Recent Advances in Detonation Techniquesfor High-Enthalpy Facilities
Frank K. Lu,* Donald R. Wilson, f W. Scott Stuessy*University of Texas at Arlington, Arlington, Texas 76019
Robert J. Bakos,§ and John I. Erdos^GASL, Ronkonkoma, New York 11779
Detonations can be used to generate a high-pressure gas of high acoustic speed to drive ashock tube. Recently, detonation-driven facilities have been Implemented for meaningfulhypervelocity testing. These facilities can be operated with the detonation wave propagat-ing downstream or upstream. The advantages and problems associated with these methodsare discussed. In addition to a performance comparison between these two modes, com-parisons with other high-performance techniques, such as free-piston and gun tunnels, isalso made. At present, detonation-driven facilities are generally of lower performance thanfree-piston tunnels. However, they appear easier to operate.
Subscriptso = stagnation conditions1, 2, 3, ... = regions in wave diagram corresponding
to different gas statesCJ = Chapman-Jouguets = shock
IntroductionA ERODYNAMIC testing using short duration facil-
•**- ities is a highly specialized technique confinedmostly to hypersonic and hypervelocity regimes withtheir high enthalpy requirements. The key character-istic of these facilities is that they are derived from
*Associate Professor and Director, Aerodynamics ResearchCenter, Mechanical and Aerospace Engineering Department. As-sociate Fellow AIAA.
fProfessor and Chairman, Mechanical and Aerospace Engi-neering Department. Associate Fellow AIAA.
^[Faculty Research Associate, Aerodynamics Research Cen-ter, Mechanical and Aerospace Engineering Department. SeniorMember AIAA.
{Principal Scientist. Member AIAA.§Vice President. Associate Fellow AIAA.
shock tube principles. Consequently, these facilitieshave short test times, typically in the 0.1-10 ms range.This short test time contrasts drastically from thoseachievable using continuous or blowdown (that is. "con-ventional") facilities.1 Despite this limitation, short du-ration facilities appear to be the primary means forachieving hypervelocity flows at present. The under-lying principle is to store energy over a long period oftime, lessening the input power requirement, and thenreleasing the accumulated energy rapidly. The trade-off between flow duration and flow enthalpy is clearlymanifested.
The alternative to impulse facilities may be extremelyprohibitive in cost, especially if large conventional tun-nels are contemplated. Additionally, one may even ar-gue that for certain purposes, there is no other reallyviable alternative, such as in simulating orbital flight orflight through planetary atmospheres.2 The latter ap-plication exploits the fact that impulse facilities can usetest gases other than air. Moreover, elaborate thermalprotection of the test model or of the instrumentationcan be dispensed with because of the short time of pas-sage of the heated test gas. Heat-transfer studies withappropriate cold-wall boundary conditions encounteredin flight can be performed using short duration facili-ties. Other than aerothermodynamic applications, im-pulse facilities can be used for instrumentation and fa-cility development.
The most serious limitation of short duration facil-ities is the test time. This difficulty has, fortunately,been greatly reduced with modern, high speed, dataacquisition systems and innovative techniques. Never-theless, a proper appreciation of the limitations of shortduration facilities is necessary to enable its advantagesto be effectively exploited.
Present interest in high-enthalpy facilities stems fromthe need to test advanced air-breathing hypersonicpropulsion systems. In particular, the facility shouldbe capable of providing post bow shock conditions
for testing concepts such as oblique detonation wavepropulsion.4 Recent proposals for achieving high streamenthalpies considered radiative energy addition to a su-personic flow5 or the incorporation of a magnetohydro-dynamic accelerator.6 Although proposed for continu-ous or blowdown operation, the feasibility of these fa-cility advances can be studied using shock tubes.
In view of the enthalpy requirements for hyperveloc-ity testing, the shock tube must incorporate a high per-formance driver. Warren and Harris7 classified highperformance drivers as (a) conventional drivers usinghigh pressure and high acoustic speed gases, (6) electricand magnetic field interaction drivers, (c) detonation,that is, explosive drivers, and (d) those that derive theirflow characteristics from the coupling of energy addi-tion and wave processes in an unconventional manner.Examples of conventional drivers, which are well de-veloped, include electrical energy discharge, internal orexternal heating of hydrogen or helium driver gas, andpiston compression. The last technique is found in freepiston8 and gun tunnels.9 Electric and magnetic fieldinteraction drivers, and detonation drivers appear tobe less well developed. The final class of techniques in-cludes the expansion tube2 and the use of shock wavesor detonation waves to achieve a high pressure, hotdriver gas.10
Although Warren and Harris listed many high per-formance techniques, only a few of these appear fea-sible. The free piston technique appears to be themost developed and have been implemented in differ-ent institutions, including the largest such facility atGottingen, Germany.11 The free piston technique isgenerally thought to be capable of achieving extremelyhigh enthalpies. However, there is recent interest inusing detonation techniques to achieve high enthalpies,albeit somewhat lower than that achievable by free pis-ton techniques. Nevertheless, detonation techniquespossess favorable features to ensure that they occupya useful niche in hypervelocity testing. They are thesubject of this review. To understand the use of deto-nation drivers, a brief summary of shock tube principleswill be given next.
