Purported SR-72 Propulsion

For some years now there have been marketing-hype disclosures about Lockheed Martin’s efforts toward the “SR-72”,  an intended follow-on to their famous SR-71 “Blackbird”.  The hype was about hypersonic speeds above Mach 5,  and some hand-waving about an advanced engine,  usually supposedly a combined-cycle gas turbine and scramjet (supersonic-combustion ramjet) engine.

I knew the hand-waving about combined-cycle turbine-scramjet was BS,  because about the fastest practical speed for gas turbine is about Mach 3.2 to 3.3 due to overheat damage,  and about the min takeover speed for scramjet is Mach 4.  Plus,  the inlet and nozzle geometries are utterly incompatible. 

What that really means is that your propulsion unit has to operate first as a gas turbine to take off and climb and accelerate to ramjet takeover speed at about Mach 2.5,  then operate as a (subsonic-combustion) ramjet to accelerate above Mach 4,  then finally operate as a scramjet to “fly hypersonically” at or above Mach 5.  The ramjet and the gas turbine share similar inlet and nozzle geometries,  but the scramjet is still utterly incompatible geometrically with the other two.  And,  you must change engine type in order to slow down for a more economical cruise. 

My suggested solution has,  up to now,  been “parallel-burn” propulsion:  do not try to combine the various propulsion types into one design,  instead install all 3 separately,  each optimized for what it is.  (Combined,  it is inevitable that performance of each component suffers greatly.)  But,  a major problem with parallel burn at higher speeds (where drag is high),  is that no one of these propulsive items is a large enough fraction of the vehicle frontal cross section area!  That severely limits the max speed attainable,  likely to less than hypersonic,  which eliminates any reason to have the scramjet at all!

Concept for Combining Gas Turbine with Ramjet and Scramjet

I have since had a sort-of hybrid idea.  The 3 systems can share one common supersonic inlet capture installation,  but nothing else!  The post-capture channels of the inlet must be made variable geometry,  so that the gas turbine and the ramjet can be fed subsonic air in a diverging channel,  while the scramjet is fed supersonic air in a constant-area channel.  The supersonic channel to the scramjet must be “straight through”,  you absolutely cannot divert a channel carrying supersonic flow,  because the turn always causes shock-down to subsonic flow!  Anybody who claims otherwise is spouting pure BS!

The gas turbine needs to be a low-bypass ratio afterburning design suitable for supersonic flight,  and also be fitted with air bypass tubes around its core big enough so that they can carry 100% of the air flow,  tapped off ahead of the compressor face,  and going directly to the afterburner.  (In the SR-71,  those engines had 25% max air bypass,  tapped from the 3^rd or 4^th stage of the compressor.)   In that way with 100% bypass,  the afterburner can also serve as the subsonic-combustion ramjet combustor,  using the very same post-capture subsonic inlet air channel as the turbine uses.  But,  we do need to stop the airflow into the compressor,  to avoid overheat damage!  And we need to stop backflow from the afterburner into the turbine!  Ramjet combustor gas temperatures are far higher than any allowable turbine inlet temperatures,  and “leaks” lower the ramjet pressure,  lowering performance drastically.

Therefore,  it is a key requirement here,  when operating as a ramjet,  to stop the backflow from the afterburner chamber from going up through the turbine into the turbine engine.  That is a serious and extremely difficult design problem to solve!  But it must be solved,  to prevent turbine overheat,  and to raise the achievable chamber pressure of the ramjet,  in order to preserve its performance.  Leaks are low chamber pressure,  and low pressure is low performance.  Period.  That was settled long ago in tests.

What you “buy” with the 100% bypass and the backflow stoppage complications,  is a gas turbine and a ramjet that share the same portion of the vehicle frontal cross section,  which then can be a much larger fraction of vehicle frontal cross section,  so that the top speed in ramjet can be higher,  reaching the scramjet takeover range at Mach 4+. 

For scramjet takeover,  you must suddenly change the inlet post-capture channel geometry to a long,  straight supersonic feed to the scramjet,  that is also the “isolator duct” required for stable scramjet operation.  This scramjet must be parallel-mounted to the rest of the propulsion,  and must be completely separate,  except for sharing the supersonic capture features.  It lets you put the scramjet on the belly of the aircraft,  and to use the vehicle aft underside as a free-expansion nozzle surface. That reduces (but does not zero) the scramjet’s fraction of the vehicle frontal cross section,  as opposed to that of the turbine/ramjet,  to about a 50-50 split.  That highly-integrated geometry in turn increases the max scramjet speed against drag,  making more-than-minimum (Mach 5) “hypersonic speed” feasible.

