Matrix effect correction in active neutron interrogation for residual fissile mass assessment in radioactive waste drums

Abstract

Metallic waste in the form of shells and nozzles remains following the reprocessing of spent fuel at La Hague plant (France). Before the final disposal, drums containing the waste are transported to the compaction facility where its volume is reduced by a factor of 5. With the objective of controlling the criticality/safety levels, an active neutron interrogation system at the entrance of the facility is used to assess the residual fissile mass remaining in the waste, which relies on prompt fission neutron detection (known as the differential die-away technique, DDA). The measured signal is proportionally linked to the fissile mass by a calibration coefficient. However, two effects can produce an inaccurate prediction. The number of induced fissions for the same fissile mass varies according to the waste matrix composition, particularly the neutron slow-down and absorption ratio. Secondly, the presence of fissile clusters can impact the neutron interrogation due to an increase in self-shielding effect. This work presents a matrix effect correction method based on the use of internal flux monitors (sensitive to the neutron absorption ratio) and the transmission signal (sensitive to the slow-down ratio). It relies on a numerical model of the whole measurement cell developed with the Monte-Carlo N-Particle transport (MCNP) code. The calibration coefficient of 72 different matrices, representative of the waste produced at La Hague, were simulated and a statistical model using a multilinear regression was then established. This simulated-based approach provides a more robust and comprehensive estimation of the residual fissile mass without impacting accuracy thanks to the flux monitors and the transmission signals.