Matrix Effect Correction With Internal Flux Monitor in Radiation Waste Characterization With the Differential Die-away Technique

Radioactive waste drums filled with compacted metallic residues (spent fuel hulls and nozzles) produced at AREVA La Hague reprocessing plant are measured by neutron interrogation with the Differential Die-away measurement Technique (DDT). The purpose is to assay fissile material quantities present in radioactive waste packages. In the future, old hulls and nozzles containing Ion-Exchange Resin (IER) will be measured. IERs provide moderating properties to the matrix, not encountered during the current measurement. In this context, the Nuclear Measurement Laboratory (NML) of the CEA Cadarache has been asked by AREVA NC to explore the possibility of implementing a matrix effect correction method, based on internal monitor (
³
He proportional counter) signal correlated to the matrix effect. In order to validate this method, a benchmark was performed with PROMETHEE 6 R&D measurement cell at the NML, with a similar cavity configuration to that of the industrial station. An experimental design on two main factors regarding the matrix effect (absorbing and moderating ratios) has been studied. Considering the variation range of both factors for old waste measurement, 5 test matrices have been defined. They have been measured in PROMETHEE 6 and simulated using the particle transport code MCNP. Tests have been carried out experimentally using
²³⁵
U platelets. Results show that the experimental internal monitor is sensitive to the matrix but not to the fissile material presence and location. In addition, differences between experiment and model are satisfactory (<; 10%), in terms of prompt calibration coefficient (useful signal of fissile materials) and internal monitor signal, considering the complexity of the measurement method and numerical model, and the large range of moderator and absorption ratios. The relationship between the prompt calibration coefficient and the internal monitor signal observed in PROMETHEE 6, both for experience and model, can be fitted with a similar function as the industrial measurement cell, the correlation of which being established by numerical simulation. Regressions from experimental and modelling are almost superimposed.