Simulation-based engineering and science

Sharon C. Glotzer (Chair)

R. D. Shelton, President

Michael DeHaemer, Executive Vice President

Geoffrey M. Holdridge, Vice President for Government Services

David Nelson, Vice President for Development

Ben Benokraitis, Assistant Vice President

L. Pearson, Project Manager

Patricia M.H. Johnson, Director of Publications

Halyna Paikoush, Event Planner and Administrative Support

Table of Contents


Appendix B. Site Reports—Asia

Site: Central Research Institute of Electric Power Industry (CRIEPI)

Dr. Tetsuo Matsumura, Deputy Director, Materials Science Research Laboratory

Mr. J. Tsutsui, Senior Research Scientist, Environmental Science Research Laboratory

Dr. T. Iwatsubo, Senior Research Scientist, Energy Engineering Research Laboratory

Dr. H. Kaieda, Senior Research Geophysicist, Civil Engineering Research Laboratory

Dr. N. Soneda, Senior Research Scientist, Materials Science Research Laboratory

Dr. K. Nakashima, Research Scientist, Materials Science Research Laboratory

Dr. N. Hashimoto, Research Scientist, Energy Engineering Research Laboratory

BACKGROUND

The Central Research Institute of Electric Power Industry (CRIEPI) is a Japanese nonprofit corporation founded in 1951 with a broad mission of “solving global environmental problems while ensuring energy security.” CRIEPI has an annual budget of ¥36.8 billion,¹ which comes mostly from the Japanese power industry; the remainder comes from the Japanese government. CRIEPI has 786 employees of whom 656 are classified as researchers. The institute is organized into one center, the Socio-economic Research Center, and 7 research laboratories: Systems Engineering, Nuclear Technology, Civil Engineering, Environmental Science, Electric Power, Energy Engineering, and Materials Science Laboratories. CRIEPI’s permanent research staff members have formal education in electrical, civil, mechanical, chemical, and nuclear engineering as well as in biological, environmental, information, and social sciences. They publish on the order of 1500 papers per year and file on the order of 300 patent applications per year.

The largest simulation efforts at CRIEPI are focused around materials science and climate/weather simulations. The materials science simulations focus on the multiscale modeling of irradiation embrittlement processes, light water reactor pressure vessel steels, and the stability and growth of oxide films on silicon carbide The multiscale aspects of the effort include molecular dynamics simulations of irradiation cascades, copper precipitation in iron, and dislocation irradiation defect reactions; kinetic Monte Carlo simulations of microstructural coarsening during irradiation; and dislocation dynamics simulations of irradiated strength change. The multiscale modeling framework relies on passing relevant information from small length scale simulations to larger length scale simulations, with the final result being a continuum rate model of irradiation damage evolution and corresponding mechanical property change. The oxide stability simulations focus on performing first principles molecular dynamics simulation of interfaces SiO₂ and SiC with defects and observing the chemical reactions of oxygen atoms across the interface. Such interface calculations are at the forefront of computational chemistry and required the use of the Earth Simulator to retain stable interfaces and be able to introduce stable point defects at those interfaces.

The CRIEPI researchers, in collaboration with NCAR (National Center for Atmospheric Research) and Earth Simulator staff, led an international team in performing unprecedented global climate simulations on the Earth Simulator that has influenced international policy on global climate change. The effort was sponsored by MEXT, and the porting of the CCSM 3.0 codes to the vector architecture of the Earth Simulator required approximately 10 person-years worth of effort. The result of the effort was the ability to perform multiple 400-year predictions of future temperature and sea level changes with different CO₂ emissions profiles. The simulations required about six months of wall clock time on the Earth Simulator for four different scenarios to be completed. A challenging high-resolution (a grid spacing of 10 km) ocean simulation was also conducted on the Earth Simulator, and eddy-resolved ocean currents were successfully simulated. Since then, the team at CRIEPI has begun performing regional climate simulations of East Asia to assess the regional impacts of global warming, high-resolution weather forecasting simulations, and global models of ocean currents. Several highlights included simulations of a typhoon making landfall in Japan, and predictions of local flooding in areas surrounding Japan’s energy infrastructure that required higher resolution than the Japanese Meteorological Agency typically performs.

CRIEPI researchers provide simulation analysis and tools to the Japanese electric power companies in other areas as well. Examples include CRIEPI’s Power system Analysis Tool (CPAT) for electric grid stability, CFD analyses of combustion in coal-fired power plants, CFD analysis of transmission line vibration, and seismic analysis of oil tanks. For the most part, CRIEPI employs commercial software in its routine simulations and outsources code development to tailor the commercial tools for problems of interest. However, the simulation tools for the high-end materials science and climate simulations are developed in-house or with a small set of international collaborators. CRIEPI management considers the codes developed in-house to be proprietary; after establishing the proprietary rights the institution shares its codes with the scientific community.

