. [31–33], Henner et al. [36], Leblanc and Fujiwara [71], Moon et al. [81], Taki et al. [107], Zhang and Taki [113], Liu et al. [78] and Bengherbia et al. [6]. It is beyond the purpose or scope of this paper to review in detail the numerous contributions in this area of ram accelerator research; the topic is worthy of a separate review paper of its own. The interested reader is directed to the listed references for additional information. Here only a very brief overview is provided of some of the more recent work performed in the U.S.

The bulk of the more recent CFD modeling of the ram accelerator in the U.S. has been conducted primarily by Li et al. and Nusca (see above-cited references). In particular, Li et al. have performed CFD analyses related to the UW 38-mm ram accelerator [72–75]. Cases they have studied include high-pressure operation, the starting process, unstarts, wave fall-offs, and aerodynamic stability of projectiles, among others. Figure 25 shows computational results for a 2-D projectile canted at 1.5° counterclockwise from the tube axis, travelling at 1250 m/s in a stoichiometric hydrogen-air mixture at 25 bar; the projectile Mach number under these conditions is 3.01 [75]. The pressure, Mach number, temperature, and water vapor concentration fields are plotted in the figure, as are the pressure distributions on the upper and lower surfaces of the projectile. It was found that the aerodynamic torque stabilizes the projectile if the normal shock is maintained on the rear part of the projectile by the thermally choked combustion, and destabilizes the projectile if this normal shock is absent. It remains to be seen if this is also the case in a 3-D computational model.

Fig. 25

Computational results for canted projectile in 38-mm UW ram accelerator. Propellant mixture is stoichiometric H₂/air, velocity is 1250 m/s, canting angle is 1.5˚ [75]

CFD modeling with a virial equation of state by Nusca [87] was focused on experiments in the 120-mm-bore ram accelerator at ARL. Viscous, 3-D flow was considered with frozen chemistry to examine the impact of fins on the flow field and the corresponding pressure profiles of axisymmetric projectiles. The influence on tube wall pressure measurements of fin orientation relative to pressure transducer location correlated well with experiment (Fig. 26a) and matched previous experimental results of Hinkey et al. [46]. Axisymmetric pressure profiles indicated reflected shock wave systems similar to those determined along the center-line between fins, but at a much lower strength. Quasi-steady flow modeling with axisymmetric projectiles and three-step finite rate chemistry was carried out to provide a more detailed look at how the combustion process affects the flowfield over a range of fill pressure (5–10 MPa) and Mach number (3.3–4.1) conditions. Even though the fins were not present in these calculations, the correlation between CFD predicted projectile base pressure and experiment was remarkably good (Fig. 26b). Unsteady calculations with axisymmetric projectiles and three-step chemistry with an obturator initially at the base of the projectile predicted velocity-distance profiles that agreed to within 3 % of experiment (Fig. 26c). The actual thrust prediction at any given Mach was actually even better, as evident by the similar slopes of the theoretical and experimental V-x profiles.

Fig. 26

Upper left 3-D wall pressure calculations with inert propellant. Upper right wall pressure calculation for axisymmetric projectile with 3CH₄ + 2O₂ + 10N₂ propellant. Left time-accurate velocity-distance profiles for axisymmetric projectile with obturator initially at its base in same propellant as above [87]

4 Future Work

Current areas of interest that continue to be investigated include operation of the ram accelerator at elevated pressures, investigation of projectile material and geometry effects, modeling of ram accelerator operation in three different velocity regimes with real gas and unsteady effects, studies of the starting dynamics, especially at high pressures and/or low initial projectile velocities, investigations of the superdetonative propulsive modes (to attain the highest possible velocities), and improved CFD modeling of all the effective propulsive modes of the ram accelerator. The baffled tube ram accelerator is being studied experimentally and computationally as well, as it holds much promise for very high acceleration performance. The ultimate aim of all ram accelerator research is to attain the theoretical velocity capability of this launcher technology, namely 6–8 km/s, at which the most interesting applications, such as hypersonic aeroballistic testing and direct space launch become practicable.

5 Conclusions

The ram accelerator is a ramjet-in-tube hypervelocity launcher, originated at the University of Washington, that uses chemical energy to propel projectiles to very high velocities. A projectile similar to the centerbody of a supersonic ramjet travels through a tube filled with high pressure combustible gas, which burns on or behind the projectile to provide thrust. The ram accelerator has demonstrated successful operation at a variety of operating conditions and scale sizes, in a wide range of gas mixtures. Propellant fill pressures of 5–200 bar, and velocities up to 2.7 km s⁻¹ have been attained with a bore size of 38 mm, while velocities of up to 2.2 km/s have been achieved with 120-mm caliber, 5-kg projectiles operating in 100-bar propellant mixtures. Three velocity regimes, centered about the Chapman-Jouguet detonation speed of the propellant gas mixture, have been identified that exhibit different acceleration characteristics, indicating the existence of several different propulsive cycles: subdetonative (characterized by thermal choking behind the projectile), transdetonative (characterized by the forward motion of the combustion process onto the projectile body and the existence of regions of mixed supersonic and subsonic combustion), and superdetonative (in which a reflected oblique shock wave induces the combustion process to occur entirely on the projectile body). Numerous computational fluid dynamic studies of the ram accelerator have also been carried out, and have shown that at fill pressures above 25 bar both real gas and non-steady flow effects significantly influence the flowfield analysis. To make better use of the full heating value of propellants and allow higher performance at lower fill pressures, a baffled-tube ram accelerator concept has been proposed and successfully tested at the University of Washington. Ram accelerator facilities have been built and successfully operated by several research groups around the world; of these, only one, at the University of Washington, is currently active (the other ram accelerator facilities are not currently operational due to a variety of factors, mostly involving fiscal constraints). The ease with which the ram accelerator can be scaled up in size offers unique opportunities for its use as a hypersonic research tool, a potentially low-cost space launcher, and other interesting applications. Progress in these areas is predicated by the further development of the velocity and scaling capabilities of this innovative launcher technology.