The determination of the effectiveness of nuclear reactor control rods by means of exponential experiments

Abstract

The conditions under which an exponential assembly may be used to determine control rod effectiveness are considered. It is concluded that a necessary condition for measurements to be analysed by simple theory is that the neutron energy spectrum perturbation produced by the rod should not extend over the whole volume of the assembly: the unperturbed asymptotic spectrum should be regained within the pile. Experiments show this to be the case for the largest rod used in the present experiments. Neutron balance equations in two energy groups for a square subcritical assembly with a cylindrical absorber along the axis are outlined. Measurements made in a graphite-natural uranium exponential pile are presented for boron steel, cadmium and mild steel control elements, all in the form of circular tubes ranging from 0.625 to 3.375 in. outer diameter. The results are discussed in terms of thermal neutron extrapolation lengths derived from Kushneriuk & McKay (1954). They indicate that absorption of thermal neutrons accounts for some 80% of the total effect in the case of boron steel, a larger proportion for cadmium and almost the whole effect for mild steel. The effect of placing moderating materials in the centre of the rods is investigated. For boron steel and cadmium increases in effectiveness of up to 20% are observed with moderating inserts of a hydrogenous material, while graphite is found to produce no significant change. Hydrogenous inserts increase the effectiveness by up to 40% with mild steel rods, but the effect appears to be attributable mainly to absorption of thermal neutrons by the inserts, rather than to moderation of fast neutrons. The results show very good internal consistency, supporting the validity of the technique. The sensitivity of the method appears comparable with that of measurements on a critical system with a similar lattice.