The fuel in a PWR is uranium dioxide in the form of solid pellets about 13.5 mm long and 8.2 mm in diameter. These are contained within zircalloy-4 (an alloy of zirconium) tubes about 3.7 m long, the full length of the reactor core. A fuel assembly consists of 264 of these tubes or pins, together with 25 other tubes used variously to guide absorber rods (control rods) or to house instrumentation, and are arranged in a 17 x 17 square array. There are about 193 such assemblies in the reactor core. The uranium is en­riched in the isotope U-235 to overcome the reactivity taken up by the natural (light) water which, unlike magnox and AGR reactors, acts both as coolant and as neutron moderator. The zircalloy is not itself a significant absorber of neutrons. The neutron absorber boron, in the form of boric acid, is added to the water coolant so that its concentration can be steadily decreased to compensate for loss of reactivity as the fuel is irradiated.

In common with other reactor types, fission pro­ducts are produced during irradiation. The majority of these, under normal conditions, remain trapped within the fuel pellets but a fraction of them escape into the narrow gap between the pellet and the zirc — ailoy clad, and between the slightly concave ends of the pellets themselves. An additional source of radio­activity which has to be taken into account, is that arising from corrosion products which become acti­vated by irradiation as they are carried through the core by the coolant. It is ultimately the purpose of the fault studies to show that the dose received as a consequence of release of some of these radioactive materials is acceptably low given the predicted fre­quency of occurrence.

Fault categories considered are as follow-s:

• Inadvertent reactor trip.

• Increase in heat removal by the secondary system.

• Decrease in heat removal by the secondary system.

• Electrical supply system faults.

• Decrease in coolant flow in the reactor coolant system.

• Reactivity faults.

• Pressure transient induced by changes of reactor coolant inventory,

• Decrease in reactor coolant inventory (loss of coolant).

• Anticipated transients without trip (ATWT).

• System-related faults.

• Radioactive releases from systems not within the nuclear steam supply system.

• Internal and external hazards.

From these broad categories a detailed schedule of initiating faults is drawn up and contains some 130 faults. Fault sequences, i. e., an initiating fault fol­lowed by other independent failures, are then identi­fied which are within the design basis; essentially those with a predicted frequency of more than about 10 ^7/ year.

To avoid the need to carry out transient analysis for every fault sequence, a prohibitive task, bounding faults or limiting design basis faults are selected. The selected sequencies are those which lead to:

(a) The most onerous demands on plant components such as the containment and pressure vessel, or

(b) The most onerous demands on the reliability of safeguards systems, or

(c) The potential for the largest release of radioactivity to the environment.

Transient analyses are carried out for the identified limiting design. The transients are assessed against defined limits, as relevant to (a), (b) or (c) above. The limits depend to some extent on the fault being considered. For example, for ‘frequent’ faults (defined as initiating faults of frequency in excess of 10 _3 per year — for which diversity of protection and safeguards systems is generally provided), it is ne­cessary to show that fuel clad failure will not occur. For less frequent faults (‘infrequent faults’), clad fail­ure is acceptable provided the fuel itself does not fail disruptively such that ability of the core to be cooled could be affected. The limits for the latter in the case of loss of coolant accidents, for example, are that the peak clad temperature does not exceed 1204°C (2200°F), that oxidation of the clad at any point does not extend beyond 17°7o of the clad thick­ness and that not more than l^o of all the clad in the core is oxidised. These limits ensure that significant embrittlement of the clad does not occur and that hydrogen production is insufficient to lead to an ex­plosive concentration.

In principle, the categories of faults that require assessment for a PWR are similar to those for mag — nox and AGRs. However, the behaviour of the PWR is, in general, very different because of the funda­mentally different design and the fact that the water coolant can change phase and become steam if pres­sure is reduced, the change in phase causing a rapid decrease in heat transfer from fuel to coolant. In simplified terms the characteristics of faults are de­scribed in the following paragraphs.