Doing these required design features is a hellaciously-difficult problem,  but does offer a potentially-feasible solution for hypersonic flight that does not involve rocket thrust to takeover speed.  I have not even touched on the thermal management issues,  which may,  in point of fact,  be fatal to the concept!  Suffice it to say the usual construction techniques for the afterburner and its nozzle cannot be used,  because for Mach 3.3+ speeds,  there is no such thing as the cooling air that those technologies require.

Finally,  if the marketing hype you see does not include a propulsion system that addresses the issues I have raised here,  and a thermal management scheme that addresses the propulsion and the inlet and the airframe,  then I suggest that you dismiss it as the BS that it quite evidently is!

A cartoon sketch of my scheme is given here as Figure 1.

Figure 1 – A Possible Means to Combine Gas Turbine Takeoff and Landing with Scramjet Dash

Rocket-Boosted Ramjet Is a Much Better Way

Actually,  I still prefer my parallel-burn,  completely separate,  rocket and ramjet solution,  and just forget the scramjet!  To take off,  climb,  and accelerate to around Mach 2.5 does not require all that big a rocket engine,  or all that much propellant.  The subsonic-combustion ramjet takes over at about Mach 2.5,  and supports supersonic cruise much more economically in the vicinity of Mach 3,  but with enough frontal cross section fraction to support supersonic dash speeds to Mach 5,  or possibly even Mach 6.  And that is hypersonic!  No scramjet required!  It just has lower specific impulse at hypersonic speeds,  as does the scramjet.  However,  you do not have to change propulsion to slow down to cruise!

If you include some small liquid rocket propulsion,  your landing is not entirely “dead-stick”.  Just fire up the liquid rockets to divert or go-around.  I find that to be a far safer and more practical solution,  manned or unmanned!

The main mass of booster propellant to reach ramjet takeover,  is likely a solid packaged within the ramjet combustor as an “integral rocket ramjet” booster (IRR booster).  There are two reasons for this:  (1) the booster needs to be big to have the very high thrust to accelerate very quickly to ramjet speed,  to reduce the aerodynamic drag losses to tolerable values,  and (2) there are no air-cooled technologies available for the combustor and nozzle internal heat protection at these flight speeds,  since there is no such thing as “cooling air” above about Mach 3.2 to 3.3;  thus the only technological solutions for combustor and nozzle are one-shot ablatives.  The IRR is proven,  existing 1-shot missile technology. 

That last says you need to pull the entire ramjet combustor unit out,  and replace it,  after every flight!  It therefore might as well contain an integral solid booster,  just like what has proved so successful in missile work.  You need the big boost to ramjet speed only once per mission!  The smaller liquid rockets let you fly the plane at speeds below ramjet speed,  for the approach and landing.

See Figure 2. 

Figure 2 – Rocket-Boosted Ramjet as a Means to Achieve Hypersonic Dash

Figure 3 shows some details about how the cartridge-loaded ramjet combustor and nozzle is also its own  integral rocket ramjet (IRR) booster.  The craft need accelerate only once to ramjet takeover speed,  and the IRR booster does that job,  then transitions to ramjet thrust in about 0.1 sec (as demonstrated by ASALM-PTV in flight).  The liquid rockets are much smaller,  and mainly serve to keep the descent and landing from being totally “dead stick” (with no go-around or divert capability).

Combustor and nozzle heat protection is by ablative materials,  which cannot be re-used.  So,  the IRR unit must be replaced for every flight.  In this concept,  there must be airframe structure to support the vertical tail,  so the IRR unit resides inside this airframe,  not exposed to hypersonic external aeroheating.  That greatly simplifies the thermal management,  to something the ablatives can easily handle for very long burns.  The case can be power-washed out,  refitted with ablatives,  and cast with another propellant charge.  On-pavement recovery has little in the way of risk to support this kind of reuse.

By making the bottom flat with the bifurcated inlet ducts,  there is little need for wing area in supersonic flight above about Mach 3,  but there is room for the small liquid rockets aft of the inlets ducts!  The wing is really sized for a tolerable landing speed,  with the delta planform allowing high angle of attack without stalling.  It is mostly just parasite drag at high speeds,  so there are many design tradeoffs here.  However,  at very high altitudes in very thin air,  the wing allows sufficient lift generation at lower angles of attack that correspond to lower drag-induced-by-lift.  This may help extend cruise range,  and certainly might help extend the service ceiling.   The “right” wing is quite likely smaller than the one sketched on the figure.  

