During this lapse of time, the pressure in the primary cooling system, which was insufficiently cooled, increased to the point where it triggered opening of the pressuriser relief valve, whose purpose is to evacuate the excess steam towards a tank and thereby reduce the pressure in the primary system. When cooling by the SGs was restored and the primary system pressure reached the pressuriser relief valve closing threshold, a third failure occurred: the pressuriser relief valve received the command to close, but remained jammed in the open position, resulting in the loss of primary coolant via this valve. The operators who checked the pressuriser relief valve position indicator saw a “valve closed” indication. But this indication was false. This is because the indicator in the control room reflected the command received by the valve and not its actual position. The loss of primary coolant activated the safety injection system. The operators in charge of operational management of the plant focused their attention on the level of water in the pressuriser to prevent it from filling up. Faced with the rapid rise in the water level in the pressuriser, and believing the relief valve to be closed, the operators manually stopped the safety injection. The mental picture the operators had of the situation was false; they lacked direct information on the state of the reactor core. Melting of the fuel, then reactivation of safety injection Given the emptying of the primary cooling system, the fuel was no longer cooled. This led to degradation of the fuel, with a significant release of fission products from the fuel into the primary coolant. Two hours and fourteen minutes after the start of the accident, the alarm signalling high radioactivity in the reactor containment was activated. From this moment, the operators could no longer ignore that the situation was serious. The pressuriser relief valve was then closed, stopping the emptying of the primary cooling system. At this stage of the incident, new radioactivity alarms were activated, some situated outside the reactor building. Nine hours and fifty minutes after the start of the accident, a localised explosion of about 320 kg of hydrogen caused a pressure peak of about 2 bars in the reactor building, without causing any particular damage. It took the next twelve hours to purge the primary system of the majority of the hydrogen created by the oxidation of the Zircaloy and the incondensable fission gases released from the fuel during the accident. Wednesday 28 March 1979, 20:00 (8 pm) The accident in itself was over. It was nevertheless necessary to let several days go by before being able to exclude the risk of a hydrogen explosion. The damage suffered by the fuel elements was far greater than that imagined for the most severe design-basis accident considered for the installation. It was not until six years later, in 1985, that it was found that 45% of the fuel had melted, taking with it cladding and structural materials, forming what is called “corium”. Part of this corium, about 20 tonnes, flowed in liquid form into the bottom of the reactor vessel, fortunately without melting through it, possibly thanks to the forming of a space between the corium and the reactor vessel which would have allowed the cooling water to circulate in the vessel. Minimal consequences for the environment Despite the partial meltdown of the reactor core and the large release of radioactivity into the reactor containment, the immediate radiological consequences for the environment were limited. The reactor containment had effectively fulfilled its purpose. The low-level releases into the environment were caused by a system for pumping the primary cooling system effluents, which was kept in service. When unit 2 (TMI-2) suffered its accident in 1979, unit TMI-1 was disconnected from the network. It was put back into service in October 1985, despite public opposition, several court injunctions and technical and regulatory complications. In 2009, its operating license was extended by 20 years, that is to say until 19 April 2034. However, as the site had been losing money for several years, the licensee – Exelon – decided to stop operating it on 20 September 2019. Nuclear accidents and developments in nuclear safety and radiation protection • 13
RkJQdWJsaXNoZXIy NjQ0NzU=