The researchers at CRIEPI working in the materials science, combustion, and climate areas are proposing ambitious simulation projects. In the nuclear materials arena, researchers are interested in simulating the process of stress-corrosion cracking of pressure vessel steels and cladding materials using a multiscale modeling framework. The challenging aspect of this endeavor is that it adds the complexity of surface chemistry and interfacial defects to simulations of mechanical behavior in irradiation environments. In the combustion arena, researchers are interested in predicting the combustion of new fuels and helping to design next-generation power plants. In the climate arena, researchers are interested in increasing the spatial resolution of the ocean current simulations to 0.1 degree such that oceanic eddies could be resolved, in adding complexity and explicit degrees of freedom in their climate models, and in developing a full “Earth system model” that simultaneously models the evolution of the atmosphere, ocean currents, and land masses.

An equally impressive experimental program was presented in the areas of nuclear materials and combustion science. CRIEPI has invested in state-of-the-art experimental hardware that is able to probe material microstructure at the same level as its kinetic Monte Carlo simulations of irradiation damage evolution. Its experimental and simulations efforts are well coordinated, and there appears to be strong feedback between the two. Validation of the kinetic Monte Carlo simulations allows CRIEPI researchers to make credible predictions of material behavior at engineering length scales with knowledge of the proper mechanisms of microstructural evolution. In their combustion experiments, CRIEPI researchers have built small test furnaces to validate the results and CFD simulations, and it appears that the simulators are now driving the experimentalists to make finer measurements of processes than they have had in the past. SBES at CRIEPI appears to have changed the manner in which experimental data is assessed.

In its hiring practices, CRIEPI tends to hire domain specialists when seeking new people for the institution’s modeling and simulation activities and does not exclusively seek out candidates with experience in computer science disciplines. The research staff preferred to hire domain specialists and train them to perform computer simulations rather than hire computer scientists and train them in the technical disciplines that require simulation support. There was an opinion shared by the staff that the universities were properly preparing students to perform the sophisticated cross-cutting research needed at industrial institutions.

CONCLUSIONS

CREIPI has an impressive array of simulation and experimental capabilities to address the most challenging problems facing the Japanese power industry. Its high-end simulations capabilities in materials science and global climate change are on par with the premier simulation institutions around the world. The institution’s mission focus enables its researchers to build strongly coupled simulation and experimental campaigns enabling them to validate their simulation results and build confidence in predictions. The close coordination of simulation and experiment should serve as an example to the rest of the world..

Dr. Taizo Sasaki, First Principles Simulation Group II Leader, CMSC, NIMS

Dr. Tsuyoshi Miyazaki, Researcher, CMSC, NIMS

Dr. Masato Shimono, Researcher, CMSC, NIMS

Dr. Toshiyuki Koyama, Researcher, CMSC, NIMS

Dr. Shigeru Okamura, International Affairs Office, NIMS

The National Institute for Materials Science was formed in 2001 by combining the National Research Institute for Metals (founded in 1956) and the National Institute for Research in Inorganic Materials (founded in 1966). The NIMS mission is (1) fundamental research and generic/infrastructural technology R&D; (2) popularization of research results and promotion of related activity; (3) common use of NIMS facilities and equipment; and (4) training of researchers and engineers and improvement of the quality of these human resources. NIMS is somewhat similar in scope to the Materials Science and Engineering Laboratory at the National Institute of Standards and Technology in the US. Their research is of the highest quality and impact. The have 124 MOUs with countries around the world, and 12 sister institutes in 7 countries. They participate in 10 graduate schools internationally, and collaborate extensively with industry. NIMS ranked 6th in the world in # of citations in materials. Six major research fields for NIMS from 2006-2010 are:

• 
  Nanotechnology-driven advanced materials research
  

• 
  Nanotech common key technologies
  
• 
  Synthesis and control of novel nanomaterials
  
• 
  Nanotech driven materials research for information technology
  
• 
  Nanotech driven materials research for biotechnology
  

• 
  Advanced materials research for social needs
  

• 
  Materials research for the environment and energy
  
• 
  Materials research for reliability and safety
  

The director of the CMSC is Dr. Takahisa Ohno. The Center is comprised of four groups, and totals roughly 20 permanent research staff and 20 postdocs and graduate students:

• 
  First Principles Simulation Group I – One group leader, six researchers and engineers, and 12 research fellows, and two administrative staff. Dr. Ohno is also the group leader of the First Principles Simulation Group I.
  