Figure 3 – Cartridge-Loaded Ramjet Combustor with IRR Booster

In cruise at about Mach 3,  the ramjet specific impulse (Isp) should be in the neighborhood of 1000-1300 secs.  Running richer at full ramjet thrust for Mach 5+ dash,  the ramjet Isp is likely nearer only 700-800 sec.  The liquid rockets are lower-pressure units that are simply pressure-fed the LOX,  and little bit of the same thermally-stable kerosene that the ramjet uses.  It would be realistic to expect about 300 sec of Isp out of them.  The solid booster,  at about 85-87% solids,  would achieve a sea level Isp near 250-255 sec.

This plane could actually take off using the small rockets,  like the “rocket racer” did,  although zero-length launch from a ramp is also very feasible,  since the integral rocket booster accelerates the airplane at 5+ gees.  Once leaving the pattern,  you pull up sharply,  fire up the solid booster and shut down the small rockets.  Seconds later,  you do ramjet takeover at about Mach 2.5 while climbing very steeply,  and at much higher altitude.  The ramjet then takes you to cruise conditions,  and also hypersonic dash. 

At mission’s end,  you start your approach in ramjet,  but shut it down as you decelerate below Mach 2.5,  making most of the rest of the approach in glide.  As you near the field,  use the small liquid rockets as necessary to divert or to go around for a missed approach.   There is only one boost to ramjet takeover per mission,  but the small rockets can be used multiple times for multiple purposes in a mission. 

You swap out the spent combustor unit for a fresh one,  and refill the kerosene and oxygen tanks.  With on-ramp recovery,  spent combustor refurbishment is also a very low risk possibility.  Easy!

None of these considerable existing-technology advantages obtain with the sort-of combined-cycle gas turbine/ramjet/scramjet craft described above.  There are still missing-technology items with it,  but not with this rocket-ramjet airplane.

Related Information:

If you want to see more about how supersonic inlets really work,  and how they are adapted to ramjet versus gas turbine,  please see on this site “Fundamentals of Inlets”,  posted 9 November 2020. 

If you want to see more about how (subsonic combustion) ramjets really work,  please see “How Ramjets Work”,  posted 1 December 2022,  and “Primer On Ramjets”,  posted 10 December 2016. 

The general issues that must be addressed for hypersonic vehicles are discussed in “About Hypersonic Vehicles”,  posted 1 June 2022.  A peculiar problem with high hypersonic flight is discussed in “Plasma Sheath Effects in High Hypersonic Flight”,  posted 18 September 2022,  which debunks some of the widely-circulating myths about “unstoppable” hypersonic missile weapons.

If you want to see what an integral solid booster is,  please see “Solid Rocket Analysis”,  posted 16 February 2020,  and concentrate on the low L/D keyhole slot grain design therein.  How the internal ballistics of solid propellant devices work is well-explained.  There is also information on achievable burn rates,  and on safety sensitivity data.

The thermal management issues are discussed in more detail in “On High-Speed Aerodynamics and Heat Transfer”,  posted 2 January 2020,  “Heat Protection is the Key to Hypersonic Flight”,  posted 4 July 2017,  and “Shock Impingement Heating Is Very Dangerous”,  posted 12 June 2017.   

Flameholding in the ramjet wasn’t an issue discussed here,  but if you are interested,  that is discussed in “Ramjet Flameholding”,  posted 3 March 2020.  Something similar applies to scramjet,  and something somewhat different (but still similar) applies to gas turbine can combustors.  That article makes clear why the usual V-gutter and can stabilizers cannot work at speeds past about Mach 3.3,  and what will work.

There is a whole catalog article,  sorted by topic area,  of many of my technical articles posted on this site.  It is “Lists of Some Articles By Topic Area”,  posted 21 October 2021.  There is some duplication from list to list,  where the topic areas overlap.  It does have topic areas for ramjet,  for rocket stuff,  and for high-speed aero-thermo-dynamics and heat transfer.  I do try to keep that article updated and current. 

You can use the navigation tool on the left side of this page to access any of these articles very quickly.  Just jot down the titles and dates.  Then click on the year,  the month,  and finally the title if more than one was posted that month. 