Background
Shock Tube Principles
A shock tube, shown schematically in Fig. la, consistsessentially of a high pressure, driver section separatedfrom the low pressure, driven section by a diaphragm.Gases fill both sections. When the diaphragm is rup-tured, the high pressure gas, at an initial state 4, ex-pands into the low pressure section, filled with gas at aninitial state 1. Propagating into the driven gas ahead ofthis expansion is a shock wave. This shock compressesthe driven gas, thereby changing its state from 1 to 2.This slug of shock processed, driven gas—between thepropagating shock and the arrival of the contact surface
Diaphragm
I® Driver tube! Driven tube (j) j
I *--;;; (3) ! ® ! -•-*• I- - . - -
Expansion Contact Primarywave surface shock
a. Schematic.
;;:::: Expansion• •V - : Wave
Primary Particle xshock path
b. Wave diagram.
Figure 1: Ideal wave processes in a shock tube.
between the driver and driven gas—forms the test gas.(The test gas can be further processed; for example, itcan be accelerated by a nozzle, as in a shock tunnel.)By using different driver gases or by having the driverand driven gases at different initial pressures and tem-peratures, the shocked state of the driven gas can bepreset. Quasi one-dimensional theory yields the follow-ing implicit expression for the pressure ratio of the testgas after shock passage as
(1)
The key parameter governing shock tube performance isthe shock Mach number Ms, defined as the shock speedreferred to the initial speed of sound in the driven gas.The shock Mach number is given by
Ms = (2)
Although actual shock tube processes are complicatedby non-ideal behavior, such as boundary layer growthand finite diaphragm opening times, the technique hasbeen sufficiently developed for shock tube performanceto be well understood. Shock tubes are used in manyscientific disciplines and these tubes have been built ina broad range of sizes, with a bewildering number ofenhancements.
The theoretical performance of a shock tube is dis-played in Fig. 2, where helium is the driver gas andair is the driven gas. Both gases are assumed perfectbecause real gas effects in the test gas up to a shockMach number of 8 are not large. In the figure, the ini-tial driver-to-driven pressure ratio p±/pi is plotted asa function of Ms for different values of a^/ai- Alsoplotted as a dotted line is the pressure ratio across thepropagating shock pz/Pi-
Figure 2: The effect of acoustic speed and initial driverpressure ratio (helium driver gas and the air test gas).
Figure 2 shows that a cold driver, with a low sonicspeed, achieves a practical shock Mach number of onlyabout 3. This limited shock Mach number led toschemes to improve driver performance in order to pro-vide adequate hypervelocity, real gas simulation. Thefigure shows that to achieve large shock Mach numberswithout incurring an exorbitant initial pressure ratio,the sonic speed of the driver gas must be raised. Inother words, the driver gas must be of low molecularweight and must be heated. For example, to achieve ashock Mach number of 8, the theoretical initial pressureratio needs to be only 1 300 and 260 for acoustic speedratios of 5 and 10 respectively. In practice, using a valueof pi = 10 kPa for instance, Ms = 8 can be achievedwith a value of p± of only 2.6 MPa for an acoustic speedratio of 10. (Such a pressure can be obtained withoutusing highly specialized industrial equipment.) Unfor-tunately, if the air is at room temperature, the heliummust be at a temperature of 3 500 K. The only practi-cal means of achieving this high temperature withoutdestroying the facility is to heat the gas in a transientmanner.
Detonation Wave Drivers
In ground test applications, the free piston techniquefor producing a hot driver gas is well-developed. Nev-ertheless, detonations can also be used to obtain a hotdriver gas with a large acoustic speed. Detonationdrivers are an inexpensive, simple, viable alternativeto free-piston drivers in obtaining high enthalpies be-
cause the technique dispenses with a fast moving, heavypiston. Detonations are also attractive compared tocombustion as a transient, heating method.12 The lowvelocity combustion depends on factors such as igni-tion energy, number of igniters, initial turbulence, flamepropagation and size of tube, all of which cause diffi-culties in reproducibility.13
The detonation-driven shock tube was first proposedby Bird,14 and was subsequently studied by severalinvestigators.15"17 This concept has recently been fur-ther developed.10'13'1 ~ A detonation process is typ-ically established in a driver tube filled with a near-stoichiometric oxyhydrogen mixture, although othergas combinations, such as acetylene and oxygen.1' arepossible. The mixture pressure can be quite low, thuseliminating the need for thick metal diaphragms. Thedetonation process produces a low molecular weightdriver gas at high temperature and pressure, all ofwhich are desirable features. The sudden pressure riseproduced by the detonation causes the primary di-aphragm to rupture, thus establishing a shock wave inthe driven tube.
