The event occurred in the steam condensing line of the residual heat removal system the nuclear reactor. Radiolysis generated hydrogen and oxygen which then accumulated in high concentration in the pipe. During the periodical manual start-up test of the High Pressure Coolant Injection (HPCI) system, a pressure spike due to pressure fluctuations in the pipe led to ignition followed by detonation. The pipe burst open and pieces were projected damaging the surrounding structure. Noble metals (Pt and Rh) used to prevent stress corrosion of the lower part of the reactor containment vessel, might have been involved in the detonation, but this is not certain (see discussion in NIT No. 90, among References).A 3-dimensional simulation of the event was simulated by finite element analysis (see Naitoh et al. in references). It showed that the detonation pressures in the straight pipes were about 120MPa and the pressure wave propagation speed was approximately 3,000 m/s. The peak pressures at the elbows were 2.02.5 times higher than those in the straight part of the line, due to reflection and overlapping of the pressure waves. These values at the elbow above the surface of the accumulated water exceeded the critical strain, explaining why the pipe ruptured there. The results were reasonably matching the actual pipe deformation observed after the accident.
Components Involved
steam condenser pipe, pressure vessel, reactor coolant,
Pre-event Summary
A manual start-up test was performed just before the accidents.The plant was built in 1977. Following safety inspections in 1993-94, a series of improvements had been executed, including the steam condensing line, which was originally vertical and straight. It was replaced with L-shaped pipes connected to horizontally-placed pipes,to create a detour for steam coming from the reactor.
Lessons Learned
The references do not report the root causes: in particularly, the excessive presence of hydrogen remains unexplained. Radiolysis is only one of the possible explanations. Also the source of ignition was not definitively ascertained. Some calculation showed that a much higher would have been necessary to ignite the hydrogen at the operative temperature of the steam. As to the pipe failure: the fact that the geometry of the steam condensing line as modified in the year 1993-94, introducing horizontal sections has certainly caused the presence of strain-stress concentration and the bending. The only measure reported by the sources available, was the instruction, by the Nuclear Authority, to drain water from similar steam condensation system in all the other boiling water reactors of he same type. The NIT reports (see references) are very critical on the measures adopted.