The Relation between Experiment and Theory in Reactor Physics

The end-product of theoretical and experimental reactor physics work is a detailed knowledge of the neutron physics behaviour of a power reactor, including the reactivity, power distribution and control characteristics. The problem may be divided into two parts. First, the properties of simple geometric configurations of homogeneous or uniform lattice structure and secondly the effects of departures from the idealized system, including the effects of control rods and other absorbers. Alternative approaches to the first problem are possible. For instance, an approximate physical model can be used and experiments carried out covering a range of all the variables. The experiments are then correlated using the simple model and adjusting the nuclear data to give agreement. This approach has been used successfully for natural uranium- fuelled graphite- moderated reactors but is difficult to apply to enriched thermal reactors particularly if more than one fissile material is to be considered. Similarly, for fast reactors good agreement, at least for the smaller systems with relatively hard spectra, can be obtained by ''one group'' calculations in which all neutrons are assumed to have an energy which is associated with a single effective cross-section. The other extreme is to use basic nuclear data and the best physical model which can be conceived with present-day knowledge. Such an approach is made possible by the increased capacity of digital computing machines and by our improved knowledge of nuclear data. The calculations are required to predict the precise physical quantities measured in the experiments. This approach has been used in the liquid moderated reactor field where the experiments are compared with the most elaborate theory and with a theoretical model sufficiently detailed to account for the physical processes involved using basic data, yet adequately rapid in application to be a useful method for design survey purposes. Similar approaches are possible for lattice irregularities. The relative merits of making measurements in subcritical and critical systems are discussed indicating those experiments for which a critical reactor is required. The discussion embraces graphite, heavy-water and light-water thermal reactors, and fast reactors. (author)