Implementation
At present, detonation driven facilities have beenreported by the University of Technology, Aachen,Germany13'23 the Institute of Mechanics, Chi-nese Academy of Sciences,18"22 GASL, Ronkonkoma.New York24'25'27 and the University of Texas atArlington.28'29 These facilities use an oxyhydrogen mix-ture as the driver gas, with helium dilution as necessary.Table 1 summarizes the major characteristics of the det-onation drivers in the four institutions mentioned, in-dicating the propagation mode of the detonation wave.The detonation wave can propagate either downstreamor upstream. Thus, the mode of propagation can serveas a means of classifying detonation wave drivers.
In the downstream propagation mode, the ignitionsource is located at the upstream end of the driven tube,producing a detonation wave that propagates down-stream. The main wave processes are shown schemati-cally in Fig. 3. The detonation wave, denoted as a solidline, propagates downstream into region 4. Momentumof the burned gas following the detonation wave is alsoin the downstream direction and produces a strongershock for a given detonation overpressure. However,the detonation wave is followed by Taylor rarefaction,shown as chained lines in region 4', which progressivelyattenuates the strength of the driven shock. The det-onation wave propagating into 4 is reflected at the di-aphragm, to yield an effective, unsteady condition givenby 4". The reflection of the detonation wave at the di-aphragm interface generates a shock which propagatesinto region 1, being driven by the high enthalpy det-onation products at state 4". The burned gas thenexhausts into the driven section to reach state 3. Thisgas is separated by an interface, shown in Fig. 3 as a
dotted line, from the post-shock driven gas at state 2.Further wave processes are not shown but may be im-portant, such as in creating a high pressure, stagnantregion 5.
An example of detonation tube pressure traces for thedownstream propagation mode is shown in Fig. 4.29 Thedetonation tube was filled with a stoichiometric oxyhy-drogen mixture at 6 atm. Transducer (3) was closer tothe igniter than transducer (4) and thus recorded anearlier arrival of the detonation wave. Time-of-flightcalculations indicated that the detonation wave eventu-ally reached CJ velocity. The pressure records showeda rapid pressure rise. However, the transducers wereunable to resolve the "von Neumann spike," the max-imum pressure of the detonation front. The recordsshow a rapid pressure drop associated with the Tay-lor rarefaction, as indicated in the figure. The Taylorrarefaction reduces the effective pressure pumping thedriven gas to below the Chapman-Jouguet pressure atthe trailing edge of the detonation front. There is areduction in shock tube performance. The records alsoshow peaks within the rarefaction region. These peaksare attributed to waves reflecting off the the diaphragminterface due to the practical limitation of having toplace the igniter slightly away from the diaphragm po-sition.
Figure 4: Pressure traces for downstream propagationmode (stoichiometric oxyhydrogen mixture at initialpressure of 6 atm, electric arc ignition).
Figure 3: Downstream propagation mode.
In the upstream propagation mode, the ignitionsource is just upstream of the primary diaphragm, pro-ducing a detonation wave that propagates upstream.The detonation tube process is shown schematically inFig. 5. The detonation wave propagates to the left intoregion 4. The pressure rise following the detonationwave is fairly constant but the momentum imparted tothe driver gas by the detonation wave is directed up-stream. This has adverse loading consequences on theshock tube.
An example of detonation tube pressure traces for theupstream propagation mode is shown in Fig. 6.29 The
Figure 7: Downstream propagation mode using shockinduced detonation.
detonation tube was filled with a stoichiometric oxyhy-drogen mixture at 6 atm. Transducer (4) was closer tothe igniter and thus recorded the arrival of the detona-tion wave before transducer (3). However, transducer(4) detected some precompression ahead of the detona-tion wave, indicating that the detonation wave did notreach the CJ velocity, a phenomenon also reported inRef. 30. Transducer (3) as well as subsequent trans-ducers upstream (not shown in Fig. 6 for clarity) didnot detect any precompression. Time-of-flight calcu-lations indicated that the detonation wave eventuallyreached CJ velocity. The pressure records indicateda rapid pressure rise after the precompression but thevon Neumann spike was not resolved. The Taylor rar-efaction and additional peaks in this region are alsoindicated in the figure.
For either propagation mode, further performance en-hancement is possible by helium dilution to the oxyhy-drogen mixture. Helium dilution raises the sonic speedin the driver gas, and also somewhat reduces the dangerassociated with premature detonation of the oxyhydro-gen mixture. Performance calculations by Yu, et al.18
indicate that the performance degradation caused bythe slight lowering of the detonation temperature dueto helium dilution is more than adequately offset by theincreased sonic speed of the driver-tube gas.
Two of the institutions which have recently imple-mented detonation drivers make use of downstreampropagation while the other two make use of upstreampropagation (Table 1). Means of implementing thesetechniques to enable them to be used for test facilitiesare now elaborated separately in the following subsec-tions.
Downstream propagation
Downstream propagation using arc ignition28 yieldedshock speeds considerably lower than those predictedby a simple, one-dimensional model, thus yielding dras-tically lower pressure and temperature levels in state 2of the driven gas. The primary reason was attributedto the Taylor rarefaction wave associated with the arc-ignition process for the downstream mode.
