Yoshitaka Wada
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yoshitaka Wada has a total of 32 co-authors. Here are the top co-authors by number of shared publications:
- Ryuji Shioya
Toyo University
2 shared publications - Akihiro Mano
Japan Atomic Energy Agency, Tokai, Ibaraki 3191195, Japan
1 shared publication Akio Miyoshi
INSIGHT, Inc
1 shared publication- Akiyuki Takahashi
Tokyo University of Science
1 shared publication - Al Mahdi Remmal
Framatome, Paris 92400, France
1 shared publication - Amane Takei
Faculty of Engineering, University of Miyazaki, Miyazaki City 889-2155, Japan
1 shared publication
Latest Publications
- Open Access
- 1 January 2025
- Published by Japan Society of Mechanical Engineers in Transactions of the JSME (in Japanese)
- Vol. 91 (949) , 25-00102
Development of regularization method for physics problems using PINNs and data augmentation using equation discovery technique
This study proposes a regularization method for high-accuracy prediction by machine learning under small data conditions by combining Physics-Informed Neural Networks (PINNs) and data augmentation based on equation discovery using Sparse Identification of Nonlinear Dynamical Systems (SINDy). PINNs incorporate governing equations into the loss function to enforce physical consistency, while SINDy extracts mathematical models from data, enabling interpolation with continuity and physical interpretability. Validation using crack propagation data (25 points) demonstrated that the combination of SINDybased augmentation and PINNs reduced the sum of squared errors (SSE) by approximately 28% compared to linear interpolation. The effectiveness of the proposed method was also confirmed for different datasets (59 points). These results indicate that the method is particularly effective in improving learning accuracy for engineering problems where only small amounts of data are available.
- 1 January 2025
- Published by Japan Society of Mechanical Engineers in Transactions of the JSME (in Japanese)
- Vol. 91 (949) , 25-00102
Open AccessDevelopment of regularization method for physics problems using PINNs and data augmentation using equation discovery technique
This study proposes a regularization method for high-accuracy prediction by machine learning under small data conditions by combining Physics-Informed Neural Networks (PINNs) and data augmentation based on equation discovery using Sparse Identification of Nonlinear Dynamical Systems (SINDy). PINNs incorporate governing equations into the loss function to enforce physical consistency, while SINDy extracts mathematical models from data, enabling interpolation with continuity and physical interpretability. Validation using crack propagation data (25 points) demonstrated that the combination of SINDybased augmentation and PINNs reduced the sum of squared errors (SSE) by approximately 28% compared to linear interpolation. The effectiveness of the proposed method was also confirmed for different datasets (59 points). These results indicate that the method is particularly effective in improving learning accuracy for engineering problems where only small amounts of data are available.
- 6 December 2021
- Published by ASME International in Journal of Pressure Vessel Technology
- Vol. 144 (1)
Benchmark Analysis of Ductile Fracture Simulation for Circumferentially Cracked Pipes Subjected to Bending
To predict fracture behavior for ductile materials, some ductile fracture simulation methods different from classical approaches have been investigated based on appropriate models of ductile fracture. For the future use of the methods to overcome restrictions of classical approaches, the applicability to the actual components is of concern. In this study, two benchmark problems on the fracture tests supposing actual components were provided to investigate the prediction ability of simulation methods containing parameter decisions. One was the circumferentially through-wall and surface cracked pipes subjected to monotonic bending, and the other was the circumferentially through-wall cracked pipes subjected to cyclic bending. Participants predicted the ductile crack propagation behavior by their own approaches, including finite element method (FEM) employed Gurson–Tvergaard–Needleman (GTN) yielding function with void ratio criterion, are FEM employed GTN yielding function, FEM with fracture strain or energy criterion modified by stress triaxiality, extended FEM with J or ΔJ criterion, FEM with stress triaxiality and plastic strain based ductile crack propagation using FEM, and elastic-plastic peridynamics. Both the deformation and the crack propagation behaviors for monotonic bending were well reproduced, while few participants reproduced those for cyclic bending. To reproduce pipe deformation and fracture behaviors, most of the groups needed parameters that were determined to reproduce pipe deformation and fracture behaviors in benchmark problems themselves and it is still difficult to reproduce them by using parameters only from basic materials tests.
- 6 December 2021
- Published by ASME International in Journal of Pressure Vessel Technology
- Vol. 144 (1)
Benchmark Analysis of Ductile Fracture Simulation for Circumferentially Cracked Pipes Subjected to Bending
To predict fracture behavior for ductile materials, some ductile fracture simulation methods different from classical approaches have been investigated based on appropriate models of ductile fracture. For the future use of the methods to overcome restrictions of classical approaches, the applicability to the actual components is of concern. In this study, two benchmark problems on the fracture tests supposing actual components were provided to investigate the prediction ability of simulation methods containing parameter decisions. One was the circumferentially through-wall and surface cracked pipes subjected to monotonic bending, and the other was the circumferentially through-wall cracked pipes subjected to cyclic bending. Participants predicted the ductile crack propagation behavior by their own approaches, including finite element method (FEM) employed Gurson–Tvergaard–Needleman (GTN) yielding function with void ratio criterion, are FEM employed GTN yielding function, FEM with fracture strain or energy criterion modified by stress triaxiality, extended FEM with J or ΔJ criterion, FEM with stress triaxiality and plastic strain based ductile crack propagation using FEM, and elastic-plastic peridynamics. Both the deformation and the crack propagation behaviors for monotonic bending were well reproduced, while few participants reproduced those for cyclic bending. To reproduce pipe deformation and fracture behaviors, most of the groups needed parameters that were determined to reproduce pipe deformation and fracture behaviors in benchmark problems themselves and it is still difficult to reproduce them by using parameters only from basic materials tests.
