Relative mass resolution technique for optimum design of a gamma nondestructive assay system

Abstract

Nondestructive assay(NDA) is a widely used nuclear technology for quantitative elemental and isotopic assay. Nondestructive assay is performed by the detection of an identifying radiation emerging from the sample, which can be unambiguously related to the element or isotope of interest. In every assay we can identify two distinct factors that lead to measurement uncertainty. We refer to these as statistical and spatial uncertainties. If the spatial distribution of the analyte and the matrix material in the sample are known and fairly constant from sample to sample, then the major source of measurement uncertainty is the statistical uncertainty resulting from randomness in the counting process. The spatial uncertainty is independent of the measurement time and therefore sets a lower limit to the measurement uncertainty, which is inherent in the assay system in conjunction with the population of samples to be measured. The only way to minimize the spatial uncertainty is an optimized design of the assay system. Therefore we have to decide on the type and number of detectors to be used, their deployment around the sample, the type of radiation to be measured, the duration of each measurement, the size and shape of the sample drum. The design procedure leading to the optimal assay system should be based on a quantitive(RMR:Relative Mass Resolution) comparison of the performance of each proposed design. For NDA system design of low level radwaste, a specific purpose Monte Carlo code has been developed to simulate point-source responses for sources within an assayed radwaste drum and to analyze the effect of scattered gammas from higher energy gammas on the spectrum of a low energy gamma-ray. We could use the well-known Monte Carlo code, such as MCNP for the simulation of NDA in the case of low level radwaste. But, MCNP is a multi-purpose Monte Carlo transport code for several geometries which requires large memory and long CPU time. For some cases in nuclear applicat...