Explosion at a testing facility for a space rocket engine

An explosion occurred at testing facility of a space rocket engine.
The incident occurred when executing a test studying the performance of a hydrogen gas flow measuring system for fuel feeding system of the rocket.
During testing of a 42 MPa high-pressure hydrogen gas flow system, an exhaust gas duct outlet was damaged; hydrogen blew out and exploded. Apparently the cause of this accident was the deterioration of the exhaust gas duct outlet by thermal stress.

A test started, The hydrogen gas begun flowing.
5.8 seconds late, the internal pressure of the exhaust gas duct reached about 22 MPa and the duct damaged, causing ejection of gaseous hydrogen.
7.3 seconds later, a hydrogen gas detector detected the leakage of hydrogen gas, and a breaker automatically shut. Then, after one second, the blowing off of the hydrogen stopped.
12.9 seconds later, the leaked hydrogen exploded.

A slate roof of the duct testing house blew off as far as about 160 m away. Just after that, the second explosion occurred at 200-300 m up in the sky.

The INITIATING cause was the embrittlement of a welding in the exhaust gas duct. More than 132 tests had been performed with high-pressure and high-temperature hydrogen gas combustion.
The fact that cracks were found and repaired before, with reduction of the thickness of the duct, hints at some shortcoming in the procedures and/or in the assessment of the possible consequences of the repair, such as the reduction of the mechanical resistance of the component. There is possibly also a shortcoming in the management of operation, failing to learn from previous near misses

fuel-feeding system for a space rocket engine

The exhaust gas outlet duct was 21 cm in diameter, 91 cm in length and made of stainless steel and nickel alloy. This duct had been used for five years and 132 tests had been carried out before the day of the accident.

After 117 tests, some cracks had been found at the inside surface of welded parts
Also more recently, after a check in August, 1990, some cracks were also found. These cracks were removed by grinding and the equipment's strength was confirmed by a pressure-proof test, an air test, and a special test. After the repair in August, 1990, the welding thickness had a maximum of 20 mm and the minimum of 6 mm.

According to the second source, this testing site for hydrogen rocket had been plagued with problems. The rocket’s scheduled launching had been recently postponed by a year, to January or February 1993.
Two similar accidents had already occurred in the past: first in September 1987,when the roof of the facility was blown off by a liquid hydrogen leak. The second time in November 1987, with only minor damage from a blast caused by a similar leak.

A pump testing room, a general testing room, and the gas generator turbine testing ceiling were all destroyed.
Almost all measurement piping in the pump testing room received blast damage. A slate roof 50 m separated from high-pressure gas containers, and walls around the exhaust gas treatment equipment 190 m away were damaged.
Neighbouring damage: eight private houses (the farthest one was about 2 km away), and one factory were damaged. At seven of the houses, windows were broken. At two of the houses, parts of walls and fixtures were moved.
The explosion was heard 4-5 km away.

The interlocked hydrogen detection system intevened, stopping the hydrogen flow. This prevented further escalation.
Despite the the quick automatic action (7.3 sec after vent rupture), it activated too late to avoid the formation of a flmmable air-hydrogen mixture , which ignited with delay and exploded.

If an equipment is repaired several times, there is a possibility that its original performance has reduced. In particularly, grinding as a method to eliminate superficial cracks in a pipe reduced its thickness and possibly also its capacity to bear stresses. The tests performed after the grinding to measure the residual strength were clearly not enough to measure the duct’s behaviour under the real testing conditions.

(1) A flange joint was installed, to allow for the inspection of the welded part from inside. The basic material and welding material were changed.
(2) An inner liner to prevent thermal stress was added.
(3) A limit to the maximal number of the use was set.
(4) The repair standard was reviewed