AT a remote air base in California鈥檚 Mojave Desert, engineers are refitting a retired SR-71 Blackbird spy plane for a series of ambitious experiments. The aircraft is being modified to carry a scaled-down version of a rocket engine that will be lighter and simpler, yet more powerful than anything that has gone before.
Later this month, with the new engine strapped to its back, the Blackbird will begin tests at altitudes of more than 25 kilometres and at speeds above Mach 3.
The engine is being designed and built by Rocketdyne, the Californian company that built the engines for the space shuttle. But the work is part of a wider programme funded partly by NASA and partly by the American aerospace company Lockheed Martin, to design a reusable rocket that could eventually replace the shuttle. If the tests are successful, larger versions of the engine could one day power launch vehicles that would make space flight as easy as air travel is today.
Advertisement
For decades, space enthusiasts have dreamed of a reliable, reusable engine that would be simple and cheap to operate, but powerful enough to blast an aircraft-like vehicle into orbit. This is what NASA had in mind when it began designing the space shuttle in the 1970s. But, at that time, the technology needed to build engines that were efficient and powerful enough did not exist. Instead, to reach space the shuttle requires two strap-on, solid fuel rocket boosters and a giant external fuel tank to hold the 2 million litres of liquid oxygen and hydrogen that flow through its three main engines during a launch.
All conventional rocket boosters work in the same way. They generate thrust by jettisoning hot gases from a bell-shaped nozzle. As the gases expand, they push against the inner wall of the nozzle forcing the rocket forward. The size and shape of the nozzle is crucial for maximising this force. If the nozzle is too small, much of the expanding gas will be pushed out before imparting any thrust to the rocket. And if the nozzle is too big, the force of the gas will be spent before it can reach the inner wall.
The trouble with this mechanism is that the rate of expansion of a hot gas depends on atmospheric pressure and this changes dramatically as a rocket gains altitude. To compensate for this change in pressure, a nozzle would have to expand as the rocket climbed. But such a system would be mechanically complex and heavy, not to mention expensive to operate and maintain.
This is not a problem for today鈥檚 expendable launchers because each stage can be fitted with a nozzle optimised to work at a particular altitude. But the next generation of launch vehicles are being designed to enter and leave the Earth鈥檚 atmosphere using only one stage, so this will no longer be possible.
The aerospike neatly sidesteps this problem. It was conceived in the 1970s as the main thruster for the shuttle, but abandoned because of the formidable technical challenges needed to make it work. The design turns the concept of thrust generation on its head. Instead of expelling the exhaust into a bell-shaped nozzle, the aerospike jettisons hot gases along the outer surfaces of its nozzle where they expand, pushing the rocket forward. The exhaust plume is open to the atmosphere and, although changes in pressure influence its size, this has a negligible effect because the thrust is created within the plume.
A full-scale aerospike would be more powerful than ever before. While the shuttle鈥檚 main engine produces a thrust of around 1.75 meganewtons, the aerospike will deliver almost 1.92 meganewtons. And it will have a greater specific impulse, the term that engineers use to describe the engine鈥檚 fuel efficiency. For every kilogram of fuel that it burns in a second, the shuttle creates 4480 newtons of thrust compared with 4500 newtons for the aerospike. This is a small difference, but it adds up to a significant advantage during the long journey into space.
Of course, an entirely different kind of nozzle is required and this is where the engineering challenges lie. Temperatures inside the plume reach 3000 掳C and, if not cooled, most materials would melt or burn. Cooling the nozzle is tricky. 鈥淥ne of the reasons we didn鈥檛 do this 10 or 15 years ago was that we didn鈥檛 have structures that could cope,鈥� says T. K. Mattingly, a former astronaut who is now in charge of Lockheed Martin鈥檚 reusable rocket programme.
The design that Rocketdyne has adopted consists of two ramp-like surfaces that form a truncated wedge pointing away from the rocket-the so-called spike (see Diagram, next page). The exhaust is expelled along each surface and the entire nozzle is fixed -a radical departure from conventional boosters. The nozzles on the shuttle engines, for example, are designed to pivot on hydraulic struts mounted on bearings known as gimbals. Pivoting allows the rocket to change the direction of thrust to make course corrections. But this system is mechanically complex and expensive to maintain.FIG-20374201.gif
By comparison, an aerospike engine can change direction by pumping more exhaust down one side of its spike than the other. And this can be controlled by a simple valve. 鈥淵ou don鈥檛 have actuators and gimbal bearings,鈥� says Mike Hampson, the Rocketdyne engineer in charge of the project, and this makes the engine simple and easy to maintain.