One Final Note:

All of this was done with open sources!  I have seen no classified information for nearly 3 decades now,  since I last held a clearance and had a need-to-know.  But it is quite likely that any “real” SR-72 vehicle will be considered a classified design by the government,  much as the SR-71 was.  About 4 decades ago,  I roughed-out a vehicle somewhat similar to the rocket-ramjet hypersonic craft outlined here,  from only open sources.  (If you really know what you are doing,  open sources are all you need.)  That design concept was confiscated by the FBI and classified by the Pentagon.  They were exploring SR-71 replacements,  even way back then.  If this current one disappears off my site,  then it happened again.

--------------

Update 9-5-2023:  I took some time to rough-out the characteristics of a rocket ramjet airplane design,  and along the way found a major choice to be made.  Since this was not already done in the original article,  see first the intended flight profiles,  given in Figure 4.  The plane could take off from a runway using its small-rocket power,  leading to the big solid booster ignition away from the airport,  or it could be launched zero-length from an inclined ramp,  directly with the big booster.  Climb and acceleration to cruise speed (and to dash speed) is by the ramjet.  Most of the approach to landing is “dead stick” glide,  but with the small liquid rockets available,  to divert,  or to go around for a missed approach.

Figure 4 – Concept Flight Profiles

I literally sized a paper liquid rocket design that uses LOX and the ramjet fuel (thermally-stable kerosene),  but is a very simple pressure-fed system.  The design goal here was simplicity above all else,  so that reliability would be highest.  This kind of thing should be utterly trouble-free,  at the cost of somewhat lower performance.  I did not choose a specific igniter,  but I did indicate that the igniter is linked to the on-off valves for the propellants.  It fires when they flow,  for some small set time interval.

The pressurant for the propellant is dry nitrogen,  commonly available in 2200 psig bottles.  It is likely an airframe-mounted vessel that is filled on the apron from standard gas bottles.  The regulators are set to deliver 700 psig to the propellant tanks,  so that a bit over half of the gas vessel pressure drop is available during the mission.  Assuming the pressure drop through the passages and injector plates is about 200 psi,  a max chamber pressure of 500 psia seems reasonable.  2:1 pressure turndown is easily achieved.   

The 15 degree conical nozzle is designed for expansion to 11.2 psia,  so that the expected separation backpressure at half pressure is still very slightly above sea level atmospheric.  That way,  nozzle flow separation is never a concern!  Expected performance data is shown in Figure 5,  including the small-rocket frontal thrust density value,  based on its exit area.

Figure 5 – Roughing Out a Small Liquid Rocket System Emphasizing Simplicity Above All

I had some old ramjet data predicted for a design with inlet shock-on-lip Mach number 2.5,  using kerosene fuel at equivalence ratio ER = 1.10 for max thrust without excessive waste. These data were for Mach numbers from 2 to 6 at 40,000 feet (40 kft) on a US 1962 standard day.  I curve-fit the variations in thrust and specific impulse vs Mach number at 40 kft,  and recorded the key area ratios and size of the sized engine.  I had no data at sea level or at 85 kft,  but instead just ratioed the thrusts by the ratio of atmospheric pressures.  That is not “right”,  but it is pretty close.  It was easy to divide the installed ramjet thrust by its nozzle exit area,  to get the frontal thrust density for the design study.  I took an educated guess for the leaned-back cruise specific impulse at Mach 3 cruise,  at 85 kft.

I also had some old vehicle drag data based on information from Hoerner’s old “drag bible”.  It includes nose pressure drag,  lateral skin drag,  aerosurface drag,  and base drag effects.  It is uncorrected for the drag area reductions associated with the chin inlet mounting,  and for the propulsion plumes coming from the base.  That makes these drag values a probable over-estimate by a few-to-several percent,  but at least the trend with Mach number is correct.

The drag and ramjet thrust density and specific impulse data are given in Figure 6. 

I had an old IRR booster grain design in my records.  It is for the wrong size,  but the L/D proportion is not too far wrong.  It was easy to compute its thrust per unit exit area,  for a scaleable frontal thrust density F/Ae = 18,350 psf to use in this study.  The detail internal ballistics are not quite right,  but the frontal thrust density is in the ballpark,  regardless.  Some selected data are shown in Figure 7.  

