Masakuni Narita

A genetic algorithm approach to optimization for the radiological worker allocation problem.

This paper describes a new approach to the radiological worker allocation problem using a multiple objective genetic algorithm. The worker allocation problem in radiological facilities involves various types of constraints and even mutually conflicting ones, such as individual dose limits, working time limits, etc. A major difficulty of this highly working time limits, etc. A major difficulty of this highly constrained problem is the way of finding an optimal solution in the huge search space where a large proportion of solutions are not feasible because some of the constraints cannot be satisfied. The paper proposes a model of evolution to establish an optimal assignment efficiently, based on the biological insights into the evolutionary process and heuristic ideas. The experimental results show a very rapid evolution to produce feasible solutions, and the application of multiple evaluation functions converges the feasible solutions to good ones. The genetic algorithm approach was found to be superior to the goal programming and simplex methods.

A hierarchical domain decomposition boundary element method applied to the multiregion problems of neutron diffusion equations

AbstractA technique is presented for solving neutron diffusion equations with the boundary element method (BEM) based on a hierarchical domain decomposition technique. In this method, the reactor domain is decomposed into homogeneous regions and the boundary condition on the common boundary of regions is initially assumed. The neutron diffusion equation is solved iteratively at two levels of hierarchical structure: First, BEM is applied to solve the neutron diffusion equation of each homogeneous region under the given assumed boundary conditions and an assumed multiplication factor. Then, these assumed values are modified to satisfy the continuity conditions for the neutron flux and neutron current.The proposed technique is useful for multiregion problems with a large number of regions of complex geometry, where the finite difference approximation cannot be applied properly.

Memorial-Index-Based Inverse Kinetics Method for Continuous Measurement of Reactivity and Source Strength

A new neutron inverse kinetics method has been proposed to compute both reactivity and source strength using only noise-containing reactor power signals in a subcritical state. The method is applicable to cases when the effective source strength varies with a change in reactivity. A newly defined kinetics parameter, “Memorial Index”, is introduced to detect properly the start and end of a transient and to exclude time intervals that are ill-posed for numerical source strength estimation analysis based on the grouping method. Results for a numerical simulation analysis demonstrate that the revised analytic procedure incorporated with the Memorial Index enables one to continuously calculate precise values of both source strength and reactivity during a complicated subcriticality transient.

Anisotropic annealing of fission fragments in synthetic quartz

Abstract A study on the thermal annealing behavior of fission fragments in synthetic quartz crystals was performed for application of synthetic quartz to track detectors under high temperature circumstances: x -, y -, and z -cut quartz plates were exposed to fission fragments from 252 Cf and were annealed up to 750°C. It was found that track retention depended on the crystallographic structure: the annealing temperature, where the number of etched tracks was sharply reduced, was 550°C for the x - and y -cut plates and 750°C for z -cut plates. Track sizes decrease and disperse with reductions in the track retention.

An application of the domain decomposition method into the boundary element method for solving the multi-region neutron diffusion equation

This paper presents a technique for solving the neutron diffusion equation with the boundary element method based on the domain decomposition method. In this technique, the domain region is decomposed into homogeneous regions. The boundary conditions on the common boundary of decomposed regions and the multiplication factor are initially assumed. The boundary conditions on the other boundary are given. The neutron diffusion equation is solved iteratively at two levels of a hierarchical structure: first, the boundary element method is applied to solve the neutron diffusion equation of each homogeneous region under given assumed boundary conditions and multiplication factor. These assumed values are then modified to satisfy the continuity conditions for the neutron flux and neutron current. The proposed technique is useful for multi-region problems with a large number of regions.

Calibration measurements for the efficiency and response function of the NE213 scintillator by a /sup 252/Cf spontaneous fission source-application of a two-dimensional pulse height analysis system

The authors describe calibration measurements for the efficiency and response function of the NE213 scintillator as an application of a two-dimensional pulse height analysis system. In this method, a /sup 252/Cf spontaneous fission source with a continuous spectrum is employed by taking each fission event as a pulsed neutron source, and the efficiency and response function of the NE213 scintillator can be estimated by two-dimensional measurements of the time-of-flight spectrum (neutron energy spectrum) and pulse height spectrum. >

Energy-Dependent Neutron Detection Sensitivity of Cellulose Nitrate Track Detector

A cellulose nitrate track detector, LR-115 (Type II), was studied experimentally to observe its neutron energy response when a radiator, such as fissionable material, is absent. The detector was exposed to monochromatic neutrons from the D (d, n) He reaction produced by a Van de Graaf accelerator. Through-etched and almost through-etched tracks in the detector were counted by two different methods: 1) optical microscope counting with an automatic image analyzing system and 2) spark counting. The detection sensitivity was obtained in the neutron energy region lower than 5 MeV and it was found that only α-particles from (n, α) reactions could be detected in both counting methods. We derived the counting probability of α-tracks induced in the detector and the threshold energies for neutron detection.

An Application of Information Theory to Pulsed Nuclear Radiation Measurement

A new method for pulsed radiation measurements is studied by an application of the information theory of discrete systems. It is derived from entropy concept that in order to measure the photon density with the maximum conversion rate of a counter subjected to pulsed photon beams, the density λ of a Poisson source must satisfy the relation of λb=ln 2 when the resolving time τ of the counter lies between the pulse width b and the period (1/f). Actual devices using feedback circuits for satisfying the above condition are proposed.

Response of a wireline pickup for subnanosecond electron beam pulses

The response of a wireline-type beam monitor was measured for electron beams with energies of 28 MeV and pulsewidths of 20 ps from an electron linear accelerator. The output waveform from the wireline pickup was calculated according to the general formulation for responses of beam monitors based on electromagnetic interaction. In the calculations, a transfer function H(/spl omega/) and an effective length l/sub eff/ are introduced to express all the effects not included in the analysis. The function H(/spl omega/) is assumed to be Gaussian, and the FWHM of H(/spl omega/) and l/sub eff/ is experimentally determined. The effects of the termination, wireline length, height, and angle between the wireline and the beam direction are discussed. The test supported the analysis and, for the standard pickup, a charge sensitivity of 20 V/nC and an FWHM of H(/spl omega/) of 55 ps are obtained.

Energy-dependent neutron sensitivity of polycarbonate recoil track detector

The energy-dependent absolute neutron sensitivity of a polycarbonate recoil track detector was measured for 4-6.5 MeV. The detector was irradiated with neutrons from a D(d,n)He reaction with the ETL Van de Graaff accelerator. After irradiation and chemical etching, the number of etched tracks was counted with an automatic track-counting system. The theoretical energy-dependent sensitivity was calculated using the local energy-loss theory of Katz and Kobeitch and agreed well with the observed results.