Reactivity faults occur if there is an uncontrolled in­crease in the reactivity of the reactor either over the entire core, or locally. Such an increase inevitably gives power and hence temperature transients which must be limited by the protective equipment to less than the melting point of magnox. Indeed, because the strength of magnox decreases substantially at a few degrees centigrade less than 640°C and because full coolant mass flow conditions usually exist, the acceptable maximum temperature is reduced to avoid distortion of the fins with a consequential reduction of heat transfer coefficient.

The reactivity increases because the control rods, either as groups of rods or locally as a few rods, run out in an uncontrolled fashion. The reactivity release rate depends upon the worth of the rods running out, and their speed of withdrawal. Both of these aspects are limited by design, by restricting the size of rod groups and the speed of withdrawal and by using in­terlocks to prevent the main rod groups being moved together. There are two types of reactivity faults to be considered, those where there is a uniform increase of reactivity over the core, symmetric faults, and those where only a few rods run out, asymmetric faults. These need to be considered separately using slightly different techniques because the protection for each 4s organised in a different way. In each case, however, it is necessary to demonstrate that the probability of exceeding the limiting clad temperature is acceptably low. As in the case of the depressurisation faults, this is done by considering the magnitude of the uncertainties in the manufacture of the fuel, the heat transfer properties and the model used in the study. For those faults where the gas circuit remains intact, the acceptable probabilities that any fuel temperature will reach the limit in the event of the fault occurring is 1 in 1000.