The decrease in pressure following the detonationfront can be overcome by adding a driver tube aheadof the detonation tube, to initiate a shock-induceddetonation.31 A detonation wave is generated in thecombustible mixture by rupture of a diaphragm be-tween the driver tube and the detonation tube. Thismethod was apparently first used by Coates andGaydon.15 These authors made use of the shock wavefrom a cold hydrogen driver to ignite a detonablemixture. Recently, shock-induced detonation was re-introduced by Bakos et al.25 and adopted by Stuessy etal.29
A schematic of the method, including an idealizedwave diagram, is shown in Fig. 7. In this figure, the det-onation and driven sections are shown with different di-
Figure 8: Different modes for shock-induced detona-tion.
ameters. This causes a slightly more complicated waveprocess. Labeling of the different regions in the wavediagram remains consistent with shock tube nomencla-ture. A high-pressure air or helium driver is placedupstream of the detonation tube. When the primarydiaphragm is ruptured, a shock wave is driven into thedetonation tube, labeled region 100 in Fig. 7b. Thisshock wave quickly initiates detonation. Unlike theclosed-end operation (see Fig. 5), the driver tube servesto reduce or eliminate the Taylor rarefaction wave, thusresulting in a higher pressure available to drive the pri-mary shock wave in the shock tube. In effect, the drivertube exhaust acts like a "gas piston" to sustain the pres-sure behind the incident detonation wave at a higherlevel than would occur if the detonation propagatedaway from a closed end-wall.
The strength of the primary shock driven into region
1 can be attenuated continuously by the Taylor rar-efaction following the detonation wave. For example,the wave diagram in Fig. 7b shows a weak detonation.This yields an "under-driven" condition.25 In this case,the extent of the Taylor rarefaction is just sufficientto balance the pressure of the expanding driver gas.The uniform region of detonation products behind theTaylor rarefaction and ahead of the driver gas inter-face is labeled 400 and appears as a pressure plateauP40o in Fig. 8a. If the driver pressure is raised (that is.if p4/pioo is raised), a point will be reached such thatthe pressure of the expanding driver gas just balancesthat at the rear of the detonation wave PCJ, annihilat-ing the Taylor rarefaction. In this "perfectly driven"mode, the full CJ pressure level can ideally be main-tained behind the detonation wave, as shown in Fig. 8b.A further increase in pt/pwo causes the expanded drivergas pressure to be higher than the CJ wave pressure.This forces the detonation to travel faster than the CJspeed, resulting in an "over-driven" detonation. Taylorrarefaction also does not exist in this case.
Pressure traces of an under-driven and a nearlyperfectly-driven detonation are shown in Figs. 9 and10.29 The driver section was filled with air and heliumrespectively, with both gases at room temperature andat 210 atm. Also, for both these cases, the detonationsection contained a stoichiometric oxyhydrogen mixtureat room temperature and at a pressure of 1.5 atm. Thedriven tube was filled with air at room temperature andat a pressure of 0.14 atm.
27 28
/, ms
Figure 9: Example of an under-driven pressure trace.
In the under-driven mode shown in Fig. 9, the drop inpressure through the Taylor rarefaction wave is clearlyindicated. The next increase in pressure recorded bythe transducer is due to the arrival of the reflecteddetonation wave. In contrast, Fig. 10 shows a nearlyperfectly-driven pressure trace with almost annihila-tion of the Taylor rarefaction. Furthermore, the pres-
sure level achieved by the reflected detonation waveis much higher, with a corresponding increase in thedriven-tube Mach number from 6.70 to 7.65. Driven-tube pressure traces for these two cases are shown inFigs. 11 and 12. Approximately 1 ms of adequatelysteady pressure downstream of the propagating shockwas obtained for these two cases. Thus it appears thatthe shock-induced detonation mode offers substantialgains in performance by reducing or possibly eliminat-ing the Taylor rarefaction wave.
Figure 10:sure trace.
Example of a nearly perfectly-driven pres-
Figure 11: Pressure trace in driven tube correspondingto Fig. 9.
The downstream propagation mode of operation canbe optimized by examining the performance of thedriver and detonation sections. First, the pressure ratioP4/Pioo must be sufficiently high to quickly initiate det-onation. It has been suggested that detonable mixtureswith diluent fractions up to 65 percent are sufficiently
Figure 12: Pressure trace in driven tube correspondingto Fig. 10.
sensitive for achieving detonations at pressure ratios aslow as 20.27
The choice of the driver-to-detonation tube fill ratiop^/Pioo and the composition of the detonable mixturecan be optimized to yield the highest-post shock pres-sures in the shock tube and maximum steady flow time.Unlike a conventional shock tube driver, the detonationproducts in this driver system propagate downstream.Consequently, the strength of the driven shock wavedepends on the propagation velocity in addition to thesound speed and pressure in region 400. These threeparameters depend on the initial tube fill ratio and thecomposition of the driver and detonable gases.
