William S. Andrews

Artificial neural network models for volatile fission product release during severe accident conditions

Abstract Artificial neural network (ANN) models have been developed to predict the release of volatile fission products from both Canada deuterium uranium (CANDU) and light water reactor (LWR) fuel under severe accident conditions. The CANDU model was based on data for the release of 134 Cs measured during three annealing experiments (Hot Cell Experiments 1 and 2, or HCE-1, HCE-2 and metallurgical cell experiment 1, or MCE-1) at Chalk River Laboratories. These experiments were comprised of a total of 30 separate tests. The ANN established a correlation among 14 separate input variables and predicted the cumulative fractional release for a set of 386 data points drawn from 29 tests to a normalized error, E n , of 0.104 and an average absolute error, E abs , of 0.064. Predictions for a blind validation set (test HCE2-CM6) had an E n of 0.064 and an E abs of 0.054. From this 14 variable ANN model, a pruned version utilizing only the 6 most significant variables was trained to provide comparable predictions. An ANN model was also developed for LWR fuel, based on data from the vertical induction (VI) series of tests (VI-2 to VI-5) conducted at Oak Ridge National Laboratory. Predictions for data not used in ANN training had an E n of 0.045 and an E abs of 0.059. A methodology is presented for deploying the ANN models by providing the algorithms for trained ANNs and the corresponding connection weights. Finally, the performance of the full ANN CANDU model was compared to a fuel oxidation model developed by Lewis et al. and to the US Nuclear Regulatory Commissions CORSOR-M.

Measurements of the Temperature Inside an Explosive Fireball

This paper discusses the development of a fiber optic probe that can obtain temperature measurements from the interior of explosive fireballs, which are generated when unreacted detonation products react with oxygen in the surrounding air. Signatures of the thermochemical environment and chemical species involved can often be deduced from their light emissions, but the limited optical depth of fireballs means that remote sensing techniques can only sample emissions from the outer shell. By developing a protected fiber optic probe that can be placed adjacent to an exploding charge, giving it the ability to become enveloped by the fireball, the thermal radiation from the interior of the fireball can be sampled. Measurement from five shots using Detasheet-C explosives were carried out and could be obtained over the course of about 20 ms. Blackbody-type radiation with temperatures in the 1600 K to 1900 K range were observed, peaking at about 1850 K after 12 ms. The magnitude and time behavior of the temperature was not significantly different when taken at different locations within the fireball, indicating that temperature is fairly uniform throughout. The lack of specific spectral emission lines implies that in the interior of the fireball any combustion that occurred was probably primarily with carbonaceous soot, though differences in optical depth at different locations in the fireball indicate that it was much more fuel-rich closer to the center.

Modeling of cesium release from light water reactor fuel under severe accident conditions

Several empirically based models of fission product release, recently developed at various laboratories for severe reactor accident conditions, have been compared with the measured cesium release from light water reactor fuel in the 6 series of experiments performed at the Oak Ridge National Laboratory. The models under consideration treat the underlying process of release by first-order kinetics or by classical diffusion theory. In addition, a state-of-the-art approach using an artificial neutral network is evaluated.

Modeling the dispersion of radionuclides released during reactor accidents aboard nuclear-powered vessels

A radioactive aerosol and vapor dispersion model, RADiM, has been developed to predict the distribution of acquired doses from radionuclides escaping into the atmosphere following a marine reactor accident. Designed as a decision aid for emergency response teams, RADiM is based on a Gaussian plume model, which considers decay and transmutation, as well as meteorological and terrain effects. An initially estimated source term can be adjusted for strength and composition, based on subsequent field survey measurements and gamma-spectroscopy. RADiM was validated using field release data and was compared to predictions from established codes. Possible accident scenarios were also examined.

A set-up, calibration, acquisition and analysis program (SCAAP) for gamma-ray spectroscopy

A computer program entitled SCAAP (Set-up, Calibration, Acquisition and Analysis Program) has been designed to make gamma-ray spectroscopy easily useable by staff in laboratories at the Slowpoke-2 Facility at RMC and in support of the Canadian Forces Nuclear Emergency Response Teams (NERTs). The former group utilises gamma-ray spectroscopy for neutron activation analysis (NAA), while the latter may need to have inhalation dose rates calculated. The intent of this program, written using Microsoft Visual Basic, is to provide a simplified interface between the operator and the spectroscopy equipment and to provide the calculations necessary to produce results quickly. There are five sections (Setup, Calibrate, Acquire, Analyse and NAA) of which the first four are linked. In these sections, a checklist of procedures is presented and automated for the user to set up and calibrate the equipment and then to analyse spectra to provide various dose rates. In the unlinked section, NAA, gamma-ray spectra are analysed to provide elemental concentrations in samples.