In the framework of a starting collaboration between CEA and Centrum výzkumu Řež, a pile noise experiment was performed in the low power nuclear research reactor LR-0, located in Husinec-Řež, Czech Republic. The LR-0 research reactor is a light-water, zero-power, pool-type reactor. It serves as an experimental reactor for measuring neutron-physical characteristics of VVER (Water-Water Energetic Reactor) type reactors.
A pile noise experiment (referred to as HLUK, which stands for “noise” in Czech) was designed based on neutronic simulations performed with CEA Monte Carlo code TRIPOLI-4. The objective of this experiment was to measure the prompt neutron response in a reference configuration of the core and then to obtain the kinetic parameters (delayed neutron fraction and generation time) using well-established pile noise techniques.
Two high sensitivity 235U fission chambers were used, one in the center of the core and the other in the reflector. The signals were recorded with a dual channel current digitizer specially developed by CEA to acquire pile noise. A reactor dosimetry campaign using metal foils was performed to measure the reference core integral fission rate. Indeed, this is a key parameter to obtain the kinetic parameters from the raw signals.
In addition to measuring the kinetic parameters in LR-0 reference configuration, the impact of the distance the reflector detector and fuel was studied by progressively moving one detector away from the fuel elements. The objective was to find the maximum distance achievable without losing the pile noise signal.
In this paper, experimental results are presented and discussed, with a focus on the uncertainty analysis. The reference experiment was analyzed based on a fit of the power spectrum density (PSD) of the signals. The kinetic parameters are compared to the calculated values issued by TRIPOLI4 with JEFF-311 nuclear data library: 781.2(10) pcm for the delayed neutron fraction and 40.80(1) µs for the generation time of prompt neutrons.
From the various detection setups, we were able to construct an array of experimental curves coming from the same post-processing method. A graph showing the results of fit along with the distance between the detector and the fuel is then discussed. Recommendations are made on the how to place the detectors to optimize the pile noise experiment in terms of results quality and measurement duration.