To simplify the optimization, an "effective" pressurepe and sound speed ae can be defined for the propagat-ing detonation products. The effective values are thoseof a conventional static driver that would deliver thesame shock strength in a given shock tube. They arerelated to the actual values in region 400 by
Oe = P400- 1
0400 1 +7-1
M,400
(3)
(4)
The above equations show that the downstream propa-gation velocity augments the pressure and sound speedrelative to the static values.
For reflected shock tunnel operation, the optimiza-tion proceeds by choosing the effective sound speed thatwill tailor the driver-test gas interface in order to yieldmaximum test time. Then the effective pressure is max-imized relative to the peak pressure that occurs in thedriver during the operating cycle. Naturally, it is thispressure which must be maintained within the designpressure limit of the driver vessel. Referring to Fig. 8,depending on the initial fill pressure ratio, the peak
pressure will be either the initial driver fill pressure, orthe pressure immediately following the CJ wave.
Figure 13 is an example from the optimization pro-cess for a particular stoichiometric oxyhydrogen mix-ture with 30 percent argon diluent. The driver gas ishelium. The effective pressure of the detonation prod-ucts, normalized by either the driver pressure or theCJ pressure (whichever is higher), is shown as a func-tion of the fill pressure ratio. At the lowest fill pressureratio, the driver does not push on the detonation prod-ucts and provides no forward velocity. This yields aneffective pressure that is less than 40 percent of thepeak pressure behind the CJ wave. At the highest fillpressure ratio shown, the driver pressure is sufficientto maintain the pressure behind the CJ wave constant;however, the effective pressure achieved is only 35 per-cent of the fill pressure. In between these extremes, ata fill pressure ratio of approximately 20, the normalizedeffective pressure reaches a maximum of 75 percent. Be-cause the pressure ratio generated by the CJ wave forthis mixture is also approximately 20, optimum per-formance occurs when the drive fill pressure and thepost-CJ pressures are equal.
Figure 13 shows that the effective sound speed of thedetonation products changes slowly with the fill pres-sure ratio for a given composition. This effective soundspeed needs to be considered for tailored shock tunneloperation. Of note is that the sound speed of the det-onation products is greater than that of the ambienttemperature helium driver gas, indicating that the det-onation technique allows tailored operation at higherenthalpy than does the helium driver alone. This isconsistent with the aims of high performance drivertechniques as elaborated earlier.
Initiation tubeDiaphragms ^ Diaphragms
0.8
0.6
0.4
0.0
1.25
1.20
1.15
1.10
1.05
1.00
fJPit
Damping Detonation Drivensection section section
a. Schematic of detonation-driven shock tube in up-stream propagation mode.13
Figure 13: Dependence of normalized pe and ae on thelight-gas driver to detonation fill pressure (4.67H2 +2.33O2 + 3Ar).
Upstream propagation
As highlighted previously, the upstream propagation
Dampingtube
Detonation | Drivensection section
Diaphragm Diaphragminterface interface
b. Wave diagram.
Figure 14: Shock tube with detonation driver, and ad-ditional initiation and damping tubes.
mode provides a fairly steady supply pressure for driv-ing the shock wave into the driven tube. However,shock reflection at the closed end of the driver tubeproduces a structural problem. The reflected wave pres-sure may exceed 200 times the initial driver pressure.32
For initial driver pressures of 10 MPa or more, the highvalue of the reflected wave pressure may destroy thefacility. This can be overcome by adding a "damp-ing tube" to the upstream end of the driver tube.13
A schematic of the test facility and an accompanying,ideal wave diagram are shown in Fig. 14. The dampingtube is separated from the driver tube by a light di-aphragm. Figure 14b indicates that an upstream prop-agating detonation wave would break this diaphragmand continue to propagate into the damping tube. Thehigh pressure that would otherwise occur due to shockreflection is well attenuated. The damping tube ab-sorbs the shock loading and helps to reduce the struc-tural load. Moreover, the damping tube increases thetest time if the tube is used to supply a shock tunnel,as is evident from the wave diagram of Fig. 14b.
The damping pressure and damping gas can be opti-mized to minimize the endwall loading. The wave struc-ture in the damping tube depends on the ratio of the
CJ pressure to the initial pressure in the damping tube.The predicted peak endwall pressure is shown in Fig. 15and the value averaged over 30 ms is shown in Fig. 16.The figures show that air is suitable as the dampingmedium. Calculations indicate that pi needs to be noless than pcj x 10~4 or, equivalently, 0.002^4. Tailor-ing to maximize the test time, unlike conventional shocktubes which is achieved by a unique value of p4/pi, isachieved by adjusting the oxyhydrogen ratio.
Figure 15: Peak endwall pressure of the damping tubeas a function of
Figure 16: Mean endwall pressure of the damping tubeas a function of pcj/Pr-
Finally, Yu et al. found that helium addition in gen-eral does not improve the operation of the detonationtube. They used a rich mixture of hydrogen and oxy-gen. This mixture is ignited by the detonation of acombustible gas in an initiation tube which breaks thediaphragm and starts a detonation wave moving left-wards. Normal shock tube behavior occurs but withCJ conditions behind the detonation wave until thereflected detonation wave interacts with the reflectedshock wave from the end of the driven section. Thereflection of the detonation wave can be influenced bythe choice of the gas and the pressure in the dampingtube.