The biggest challenge is cooling the aerospike nozzle. In the test engine, which is only a tenth of the size of the real thing, the nozzle is made of a solid steel block drilled with holes through which water will be pumped to carry away any excess heat. Earlier this year, however, engineers discovered that it was not possible to cool the spike properly when it was operating at full thrust and the highest temperatures. 鈥淎pparently, the passages were too narrow and didn鈥檛 allow enough cooling water to flow,鈥� says the NASA manager who oversees the project. Tests had been due to begin in April, but fixing the problem has set the programme back several months.
The flights are designed to measure the manoeuvrability and flight stability of a reusable rocket that is powered by an aerospike engine. One fear is that the exhaust plume from the engine will interfere with the air flowing over such a vehicle as it passes the sound barrier. This could make it unstable. To find out, the engine will be housed in a scale version of the rocket鈥檚 airframe and the airflow carefully monitored.
To perfect the design, Lockheed has carried out extensive wind tunnel tests at its Skunk Works, the top-secret factory in California where the Blackbird, the U-2 spy plane and the F-117 stealth fighter were developed. The results have been promising and a reusable rocket should be stable-most of the time. 鈥淏ecause the aerospike nozzle is open you get some upstream effects along the vehicle,鈥� says Hampson and he expects to have to tweak the design. 鈥淲hat we want is real flight data in the real world,鈥� says Mattingly.
To get these data, NASA is paying $11 million for a series of 26 flights that will be carried out over 13 weeks this summer. The tests will involve turning the engine on in three-second bursts and then turning it off, says Hampson. Back on the ground, he will be able to watch the engine鈥檚 performance with the aid of remotely operated cameras attached to the back of the aircraft, while sensors on supporting struts will measure the amount of thrust the engine produces.
If the early tests prove that the aerospike concept is sound, the next task will be to incorporate a full-scale version of the engine into Lockheed鈥檚 reusable launcher, Venture Star. When 快猫短视频 went to press Lockheed was on tenterhooks, waiting to hear if Venture Star will be chosen as the successor to the space shuttle. With the aid of seven aerospike engines, Venture Star will take off vertically like a conventional rocket and glide back to Earth like the shuttle. However, Venture Star has no wings. Instead, it is a lifting body-its shape is designed to create lift without the extra weight of wings. For stable flight, the centre of gravity must be carefully positioned relative to its centre of aerodynamic lift. But with seven engines at the rear, the vehicle is in danger of becoming bottom-heavy. 鈥淲e don鈥檛 want to have to load the nose with lead because the engines weigh too much,鈥� says Hampson.
For this reason, keeping the weight of the engines down is another daunting design hurdle. Rocketdyne intends to make the full-scale aerospikes using composite materials that are as strong as steel but a fraction of the weight. Hampson says there are three or four candidate materials, but the most promising is carbon silicon carbide. 鈥淚t鈥檚 never been done before with composites but it significantly reduces the weight of the engine,鈥� he explains.
And instead of cooling the spikes with water, which requires the extra weight of a pump and reservoir, the full-scale engines will be cooled using liquid hydrogen from the fuel tanks. The nozzles on the shuttle鈥檚 main engines are cooled in a similar way. With these savings, each full-scale aerospike engine will weigh only 2300 kilograms, roughly two thirds the weight of one of the shuttle鈥檚 main engines.
Incorporating the engine into the structure is another challenge. The shuttle engines are separated from the craft by bulky heat shields, but Lockheed engineers believe this may not be necessary with Venture Star since the engines are cooled. 鈥淲hat we鈥檙e doing is replacing the heat shield with the engine itself,鈥� says Mike Rankin, the Skunk Works engineer in charge of integrating the engine with the vehicle.
In fact, the Lockheed blueprint requires the engines to be part of the airframe. 鈥淥n most rockets, the engines are just hanging out the back end and you have real estate that鈥檚 doing nothing but creating drag,鈥� says Hampson. The problem is that connecting the engine directly to the airframe influences the aerodynamic properties of the vehicle. 鈥淲e鈥檒l pay a flight penalty to do that. What we鈥檙e trying to do with our test programme is prove exactly what that penalty is,鈥� explains Rankin.
Venture Star will stand 38-metres high, smaller than the shuttle and half its weight-only 900 tonnes. A sub-orbital demonstrator could even be built by 1999 and the aerospike engine could be propelling a full-scale vehicle into space early next century.
鈥淚 saw a version of the aerospike at Rocketdyne back in 鈥�72 and they said it was the wave of the future,鈥� says Mattingly. 鈥淲ell, I鈥檝e been expecting the wave of the future for a long time.鈥� The next series of tests will show whether it was worth the wait.