Entries with tag computer modeling.

Computer Simulations Aid in Additive Manufacturing

Lawrence Livermore National Laboratory researchers have developed a new approach for additive manufacturing that promises to be more efficient for making metal parts. Their selective laser melting process makes a 3D object in layers using a high-energy laser beam to fuse metal powder particles. Parts made using conventional selective laser melting typically have weaknesses that could lead to failures. Their method avoids these problems by using high-power machines and working with computer designs and simulations in which the problems can be worked out before making a part, allowing users to select the appropriate parameters – the laser power or distance between scan lines, for example, or powder thickness – to create high density parts consistently. The key with selective laser melting is computing the dimensions of the pool of liquid that is formed when the laser melts the metal powder particles. The pools should be sufficiently deep to melt through the powder into the substrate, which results in a high-density material and thus a high-density part. The Lawrence Livermore work is part of its Accelerated Certification of Additively Manufactured Metals Strategic Initiative, which is investigating how computer simulations can help efficiently determine additive-manufacturing process parameters. The researchers published their findings in the International Journal of Advanced Manufacturing Technology. (EurekAlert)(Lawrence Livermore National Laboratory)(The International Journal of Advanced Manufacturing Technology)

Engineers: Proteins Could Be Used in Circuits

Researchers have identified the electrical properties of proteins that could enable the substances to be used for switching, indicating they could be used in circuits or electronic devices. The scientists in the University of Salento’s Department of Engineering for Innovation were studying these properties, an emerging discipline which they call proteotronics, in the protein known as OR-17, a receptor protein in rats. Previous research had determined proteins could be used as switches, so they created a model to describe the protein’s characteristics that enabled them to identify various properties computationally. With the ability to accurately model and predict OR-17’s electronic behavior, including its impedance, the engineers say, they could use it as a component in various devices. Being able to reliably model the behavior of the protein means it can be used as a reliable component in systems such as a circuit, they say. They published their work—“Proteotronics: Electronic Devices Based On Proteins”—online at arXiv. (SlashDot)(The Physics arXiv Blog)(“Proteotronics: Electronic Devices Based On Proteins”)

New Techniques Perform 3D Modeling of the Human Heart

A team of University of Minnesota surgeons and biomedical engineers are using new technologies to create a digital library of human heart specimens and enable 3D computer modeling and mapping of hearts. This capability could let researchers see the structure and function of cardiac tissue, enabling them to better understand variations in the heart and how it changes in the presence of disease. It could also aid in the design of new cardiac devices. The University of Minnesota techniques use contrast-computed tomography, which uses dyes in the imaging process to allow the blood vessels and other structures to be better seen. The researchers are using human heart specimens from organ donors that have been found not to be usable for transplant. They published their work as a Journal of Visualized Experiments video article. (EurekAlert)(The Journal of Visualized Experiments  @ EurekAlert)
 

Researchers Create Human-Metabolism Model

An international consortium of researchers are using the power of supercomputing to produce what they say is the most comprehensive virtual reconstruction of human metabolism available, which could help identify and create individualized treatments for a wide range of diseases. Recon 2 is based on work by University of California, San Diego, researchers and could help study and treat of cancer and diabetes, as well as psychiatric and neurodegenerative disorders. Metabolism is the conversion of food sources into energy. Metabolic imbalances can play a role in disease. In turn, genetics, environment, and nutrition play a role in metabolism. High-powered computing has enabled scientists to construct huge interactive databases of biological data regarding metabolic pathways and analyze the information they contain. UC San Diego researchers built Recon 1 in 2007, basing it on more than 3,300 known biochemical reactions documented in more than 50 years of metabolic research. Recon 2, by contrast, includes information on about 7,400 reactions contributed by researchers worldwide. The researchers presented Recon 2 in a paper published online this month in the journal Nature Biotechnology. (EurekAlert)(University of California, San Diego @ EurekAlert)(Recon 2)

New Supercomputing Record Set

Scientists at the Stanford University-based Center for Turbulence Research, which is operated by both the school and NASA, set a new supercomputing record by using a million processing cores to model supersonic jet noise. They used the US Lawrence Livermore National Laboratory’s IBM Sequoia Blue Gene/Q system to solve the complex fluid-dynamics problem. Their work could not only help develop quieter aircraft engines, but also proves that million-core simulations are possible. (EurekAlert)(Stanford University) 