Ignition
There are two methods of ignition.33 a slow or "self"
initiation by a low energy source and a fast or "direct"initiation by an energetic source. In self initiation, a de-flagration front transitions to a detonation. The under-lying mechanisms for deflagration-to-detonation (DOT)remain unclear and the DOT length can vary substan-tially depending on how well the gases are mixed andother factors. Hence, this process is not considered vi-able for facility development because it may not yieldreproducible conditions.
Development of a viable technique for detonationdrivers centers on direct initiation where the detonationforms almost instantaneously at the immediate vicinityof the igniter. The ignition process is crucial to thesuccess of detonation drivers. In general, the ignitionenergy must exceed a certain threshold. This criticalamount of energy for direct initiation depends on themixture ratio for given pressure and temperature, andigniter. Its minimum occurs at stoichiometric ratios.However, Yu and Zhao21 suggested that a gain in shockMach number can be achieved with a hydrogen-richmixture up to a H2/02 ratio of 5 at high initial pressureP±\ ~ 100. Obviously, an energetic igniter is necessaryin such a situation.
Methods used for direct initiation include the use ofan exploding wire, electric arc ignition (sparks), ex-plosives, lasers, and shock-induced detonation. Lee17
exploded a 75 p,m copper wire near the face ofa diaphragm to initiate detonation in an equimolaracetylene-oxygen mixture at atmospheric pressure. Ashock Mach number up to about 8 was obtained at apressure ratio P5i of 50. Higher shock Mach numberswere thought possible by increasing the driver pressure.Yu and Zhao21 found that sparks and exploding wiresyielded weak detonations. Instead, they directly initi-ated detonations in oxyhydrogen mixtures using a 20mg tetryl pellet. However, there was significant con-tamination and erosion of the facility. Stuessy et al.28
used an energetic arc source capable of delivering 20-25J in 60 /js.
In shock-induced detonation, the shock is gener-ated either by using an auxiliary driver filled withhigh pressure gas such as air, nitrogen, helium,26'29 orhydrogen,35 or it is generated by a detonation usingan "initiation tube," or pre-igniter.13'21 Shock-induceddetonation appears to be the most attractive for thedownstream propagation mode, with the potential ofminimizing the adverse Taylor rarefaction.
Comparison of High Performance Facilities
The fundamental difference between the upstream anddownstream modes is the direction of the detonation-induced velocity relative to the detonation wave prop-agation direction. To assess the effect of this differenceon performance, the effective static driver concept canbe used. For the upstream mode, the effective driver
Performance of the two cycles, with the same initialconditions as in Fig. 13, is compared in Fig. 17. Forthe upstream mode, the peak cycle pressure is the CJpressure while for the downstream mode, the cycle isassumed optimized as shown in Fig. 13 such that theCJ pressure and the light-gas driver pressure are equaland represent the peak cycle pressure. The downstreammode delivers approximately double the pressure per-formance over the effective sound speed range of inter-est in comparison to the upstream mode. Note thatthe temperatures are greater than room-temperaturehelium.
io.i080 ">0 6
0.5
0.4
Room temperaiure_. helium
Downstream propagation mode
Upstream propagation mode
1200 1500
at. m/s2100 2400
Figure 17: Pressure recovery of upstream and down-stream modes.
Table 1 summarizes the detonation-driven facilitiesknown to the authors. The UTA facility is a shock tubeand is currently used for high-pressure plasma and com-bustion research. The other facilities are devoted toaerodynamics testing. Table 2 shows that, amongst thedifferent techniques for generating high-enthalpy flows,detonation-driven facilities occupy an important niche.Detonation-driven facilities produce enthalpies belowthat of free piston tunnels but above that of gun tun-nels. The enthalpies, when expressed in terms of thebinary scaling parameter, are in the range of the re-entry corridor.23 Even though the performance is be-low that of the free piston tunnel, detonation-drivenfacilities have a few attractive features. First, they arerelatively easy to operate, without the problems asso-ciated with a heavy piston, such as piston erosion andrebound. Also, there is no need for thick diaphragmsto contain high pressure gas. However, the safe han-
dling of large amounts of hydrogen must be carefullyconsidered.
Outlook and Conclusions
Recent developments in detonation drivers have indi-cated that high performance can be obtained for mean-ingful hypervelocity testing. There are two preferredways of implementing detonation techniques, namely,a downstream and an upstream detonation mode, de-pending on the direction of the detonation wave. Thedownstream propagation mode makes use of shock-induced detonation and can be operated as under-driven, perfectly driven or overdriven, this classifica-tion being governed by the subsequent, wave structure.Ideally, the Taylor rarefaction should be annihilated bythe high-pressure driver. In the upstream propagationmode, the Taylor rarefaction problem does not trulyexist. However, structural problems require the use ofa damping tube. The wave processes in the upstreampropagation mode also results in a potentially longerrun time than the downstream mode. Theoretical con-siderations indicate that a higher performance can beobtained by the downstream propagation mode.
