A Bayesian approach for measuring nuclear material in radioactive waste using photofission

Abstract

Photofission provides a mean for measuring nuclear material in large and dense radioactive waste packages. High-energy interrogating bremsstrahlung photons produced by a LINAC are indeed able to induce fissions in the depth of the waste matrix. In addition, a scintillation detector can detect high-energy delayed gamma rays that are able to escape the waste package. Scintillation detectors also allow performing measurements during photon irradiation, between LINAC pulses, to measure delayed gamma rays from short-lived fission products. Photofission measurement with a LaBr3 scintillation detector can help determining, in a Bayesian frame, the mass and position of actinides in radioactive waste packages. In addition to counting statistics, the model takes into account photofission yields and waste matrix density uncertainties. The model performances are numerically illustrated in the case of a 235 g238U lump placed inside a 50 cm radius drum homogeneously filled with 2.35 g/cm3 density concrete. In the most unfavourable angular position, with 32 drum rotations, when the 238U radial position is 10% and 90% of the waste drum radius, the expected relative uncertainty on the 238U mass is respectively 32% and 45%. The 238U angular position uncertainty is 10 mrad and its radial position can be determined with an expected uncertainty of 2 cm and 3 cm, when the 238U is respectively placed at 10% and 90% of the waste drum radius.