Figure 6 – Rescaling Ramjet Performance From Some Old,  Limited Data

Figure 7 – An Older Grain Design Used To Rescale IRR Booster Performance

The original notion of the flat-bottomed airframe with the bifurcated inlet,  finally sized-out capable of reaching Mach 5,  with the ramjet exit area A6 proportion to the vehicle frontal blockage area Sx reaching A6/Sx = 0.623,  as indicated in Figure 8.  This is less ramjet frontal thrust density than originally desired,  which is what limited the max dash speed to Mach 5.  The data include a preliminary weight statement and some estimated component lengths.  Gross cruise range exceeds 3000 nmi,  at 85 kft. 

Figure 8 – Results for the Flat-Bottomed Airframe With Bifurcated Inlet

The radial distance from vehicle outer mold line to the case or fuel tank OD is a critical variable,  as well.  There must be some such distance,  to isolate thermally the hot lateral skins from the vessels containing fuel or solid propellant.  That would include some high-temperature mineral wool insulation. 

Initially I set this at 6 inches,  and could not exceed Mach 4.  Setting it to 3 inches got me not quite to Mach 5.  Resetting it to 2 inches actually got me to Mach 5.  But that is about all I can realistically squeeze out of this design concept!  The strakes containing the bifurcated inlet and small liquid rocket equipment are just too large,  driven by the required air inlet duct branch sizes. 

That trend illustrates the crucial role frontal thrust density plays in high supersonic,  low hypersonic flight.  There is no getting around this,  it is quite fundamental.

An alternative design concept would not bifurcate the inlet.  Instead it would pass through the fuel tank on its way to the engine,  within an airframe of round cross section.  That makes the tank longer.  There would be no plenum,  but there would need to be a space in which to S-duct the inlet from the bottom up to the central axis.  The wing would have to move up to a mid-wing mount,  likely just a double delta planform.  The small rocket system would have to be mounted in the base of the vertical tail fin,  much like the one used in the NF-104 design.   

I re-ran this alternate configuration,  getting the results shown in Figure 9.  The top dash speed reached Mach 5.5,  reflecting the much larger ramjet frontal thrust density associated with A6/Sx = 0.844.  It packages less fuel mass,  but it also has less cross section area producing drag,  so the drag (and thrust requirement) is lower.  The gross cruise range figure is then just about the same,  as a result. 

If the ramjet propulsion were exposed at the rear,  being all of the aft airframe cross section,  A6/Sx would be a bit higher still (very nearly 1.0),  and the top dash speed would then approach Mach 6,  the same way it did with ASALM-PTV on the one flight test in 1980.  But such exposed propulsion is a much tougher thermal problem to solve for long-duration burns.

Figure 9 – Results for the Round Section Mid-Wing Airframe with Inlet Through Fuel Tank

Bear in mind that all of these are crude estimates,  only within about 10%,  at best.  However,  that is good enough to determine that dash speed nearer Mach 6 will trade off against the far-more severe thermal management problems with exposed propulsion.  Meanwhile,  if Mach 5 dash is “good enough”,  the flat-bottomed low-wing airframe with the bifurcated inlet is quite feasible. 

Or if Mach 5.5 dash is absolutely required,  the better choice is the round airframe with center-duct inlet and a mid-mounted wing.  That one will be somewhat more challenging to detail-design,  and it will have less volume available within its nose.  (You get what you pay for.) 

Also bear in mind that the next most important feasibility item is thermal management.  Those calculations have yet to be explored.  

---------------

Update 9-18-2023:  As it says in the previous update,  the thermal management issues still need exploration,  in order to determine feasibility.  Here is an initial exploratory look. 

First,  look at “typical” lateral skins parallel to the oncoming stream.  These could be on aerosurfaces away from leading edges,  or on fuselages away from nose tips and inlet capture features.  This uses a flat plate convection model that accounts for both compressibility and the effects of viscous dissipation.  Overall setup and results are given in Figure 1.  Conditions at the edge of the boundary layer would not be very far from freestream conditions,  not enough to make a great deal of difference in the film coefficient,  so this analysis just uses free stream.  

Figure 1 – Thermal Analysis of a “Typical” Lateral Skin Panel

In the figure are plotted total temperature Tt,  recovery temperature Trec,  two curves representing equilibrium panel temperatures,  and the recommended max service levels for several possible panel materials. 