Computer Modeling Helps in Microrobotics Design

 

Georgia Institute of Technology researchers have used complex computational models to help design swimming microrobots. The engineers say their modeling shows what factors would be important in constructing a real system. The researchers modeled a robot that was 10 microns long, had two flaps, and was made from a gel that responds to chemical reactions or changes in temperature or in magnetic or electric fields, or temperature changes. These changes would enable the body of the swimming robot to expand and contract, which would also make the hydrogel-based flaps move and help propel it. The researchers said microswimming devices could travel a few micrometers per second and be useful for targeted drug delivery within the human body, to help with small-scale assembly tasks, and in microfluidic chips for lab-on-a-chip applications. They published their work in the journal Soft Matter. (Science Daily)(Georgia Institute of Technology)(Soft Matter)

Supercomputer Calculates Possible Isotope Combinations

Oak Ridge National Laboratory and University of Tennessee researchers used the Jaguar supercomputer to calculate the number of isotopes that the laws of physics allow. They used six nuclear-interaction models and found about 7,000 possible combinations. Of these 7,000, scientists have observed or produced about 3,000. The other combinations are created in massive stars or in violent stellar explosions, say the researchers. In their calculations, they quantified the so-called drip lines—the maximum number of neutrons and protons the laws of physics allow in a nucleus—that determine nuclear existence. The drip lines become uncertain among heavier elements. The calculations for each possible nuclei require about two hours of supercomputer processing time and include about 250,000 possible nuclear configurations. The researchers say they could not have done this work two or three years ago because they wouldn’t have had access to such a powerful supercomputer. The researchers say their work will create numerous scientific insights and someday could yield benefits such as cancer treatments that irradiate malignant cells without damaging healthy ones. They published their research in the journal Nature. (EurekaAlert)(Oak Ridge National Laboratory)
 

Modeling Helps Understand Human Immune System

Virginia Polytechnic Institute and State University researchers have created a computer model that helps scientists better understand the human immune system. They specifically looked at macrophage-cell responses. These cells typically identify and respond to pathogens that cause infection. However, a lengthy immune response can cause inflammatory diseases, such as multiple sclerosis or type 2 diabetes. To understand this phenomenon, the Virginia Tech scientists say, they studied how sequential pathogen attacks alter or reprogram macrophage immune responses. The scientists utilized the Metropolis algorithm—a statistical approach commonly used in physics and chemistry—in their model to determine the molecular mechanisms involved in such activities. The researchers, who published their work in PLoS Computational Biology, say more studies and testing are needed. (PhysOrg)(Virginia Tech)(PLoS Computational Biology)

Researchers Detect Migration Trends from E-Mail

New research finds that e-mail is a good source of human-migration data for almost every country of the world. Official records from different nations are typically inconsistent and often are missing key information. Also, migrants frequently never register with government agencies. Max Planck Institute for Demographic Research scientist Emilio Zagheni created a migration database by compiling the global flow of millions of e-mails. He and Ingmar Weber, a researcher with Yahoo! Research, traced 43 million anonymized e-mails based on their IP addresses between September 2009 and June 2011. When a person started sending e-mail from a new location permanently, the researchers assumed it was due to a residence change. This approach could track migrants within or between countries. The researchers adjusted their model to compensate for underrepresented populations such as the elderly, who tend to use e-mail less frequently. Zagheni said his approach could mine data to address many types of population dynamics, including mobility patterns following natural disasters. “This research has the most potential in developing countries,” he said, “where the Internet spreads much faster than [migration] registration programs develop.” The scientists published their findings in the ACM Web Science 2012 conference proceedings. (PhysOrg)(Max Planck Society)

US Researchers Refine Computer Modeling for Nuclear-Weapons Testing


Purdue University and Lawrence Livermore National Laboratory scientists are using supercomputers to refine the models that the US government uses to conduct critical nuclear-weapons performance simulations. The researchers are developing models that are accurate to the molecular level. These models would improve the accuracy of the highly complex simulations that are critical for the US to verify the performance of nuclear weapons without testing them via explosions. High precision levels are critical because many of the calculations are executed in parallel on separate machines in large clusters, which could turn small errors into large ones. The researchers also developed an automated tool that detects an error in the calculation process after it occurs so that they can correct it. They will present their work at the IEEE/IFIP International Conference on Dependable Systems and Networks in Boston. (EurekAlert)(Purdue University)

Showing 10 results.