The operation of detonation-driven facilities isthought to be simpler than that of free piston tunnels.However, the enthalpies that, are obtained are some-what lower. Nevertheless, detonation-driven facilitiescan make a useful contribution to the study of variousgasdynamics problems. There remains the potentialfor developing new ways of operating detonation-drivenfacilities with better understanding of detonations.'' "!'Finally, the possibility of hybrid drivers, combining pis-tons and detonations may open up a new class of high-performance facilities.
Acknowledgements
The authors would like to thank Professor Hans Gronigof the University of Technology, Aachen, Germany andProfessor Hongru Yu of the Institute of Mechanics, Chi-nese Academy of Sciences, Beijing, China, for gener-ously sharing their papers and results with us.
References
'Trimmer, L. L., Gary, A. M., Jr. and Voisinet, R. L.."The Optimum Hypersonic Wind Tunnel," A1AA Paper 36-0739, 1986.
2 Morgan, R. G., "A Review of the Use of Expansion Tubesfor Creating Superorbital Flows," A1AA Paper 97-0279.1997.
3Chinitz, W., Brdos, J. [., Rizkalla, O., Anderson, G. Yand Bushnell, D. M., "Facility Opportunities and Associ-ated Stream Chemistry Considerations for Hypersonic Air-Breathing Propulsion," A! A A Journal of Propulsion andPower, Vol. 10, No. 1, 1994, pp. 6-17.
"Pratt, D. T., Humphrey, J. W. and Glenn, D. E., ''Mor-phology of Standing Oblique Detonation Waves," A1AA
Journal of Propulsion and Power, Vol. 7, No. 5, 1991,pp. 837-845.
'Miles, R B., Brown, G. L., Lempert, W. R, Yetter,R, Williams, G. J., Jr., Bogdonoff, S. M, Natelson, D.and Guest, J. R, "Radiatively Driven Hypersonic WindTunnel," AIAA Journal, Vol. 33, No. 8, 1995, pp. 1463-1470.
6Nelson, G. L. and Simmons, G. A., "Augmentation ofHypersonic Propulsion Test Facilities Using MHD," AIAAPaper 95-1937, 1995.
7 Warren, W. R. and Harris, C. J., "A Critique of High Per-formance Shock Tube Driving Techniques," in Shock Tubes.Proceedings of the 7th International Shock Tube Symposium,edited by I. I. Glass, University of Toronto Press, 1970,pp. 143-176.
BGai, S. L., "Free Piston Shock Tunnels: Developmentsand Capabilities," Progress in Aerospace Sciences, Vol. 29,1992, pp. 1-41.
9Cox, R. N. and Winter, D. F. T., "The Light Gas Gunat ARDE, Fort Halstead, Kent," AGARD Report 139,1957.
10 Bakos, R. J. and Erdos, J. [., "Options for Enhance-ment of the Performance of Shock-Expansion Tubes andTunnels," AIAA Paper 95-0799, 1995.
"Eitelberg, G., Mclntyre, T. J., Beck, W. H. and Lacey,J., 'The High Enthalpy Shock Tunnel in Gottingen," AIAAPaper 92-3942, 1992.
12Bogdanoff, D. W., Zambrana, H. A., Cavolowsky, J. A.,Newfield, M. E., Cornelison, C. J. and Miller, R. J., "Re-activation and Upgrade of the NASA Ames 16 Inch ShockTunnel," AIAA Paper 92-0327, 1992.
l3Yu, H. R, Esser, B., Lenartz, M. and Gronig, H.,"Gaseous Detonation Driver for a Shock Tunnel." ShockWaves, Vol. 2, 1992, pp. 245-254.
14Bird, G. A., "A Note on Combustion Driven ShockTubes," in "Hypersonic Facilities in the AerodynamicsDepartment, Royal Aircraft Establishment," edited byP. A. Hufton, AGARD Report 146, 1957.
15Coates, P. B. and Gaydon, A. G., "A Simple Shock Tubewith Detonating Driver Gas," Proceedings of the Royal So-ciety, Vol. A283, 1965, pp. 16-32.
16Balcarzak, M. Journal of and Johnson, M. R, "TheGaseous-Detonation Driver and Its Application to ShockTube Simulation Techniques," in Proceedings of the 5th In-ternational Shock Tube Symposium, edited by Z.I. Slawsky.J.F. Moulton and W.S. Filler, 1966, U.S. Naval OrdnanceLaboratory, White Oak, Maryland, pp. 1111-1128.
17Lee, B. H. K., "Detonation-Driven Shocks in a ShockTube," AIAA Journal, Vol. 5, No. 4, 1967, pp. 791-792.
