For the measurement of point kinetics parameters of zero power reactors many different methods have been widely used and described in the literature. These include Rossi-α, probability distribution, variance-over-mean, frequency analysis, and polarity correlation of neutron detector signals. Up to now kinetics parameters of a coupled two-point reactor model have been measured only by two noise analysis techniques: Rossi-α (correlation functions) and frequency analysis (power spectral density, coherence functions) using two detectors to obtain space-dependent neutron signals. This paper describes for the first time investigations of two-detector covariance measurements and their application to the determination of coupled kinetics parameters. This technique is compared with the Rossi-α method by theoretical considerations and measurements using identical detector signals in both analysis techniques. Results were also compared to those measured previously at the same reactor by frequency analysis. Theoretical formulas for the Rossi-α experiment and variance and covariance measurements in a symmetrical two-point reactor model were derived. The material properties and neutron lifetimes for each reactor zone were assumed equivalent. The transport time of neutrons from one zone to the other was neglected. The parameters of the model include the decay constant of prompt neutrons and the coupling reactivity of the two core zones. These parameters were determined by least-squares fitting the theoretical curves to the experimental data. The measurements have been performed at the Argonaut Reactor Karlsruhe (ARK) with a symmetrical two-slab core loading for different subcritical levels. Variances and covariances of neutron counts from detectors placed in both core regions were measured in two different ways. In the first method the two-dimensional probability distribution of the number of counts from the detectors was measured using a small digital computer. The computer was used to control the gating times of two specially designed counting registers and to calculate the first and second moments. From that the covariances were calculated in real time and displayed on a CRT as a function of the counting time interval. In the second method the variance and the covariance of the detector pulses were directly measured without using the probability distributions. For this purpose a new analyser based on a special calculational algorithm was developed. This analyser can be built easily with low hardware expense. The results obtained by the Rossi-α and variance method agreed quite well. In contrast to point reactor parameter determination, where Rossi-α and variance techniques are equivalent, the variance and covariance method appear superior to the Rossi-α method in applications to coupled reactors. Kinetics parameters of a two-node reactor model can be determined more accurately by the variance technique. The effect of dead-time losses in the signal channels or in the analyser on the measured variances was also studied. The variance technique was found to be more sensitive to dead-time than the Rossi-α technique. Therefore variance measurements at fast reactors yield incorrect results except special fast electronics is used in the pulse channels.
Read full abstract
Jun 1, 1977
H.H Krause + 6
Open Access