Optimal estimation of parameters of dynamical systems by neural network collocation method

In this paper we propose a new method to estimate parameters of a dynamical system from observation data on the basis of a neural network collocation method. We construct an object function consisting of squared residuals of dynamical model equations at collocation points and squared deviations of the observations from their corresponding computed values. The neural network is then trained by optimizing the object function.The proposed method is demonstrated by performing several numerical experiments for the optimal estimates of parameters for two different nonlinear systems. Firstly, we consider the weakly and highly nonlinear cases of the Lorenz model and apply the method to estimate the optimum values of parameters for the two cases under various conditions. Then we apply it to estimate the parameters of one-dimensional oscillator with nonlinear damping and restoring terms representing the nonlinear ship roll motion under various conditions. Satisfactory results have been obtained for both the problems.     

The recovery and recrystallization of cold rolled V-W-Ti alloys

Several V-W-Ti alloys with about 50% reduction in thickness by cold rolling were isochronally annealed at temperatures from 100 to 1100 °C. Hardness was measured after the annealing. By comparison with V4Cr4Ti and V4Ti alloys, the V6WTi alloy was found to begin recovery at a temperature about 100 °C higher, while the full recrystallization temperature was around 900 °C for all of the alloys. Hardness decreased successively above 500 °C in V8W, as is the case in pure V. Impurity induced hardening was observed around 300 °C only in the alloys without Ti. Precipitation and the interaction between interstitial impurities and dislocations were supposed to be the main contributors to the different recovery and recrystallization behaviors of the alloys.     

Corrosion characteristics of reduced activation ferritic steel, JLF-1 (8.92Cr–2W) in molten salts Flibe and Flinak

Static corrosion tests were performed in molten salts, LiF–BeF₂ (Flibe) and LiF–NaF–KF (Flinak), at 500 °C and 600 °C for 1000 h. The purpose is to investigate the corrosion characteristics of reduced activation ferritic steels, JLF-1 (8.92Cr–2W) in the fluids. The concentration of hydrogen fluoride (HF) in the fluids was measured by slurry pH titration method before and after the exposure. The HF concentration determined the fluoridation potential. The corrosion was mainly caused by dissolution of Fe and Cr into the fluids due to fluoridation and/or electrochemical corrosion. Carbon on the surface might be dissolved into the fluids due to the corrosion, and this resulted to the decrease of carbide on the surface. The corrosion depth of the JLF-1 specimen, which was obtained from the weight losses, was 0.637 μm in Flibe at 600 °C and 6.73 μm in Flinak at 600 °C.     

LiErO2 formation on Er2O3 in static and natural convection lithium

Er₂O₃ is candidate material for insulating coating to prevent the magnetohydrodynamic (MHD) pressure drop in the self-cooled liquid Li blanket system. Although Er₂O₃ is stable material, detailed chemical behavior in liquid Li is not clear. Corrosion behavior of bulk Er₂O₃ in Li is investigated in static and flowing condition in the present study. After these tests, good compatibility of Er₂O₃ was confirmed and slight formation of LiErO₂ was detected by XRD analysis. This chemical behavior did not change in a static and flowing tests, however some of the corrosion product of LiErO₂ was removed easily by the Li flow. Intensity of LiErO₂ peaks in XRD spectrum suggests that the temperature gradient may affect the reaction rate in the natural convection loop. Since corrosion rate of Er₂O₃ is very small, slight change in state will be important information to evaluate lifetime of coating.     

Growth of β-FeSi2 thin films on β-FeSi2 (110) substrates by molecular beam epitaxy

We have investigated the preparation of β-FeSi₂ substrate and growth condition of β-FeSi₂ thin film on β-FeSi₂ (110) substrate by molecular beam epitaxy. The surface of the substrate was prepared by a wet-etching using HF(50%):HNO₃(60%):H₂O = 1:1:5 solution at 25 °C. It is clear that the optimal etching period to obtain a flat surface was 3 min. The β-FeSi₂ thin film with streak RHEED pattern was obtained at Si/Fe flux ratio of 2.9. Average surface roughness (Ra) of the β-FeSi₂ film was about 0.5 nm in 5 × 5 μm² area.     

Overview of the TITAN project

This is an overview paper of the joint TITAN (Tritium, Irradiation, and Thermofluid for America and Nippon) project of the Japan–USA fusion cooperation program. The objectives, tasks structure, and technical highlights of the TITAN project are presented, as well as the direction of the project toward its end in March 2013.     

Genomic classification of fish nodaviruses by molecular phylogenetic analysis of the coat protein gene

A molecular phylogenetic analysis of 25 isolates of fish nodaviruses, the causative agents of viral nervous necrosis of marine fish, was performed based on the nucleotide sequences (427 bases) of the coat protein gene. These fish nodaviruses were classified into four clusters: tiger puffer nervous necrosis virus, striped jack nervous necrosis virus, berfin flounder nervous necrosis virus, and red-spotted grouper nervous necrosis virus.     

Parallel multipliers with NOR gates based on G-minimum adders

Based on the adders with a minimum number of NOR gates discussed in Ref. 1, parallel multiplers are designed in Ref. 2, using fewer gates, fewer connections, and faster operation than conventional multipliers based on carry-save adders. In this paper, parallel multipliers of NOR gates are designed, by expressing two numbers to be multiplied in the sign and magnitude representation unlike those in Ref. 2. The multipliers in this paper are advantageous over the carry-save adder-type multipliers known to date, in terms of the number of gates and connections.This work is supported in part by the National Science Foundation under Grant No. MCS77-09744 and MCS81-08505.