Speaker
Description
A characterisation of spent nuclear fuel (SNF) is essential for a safe, secure, ecologic and economic handling, transport, interim storage and final disposal of SNF assemblies. Several quantities are of importance: neutron and γ-ray emission rates and spectra, decay heat and the inventory of fissile nuclides. These sources can only be obtained by theoretical calculations. Evidently, the theoretical calculations need to be validated and realistic confidence limits have to be determined. Unfortunately, experimental data to directly validate results of decay heat rate calculations are extremely scarce.
At the interim storage facility CLAB in Sweden a calorimeter is installed that is designed to measure the thermal heat output or decay heat rate of SNF assemblies. The calorimeter is operated in static mode using the rate of water temperature increase to establish the thermal output of an SNF assembly. The translation of the temperature increase to heat output is based on calibration measurements using an electrical heater simulating a SNF assembly.
In Ref. [1] results obtained with this calorimeter were used in a blind test exercise. Calculations of the decay heat rate of five assemblies were performed by different organisations using different codes and nuclear data libraries resulting in 31 results for each assembly. This exercise suggests a systematic difference between calculated and experimental decay heat rate. The experimental results are given with an uncertainty of about 5%, without a detailed evaluation of the different random and systematic effects contributing to the measurement uncertainty. Therefore, the origin of this difference, e.g. a possible bias due to nuclear data, cannot be identified.
To improve the predictive power of the calculation it is of primary importance to assess the performance of this calorimeter and to define reliable uncertainty margins. In this contribution a detailed study of the measurement principle of the calorimeter is presented by identifying the main metrological parameters of the measurement process. The study includes an analytical model to describe the heat transport from the assembly to the surrounding pool and Monte Carlo simulations to describe the gamma-ray transport from the assembly to the pool. The analytical model is used to optimise the parameterisation of the temperature increase with time in the water of the calorimeter and to estimate the correction factor due to difference in heat capacity of the assembly and electrical heater. The MC simulations are carried out to estimate an additional systematic effect due to heat losses caused by gamma-rays escaping from the calorimeter. In addition a variance analysis is applied to quantify additional uncertainties due to random effects.
[1] Jansson et al., “Blind Benchmark Exercise for Spent Nuclear Fuel Decay Heat”, submitted to Nucl. Sci. Eng.