Abstract :
The necessity to cool a reactor safely after it is tripped has always made the heaviest safety demands on the electrical system, independent of reactor type. In earlier gas-cooled reactor stations sufficient essential plant was provided to cope with the maximum breach which could credibly occur in the reactor pressure circuit, with a spare plant allowance to cover maintenance and fault outage assuming loss of grid at this time. Now, cooling plant is designed to meet reliability criteria. To make sure it does this, its performance is examined systematically, over the range of possible faults, and its failure modes analysed. These modes include random and common-mode component failure, in combination with hazards such as high winds, earthquakes, fires or turbine disintegration. To safeguard the cooling function in such circumstances it has been necessary to subdivide the essential electrical plant and its associated mechanical drives into functionally independent subsystems, called trains. Because the changed safety requirements have involved more plant, more widely separated, they have increased costs. Such cost increases have been minimised by balancing the risks over the systems, integrating electrical and mechanical plant layout in the reactor buildings, and reducing the size of the essential prime movers. This size reduction has been made possible by interposing frequency converters between them and the largest (gas circulator) drives. Expensive site fitting is reduced and the possibility of fouls, or failures to segregate is avoided by means of a comprehensive (1:20) scale model
