National Aeronautics and Space Administration

Glenn Research Center

History Banner
NERVA Engine Testing (1964 - 1967)
Photograph of NERVA engine being lifted by crane into the B-1 test stand
NERVA Engine at B-1
The Nuclear Engine for Rocket Vehicle Applications (NERVA) was a joint NASA and Atomic Energy Commission endeavor to develop a nuclear-powered rocket for both long-range missions to Mars and as a possible upper-stage for the Apollo Program. Los Alamos possessed the primary test facilities in Nevada and New Mexico, but NASA Lewis had been involved from the start with both the design of the engine’s reactor and the liquid-hydrogen fuel system. The turbopump, which pumped the fuels from the storage tanks to the engine, was the primary tool for restarting the engine in space. The complex system had to operate flawlessly in a hostile environment. Variations in flow pressure would affect the engine’s performance and ability to return from Mars. B-1 and B-3 would test the turbopump system extensively in simulated space conditions.
NERVA Program
Photograph of a researcher with large model of a nuclear rocket
Model Nuclear Rocket
NERVA began as Project Rover in 1955 as the Atomic Energy Commission instituted a study to develop nuclear-powered missiles for the U.S. Air Force. In 1959 NASA replaced the Air Force in this role, and the mission changed from a nuclear missile to a nuclear rocket for long-duration space flight. The Rover portion of the program consisted of basic reactor and fuel system research. This was followed by a series of Kiwi reactors built to test basic nuclear rocket principles in a non-flying nuclear engine. The next phase, NERVA, would create an entire flyable engine. The final phase of the program, called Reactor-In-Flight-Test, would be an actual launch test.
A series of 300-megawatt Kiwi-A reactors were tested at the Nevada Test Site in 1959 and 1960. The Kiwi-B reactors, which dramatically increased the power without increasing the overall size, were tested between 1961 and 1964. Aerojet was simultaneously incorporating one of the Kiwi-B reactor designs into its NERVA NRX (NERVA Reactor Experiment) engine. The first NERVA NRX test was run in September 1964 in Nevada, just months after B-1 became operational.
While work progressed on the reactor itself, other non-nuclear components and systems for the engine had to be checked and studied. The engine’s propellant feed system consisted of a Rocketdyne Mark IX pump driven by a Mark III turbine. Hot gas from the nozzle was cooled by cold hydrogen then passed through the turbine to spin the pump. By February 1962 NASA Lewis had made extensive plans to test the NERVA engine in the B-1 test stand.

Photograph of NERVA engine being prepared for a test at Nevada Test Site
NERVA Nevada Test
Cutaway drawing of NERVA engine
NERVA Engine
Photograph of Kiwi A reactor on display in Nevada
Kiwi A Reactor
Bootstrap Tests
Drawing of NERVA test equipment setup in B-1
Test Setup
The NERVA rocket had to be able to restart in space on its own using a safe preprogrammed startup system. Lewis researchers endeavored to design the system. The B-1 cold flow (non-fission) nuclear rocket simulation test program was created to obtain the data at the initiation and in the immediate aftermath of propellant flow. The tests were designed to start an unfueled Kiwi B-1-B reactor and its Aerojet Mark IX turbopump without prechilling the pumps or propellant lines. It was crucial for the planned long-term missions that the NERVA rocket be able to vary its speed and restart its engine without any external power. The latter was called bootstrapping.
Bootstrapping was accomplished by allowing a small amount of liquid hydrogen to flow through a valve into the reactor. The hydrogen vaporized then started the turbine that drove the turbopump. The turbopump then pumped additional liquid hydrogen to the reactor. B-1 test runs in September and October 1964 determined that the turbine could achieve bootstrap acceleration during flow initialization. This was also successfully demonstrated at Los Alamos in October. Further B-1 studies in early 1965 showed that the Mark IX turbopump accelerated as needed and did not stick. The expected pressure fluctuations in the reflector and nozzle were not as severe as in other tests. The separation of flow from the nozzle surface resulted in a large amplitude vibration in the nozzle.

Photograph of a trailer truck carrying reactor for test
Reactor Delivery
Photograph taken from upper level of test stand as reactor was hoisted up
Installing Reactor
Photograph of NERVA reactor installed in B-1
B-1 Test Section
Startup Sequence
Photograph of NERVA engine installed in B-3
NERVA in B-3
In April 1965 Lewis management decided to transfer the second series of NERVA tests from B-1 to the newly completed B-3. The goal of the program was to identify the best way to start a nuclear reactor in space. Delays in setting up the equipment at B-3 and pre-test runs pushed the tests back to February 1966. A series of turbopump tests that spring were followed by chilldown and bootstrap tests during the summer and fall.
The B-3 tests established the proper startup procedure, which included liquid hydrogen flow rates, power-cycle time delay, and the powering of the turbine. The use of a realistic feed system helped define the centrifugal turbopumps’ overall system performance and mechanical characteristics. During the tests a reheater system was installed at B-3 to quickly return the test stand to ambient temperatures following the cryogenic test runs. It was determined that the $3000 reheater shortened the estimated length of the program by 3 months and saved $50,000 worth of propellants.

Photograph of trailer tank hooked up to B-3
Inserting Gas Trailer
Cutaway drawing of turbopump
Turbopump Drawing
Photograph taken from above of NERVA turbopump
Turbopump Setup