• 
  First Principles Simulation Group II - – One group leader, five researchers and engineers, four research fellows, and one office assistant.
  
• 
  Strong Coupling Modeling Group – One group leader, three researchers and engineers, three research fellows, and one office assistant.
  
• 
  Particle Simulation and Thermodynamics Group – One group leader, six researchers and engineers, four research fellows, and one office assistant.
  

The visiting panel heard an overview of NIMS and CMSC research from the Dr. Kitagawa, VP of NIMS, and Dr. Sasaki, Group Leader in CMSC.

SBE&S methods within the Center span multiple scales. First principles simulations include order-N DFT simulations, hybrid methods, and TDDFT. Nanoscale functional analysis includes electron transfer, spin transport, and diffusion quantum Monte Carlo. Strong coupling modeling treats quantum transport, many-body effects, and thermal fluctuations. Multiscale modeling and simulation combines molecular dynamics, Monte Carlo, and phase-field methods. The phase field work is particularly highly regarded within NIMS. Number 2 of the top 3 materials infrastructure accomplishments at NIMs is the structural materials database (see Software below).

The panel was impressed with the quantity and quality of the SBE&S research we heard about. Overall, the computational research is highly sophisticated and on par with the best work in the US and Europe. Current projects within each of the four groups of the CMSC include:

1. 
   First-principles simulation group I:
   

(b)Investigation of Electron Transport Properties of Nanostructures

(c)First Principles Simulation of Redox Reactions

(d)Control of the States of Materials

2. 
   First-principles simulation group II:
   

(b)Prediction of Properties of Materials

(c)Response of Materials (including magnetic, superconducting and optical materials)

3. 
   Strong-coupling modeling group
   

(e)Investigation on Phase Transition and Dynamics of Vortices in Superconductors

(f)Research on Quantum Transport Phenomena

4. 
   Particle simulation and thermodynamics group
   

(h)Reaction Diffusion Behavior between Ni-Al-Ir Alloy

(i)Research on Nanostructure of Materials by MD Simulations

(j)Investigation of Dynamics in Phase Transitions of Superconductors

The main supercomputer is the Hitachi SR11000/62 with 992 processors and 2 Tb storage. A new major acquisition from a Japanese computer vendor is planned for 2008. They are not currently addressing multicore computing future architectures.

Many of the codes used by the CMSC are developed in house. The CMSC has developed several impressive SBE&S software platforms now used by others. These include:

• 
  A thermodynamics database and simulation software developed in collaboration with AIST, Tohoku Univ., Kyushu Inst. Tech., and InterScience Ltd. The software company name is Materials Design Tech. CO. LTD, a venture company from NIMS.
  
• 
  MatEX: A Materials Design Platform (http://matex.nims.go.jp/). Demonstration programs for the phase-field modeling of microstructure evolution in engineering materials on the web.
  
• 
  Nanoscale Device Simulation Software (http://www.rss21.iis.u-tokyo.ac.jp). They offer services to predict and design the materials properties and functions for the next-generation semiconductor nano-devices on the web.
  

During the talks there was good opportunity for discussion of both scientific issues directly related to the content of the talks, and also more general issues surrounding simulation-based engineering and science. Some of the additional issues that arose included:

• 
  Education and HR in SBE&S: Potential postdocs and research staff with appropriate training in the development of algorithms and programs are hard to recruit. The main cause of this issue is that too many graduate students are being trained primarily to run existing codes to solve applied problems rather than learning the skills necessary to create a new application. NIMS recently created the International Center for Young Scientists (ICYS) to become a more attractive workplace for foreign young scientists. The concept is based upon (1) international: English as working language; (2) independent: autonomous research; (3) interdisciplinary: a fusion of different cultures and fields; and (4) innovative: strategic research. Benefits to participants in this program include a high salary, research grant support (5 M yen/year), and ideal research environment with cutting edge facilities and single occupancy cubicles. In five years, CMSC hired 10 simulators. Most of their backgrounds are in materials science, some physics. Most know how to run codes, but not how to innovate and develop new codes.
  
• 
  Appreciation of SBE&S: SBE&S activities at NIMS started 15 years ago, and CMSC started in 2001 with support from the Director-General. With new emphasis on nanoscience, experimentalists are becoming more interested in collaborating with simulators. Experimentalists now seek out input from simulators, unlike a several years ago. Simulation now having more and more impact on real problems.
  
• 
  Interaction with industry: No current interactions with industry.