- 7 December 2018
- Published by Springer Nature
- p. 183-205
Numerical Library Based on Hierarchical Domain Decomposition
- 7 December 2018
- Published by Springer Nature
- p. 183-205
Numerical Library Based on Hierarchical Domain Decomposition
- 22 May 2016
- Published by Springer Nature
- p. 83-94
High-Resolution Visualization Library for Exascale Supercomputer
- 22 May 2016
- Published by Springer Nature
- p. 83-94
High-Resolution Visualization Library for Exascale Supercomputer
- 1 August 2015
- Published by ASME International in Journal of Pressure Vessel Technology
- Vol. 137 (4) , 041401
Simulation of Stress Corrosion Cracking in In-Core Monitor Housing of Nuclear Power Plant
In the in-core monitor (ICM) housing of a reactor pressure vessel (RPV), residual stress has been widely reported to cause stress corrosion cracking (SCC) damage in the weld heat-affected zone. For this reason, it is important to evaluate the crack growth conservatively, and with high confidence to demonstrate fitness for service. This paper presents crack growth simulations in an ICM housing, which is welded at two different angles to the RPV. One weld angle is at the bottom of the RPV, and the welding area of the ICM housing is axisymmetric. The other angle is at the curved position of the RPV, and the weld area of the ICM housing is asymmetric. In these weld areas, crack growth behavior is estimated by superposed-finite element method (S-FEM), which allows generation of a global finite model and a detailed local mesh representing the crack independently. In the axisymmetric weld area, axial, slant and circumferential surface cracks are assumed at two locations where the residual stress fields are different from each other: one is isotropic and the other is circumferential. It is shown that crack growth behaviors are different under different residual stress fields. The results of S-FEM are compared with those of the influence function method (IFM), which assumes that an elliptical crack shape exists in a plate. It is shown that the IFM result is conservative compared to that of S-FEM. Next, an axial surface crack is assumed at the uphill, downhill, and midhill asymmetric weld areas. The midhill crack growth behavior is different from the uphill and downhill behaviors. Finally, two surface cracks are simulated in the asymmetric weld area and two initial crack arrangements are assumed. These results show the differences of the crack interaction and the crack growth process.
- 1 August 2015
- Published by ASME International in Journal of Pressure Vessel Technology
- Vol. 137 (4) , 041401
Simulation of Stress Corrosion Cracking in In-Core Monitor Housing of Nuclear Power Plant
In the in-core monitor (ICM) housing of a reactor pressure vessel (RPV), residual stress has been widely reported to cause stress corrosion cracking (SCC) damage in the weld heat-affected zone. For this reason, it is important to evaluate the crack growth conservatively, and with high confidence to demonstrate fitness for service. This paper presents crack growth simulations in an ICM housing, which is welded at two different angles to the RPV. One weld angle is at the bottom of the RPV, and the welding area of the ICM housing is axisymmetric. The other angle is at the curved position of the RPV, and the weld area of the ICM housing is asymmetric. In these weld areas, crack growth behavior is estimated by superposed-finite element method (S-FEM), which allows generation of a global finite model and a detailed local mesh representing the crack independently. In the axisymmetric weld area, axial, slant and circumferential surface cracks are assumed at two locations where the residual stress fields are different from each other: one is isotropic and the other is circumferential. It is shown that crack growth behaviors are different under different residual stress fields. The results of S-FEM are compared with those of the influence function method (IFM), which assumes that an elliptical crack shape exists in a plate. It is shown that the IFM result is conservative compared to that of S-FEM. Next, an axial surface crack is assumed at the uphill, downhill, and midhill asymmetric weld areas. The midhill crack growth behavior is different from the uphill and downhill behaviors. Finally, two surface cracks are simulated in the asymmetric weld area and two initial crack arrangements are assumed. These results show the differences of the crack interaction and the crack growth process.
- 1 January 2013
- Published by Japan Society of Mechanical Engineers in The Proceedings of The Computational Mechanics Conference
- Vol. 2013.26, _1313-1_
1313 Study of Ray-casting Visualization for Large-scale Analysis
- 1 January 2013
- Published by Japan Society of Mechanical Engineers in The Proceedings of The Computational Mechanics Conference
- Vol. 2013.26, _1313-1_
1313 Study of Ray-casting Visualization for Large-scale Analysis