Soichiro Ikuno

Iterative Solver for Linear System Obtained by Edge Element: Variable Preconditioned Method With Mixed Precision on GPU

The variable preconditioned (VP) Krylov subspace method with mixed precision is implemented on graphics processing unit (GPU) using compute unified device architecture (CUDA), and the linear system obtained from the edge element is solved by means of the method. The VPGCR method has the sufficient condition for the convergence. This sufficient condition leads us that the residual equation for the preconditioned procedure of VPGCR can be solved in the range of single precision. To stretch the sufficient condition, we propose the hybrid scheme of VP Krylov subspace method that uses single and double precision operations. The results of computations show that VPCG with mixed precision on GPU demonstrated significant achievement than that of CPU. Especially, VPCG-JOR on GPU with mixed precision is 41.853 times faster than that of VPCG-CG on CPU.

Detection of Critical Current Distribution of YBCO-Coated Conductors Using Permanent Magnet Method

We developed a non-destructive and contactless system for measuring the critical current (I_c) in YBa₂Cu₃O_7-δ (YBCO)-coated conductors by using a permanent magnet (Sm₂Co₁₇). This I_c measurement method is based on the repulsive force (F_r) between the magnet and the shielding current in high-temperature superconductor-coated conductors. We measured F_r using a high-resolution load sensor and found that accurate F_r could be determined without the effect of thick copper film on the YBCO thin film and Hastelloy tape of the substrate. We can determine I_c from F_r0, which is the maximum repulsive force determined from an extrapolated value of the F_r vs. L curve for L = 0 mm, described in our previous paper. We show that the permanent magnet method can be used to determine the longitudinal I_c distribution and large-scale defects in YBCO-coated conductors. In addition, it turned out from this experiment that the permanent magnet method was effective to rapidly measure the longitudinal I_c distribution of long-scale coated conductors.

Analysis of Measurement Method for Critical Current Density by Using Permanent Magnet

The Permanent Magnet method for measuring critical current density in high-temperature superconducting thin film is numerically and experimentally investigated. Numerical results show that the critical current density is approximately proportional to the maximum repulsive force. Furthermore, the maximum repulsive force increases as the value of diameter of permanent magnet increases. These tendency is also observed in the experimental investigation. Finally, the spatial distributions of the shielding current density are reproduced by the system experimentally.

Numerical Simulation of Inductive Method and Permanent-Magnet Method for Measuring Critical Current Density

Two types of contactless methods for measuring the critical current density have been investigated numerically. For the purpose of reproducing the experimental results numerically, the behavior of the shielding current density is formulated by use of the current-vector-potential method. A numerical code has been developed for solving an initial-boundary-value problem of the resulting integral-differential equations and, by use of the code, both the inductive method and the permanent-magnet method are simulated successfully.

Large-Scale Simulation of Electromagnetic Wave Propagation Using Meshless Time Domain Method With Parallel Processing

The large-scale simulation of the electromagnetic wave propagation using meshless time domain method (MTDM) is numerically investigated. Moreover, compute unified device architecture (CUDA) and OpenMP is adopted for parallelization technique to reduce the computation time. The results of computation show that the execution time of the time evolution calculation on GPU is 8.8 time faster than that of CPU. In addition, the execution time of the shape function generation procedure can be speedup about 7842 times by proposed scheme and OpenMP.

Parallelization of Finite Element Analysis of Nonlinear Magnetic Fields Using GPU

The acceleration of a nonlinear magnetic field analysis by parallelizing the finite element method (FEM) is examined using the graphics processing unit (GPU). It is shown that the speedup of the magnetic field analysis is realized by parallelizing the variable preconditioned conjugate gradient (VPCG) method. The Jacobi over-relaxation (JOR) method, conjugate residual (CR) method and conjugate gradient (CG) method are also applied in the variable preconditioning. The results of computations demonstrate that VPCG using the GPU significantly improve the performance. Especially, CG applied by variable preconditioned on GPU is 39 times faster than ICCG on a CPU.

Virtual Voltage Method for Analyzing Shielding Current Density in High-Temperature Superconducting Film With Cracks/Holes

An accurate numerical method is proposed for calculating the shielding current density in a high-temperature superconducting film containing defects. If the initial-boundary-value problem of the shielding current density is formulated by the

Influence of Method for Imposing Essential Boundary Condition on Meshless Galerkin/Petrov–Galerkin Approaches

T

Investigation on 3-D implicit FDTD method for parallel processing

-method, integral forms of Faradays law on defect surfaces are also imposed as natural boundary conditions. However, the conditions are not satisfied exactly by a numerical solution and their residuals develop intolerably with a decrease in the film thickness. In order to resolve this problem, the following method is proposed: virtual voltages be applied along the defect surfaces as to have the natural boundary conditions numerically satisfied. A numerical code is developed on the basis of the proposed method, and the influence of a crack on the inductive method or the permanent-magnet method is numerically investigated.

Parallel algorithm for meshfree radial point interpolation method on graphics hardware

Modified meshless methods are proposed by choosing trial and test functions from a variety of functional spaces. The accuracy of the proposed methods is evaluated for two ways of imposing essential boundary conditions. The results of computations show that the types of the test functions do not influence on the accuracy. In contrast, the weight functions for the moving least square approximation should not be chosen as trial functions to ensure the accuracy. Furthermore, the influence of choosing the ways for imposing essential boundary conditions on the accuracy was not detected