18Yu, H.-R, "Recent Developments in Shock Tube Ap-plications," in Proceedings of the National Symposium onShock Wave Phenomena, edited by K. Takayama, Septem-ber 28-30, 1989, Sendai, Japan, pp. 1-9.
19Yu, H.-R, "Experimental Study of Oxyhydrogen Det-onation Driver for a Shock Tunnel," International ShockWave Symposium, March 16-18, 1995, Yokohama, Japan.
20Yu, H.-R., Zhao, W. and Yuan, S., "Performance ofShock Tunnel with H2-O2 Detonation Driver," Aerody-namic Experiment and Measurement & Control, Vol. 7,1994, pp. 38-42 (in Chinese).
21 Yu, H.-R and Zhao, W., "The Use of Oxyhydrogen Det-onation Driver for Generation of High Enthalpy Flow," 20thInternational Symposium on Rarefied Gas Dynamics, Au-gust 19-23, 1996.
22Yu, H.-R, Lin, J. M., Zhao, W., Li, Z. F. and Gu,J. H., "A New Shock Tunnel with Oxyhydrogen Detona-tion Driver," Paper 6070, 21st International Symposium onShock Waves, July 20-25, 1997, Great Keppel Island, Aus-tralia.
23Lenartz, M., Wang, B. and Gronig, H., "Development ofa Detonation Driver for a Shock Tunnel," 20th InternationalSymposium on Shock Waves, Pasadena, California. USA.1995.
24Engers, R J., Calleja, J. F. and Bakos, R. J., "A Det-onation Driven Injectant Heater for Pulse Facility TestingApplications," AIAA Paper 95-3153, 1995.
"Bakos, R. J., Castrogiovanni, A., Calleja, J. F.. Nucci.L. and Erdos, J. I., "Expansion of the Scramjet Ground TestEnvelope of the HYPULSE Facility," AIAA Paper 96-4506.1996.
26Erdos, J. I., Bakos, R. J. and Rogers, R. J., "Dual ModeShock-Expansion/Refiected-Shock Tunnel." AIAA Paper97-0560, 1997.
27Bakos, R J. and Erdos, J. L, "Optimizing Pressure Re-covery in a Detonation Driven Reflected Shock Tunnel.' Pa-per 5999, 21st International Symposium on Shock Waves.July 20-25, 1997, Great Keppel Island, Australia.
28Stuessy, W. S., Liu, H.-C., Lu, F. K. and Wilson. D. H.."Initial Operation of a High-Pressure Detonation-Di ive i iShock Tube Facility," AIAA Paper 97-0665, 1997.
29Stuessy, W. S., Lu, F. K. and Wilson, D. R, ''Shock In-duced Detonation Wave Driver for Enhanching Shock TubePerformance," AIAA Paper 98-0549, 1998.
30Helman, D., Shreeve, R P. and Eidelman. S.. ' 'Detona-tion Pulse Engine," AIAA Paper 86-1683, 1986.
31 Edwards, D. H., Thomas, G. O. and Williams. T. L"Initiation of Detonation by Unsteady Planar IncidentShock Waves," Combustion and Flame, Vol. 43. 19-i .pp. 187-198.
32Edwards, D. H., Williams, G. T. and Breeze J. C.."Pressure and Velocity Measurements on Detonation Wavesin Hydrogen-Oxygen Mixture," Journal of Fluid Mechanics.Vol. 6, Part 4, 1959, pp. 497-517.
33Lee, J. H. S., "Dynamic Parameters of Gaseous Deto-nations," Annual Review of Fluid Mechanics. Vol. 16. 19*4.pp. 311-336.
34Nettleton, M. A., Gaseous Detonations, Chapman andHall, London, 1987.
36Gier, H. L. and Jones, T. G., "An Investigation of a Dou-ble Diaphragm Shock Tube with a Detonable Buffer Gas.in Shock Tubes. Proceedings of the 7th International ShockTube Symposium, June 23-25, 1969, edited by 1. I. Glass.University of Toronto Press, 1970, pp. 272-290.
36Eckett, C. A., Quirk, J. and Shephard, J. E., "A NewModel for Direct Initiation of Gaseous Detonations," Paper2100, 21st International Symposium on Shock Waves. July20-25, 1997, Great Keppel Island, Australia.
37Gelfand, B. E., Popov, O. E., Medvedev, S. P.. Khomik.S. V., Kusharin, A. Y. and Agafonov, G. L., "Self-Ignition olHydrogen-Oxygen Mixtures at High Pressure," Paper 2400.21st International Symposium on Shock Waves, July 20-25.1997, Great Keppel Island, Australia.
38Lenartz, M., "Entwicklung eines Detonationstreibersfur den Stofiwellenkanal," Ph.D. dissertation, University ofTechnology, Aachen, Germany, 1996.
Notes:1. Dimensions and fill conditions are representative. These may have changed.2. GASL: light gas driver = 2.4 m long; pressure up to 140 MPa.3. TU Aachen: damping tube = 140 mm dia., 6.1 m long; initiation tube = 30 mm dia., 900 mm long.