The analysis included not only convection to the panel,  but also thermal re-radiation from the panel,  as its primary method of cooling.  This was done for a typical low emissivity,  and a typical high emissivity.  Also included were two paths for minor cooling effects due to conduction into the interior.  One was through a low density mineral wool insulation layer,  occupying nearly the same area as the panel.  The other was through a minor area representing the conduction path through whatever structures attach the skin panel to the rest of the airframe,  presumed metallic,  and of a length comparable to the insulation thickness.

For reference,  a completely uncooled panel would soak out to the recovery temperature.  At speeds under roughly Mach 4,  the panel’s surface thermal emissivity does not make much difference,  since the temperatures are low enough that there is not much thermal re-radiation.  However,  above Mach 4,  the panel emissivity makes a great deal of difference,  with high emissivity (dull black surface) much better.

Note how organic composite panels are no good above (at most) Mach 2,  and that presumes adequate strength at the max temperature of about 200 F,  which presumption is seriously in question.  Aluminum is useless above about Mach 2.5,  which explains very well why most fighters made of it,  have max dash speeds of only just about Mach 2.5.    

A lot of folks think titanium is a high temperature material,  but that is mistaken.  Its max service temperature is 600 to 800 F (800 F shown),  which is good to a most about Mach 3.5-ish,  presuming a highly-emissive surface.  That explains very neatly the max flight speeds of about Mach 3.2 for the SR-71,  which had a dull black finish. 

Above 1500 F capability,  there are only some stainless steels,  and 3 exotic alloys that are not steels.  Of these,  only one has truly high temperature capability at 1800 F plus high tensile strength:  Inconel X-750 (formerly simply known as “Inconel-X”).  Which neatly explains the choice of “Inconel-X” skins on the X-15 rocket plane.  The difference between the low and high emissivity effects is the difference of about a full Mach number for survival of lateral skins at full strength:  Mach 6 if high emissivity,  only Mach 5 if low.  Which in turn neatly explains why the X-15 had a dull black finish.

Thermal analysis of nose tips and leading edge pieces is much harder to approximate with these simple by-hand techniques.  The actual stagnation zone seeing full stagnation heating is quite small.  The large lateral areas also see convection approximatable with the flat plate model,  but at edge of boundary layer conditions crudely approximated as those behind the oblique shock corresponding to a 10 degree flow deflection.  There is thermal re-radiation cooling from both the stagnation zone,  and the lateral surfaces.  There is even conduction cooling through the thickness of the part,  moving toward where it attaches to the rest of the structure. This concept is illustrated in Figure 2.  

Figure 2 – “Typical” Thermal Equilibrium Considerations for a Leading Edge Piece

The results did not validate the equilibrium model.  In all cases attempted,  the convection into the lateral surfaces (both top and bottom together) simply overwhelmed the effects of stagnation heating convection,  and also the numbers for all three of the cooling paths.  The “equilibrium” temperatures to balance the mathematical model were above the oncoming stream total temperature,  which is the maximum soak-out temperature the part could see.  We must therefore conclude that in the absence of active cooling means,  these leading edge parts will rather quickly soak out to the oncoming stream total temperature,  or very near to it. See Figure 3.  

Figure 3 – Leading Edge Piece Results

The Inconel-X material as a leading edge piece may or may not need its full strength to withstand the local wind pressures upon it.  Roughly speaking,  it reaches its max service temperature limit,  or a bit above,  at about Mach 5.  Mach 6 is very near the melting point for the material. This very neatly explains why the X-15A-2 vehicle was coated with a pink silicone rubber ablative and white ceramic paint topcoat,  for high-speed flights past Mach 5.  On flight 188,  with Pete Knight flying it,  it reached Mach 6.7.  There was extensive airframe damage from simple overheat in multiple stagnation regions,  and near-fatal shock-impingement heating underneath the tail section. 

What that really tells us is that for long flights beyond Mach 5,  one must either do high-capability active cooling,  or else use ablative materials for the leading edges and nose tips.  Active cooling will be very heavy,  and very expensive in terms of the power to run it.  Ablatives will require replacement,  at worst after every flight,  or at best after every few flights.  The ablative approach is exactly what was done with the Space Shuttle and its derivative the X-37B,  and also the old X-20 design never built.

Remember:  if you have airbreathing propulsion,  the inlet capture features are even more challenging than leading edges and nose tips,  and the buried ducts simply will require active cooling. 

If you have no thermal management solution,  you do not have a viable design for hypersonic flight!