Research Digest | Search | MIT News | Comments | MIT
These brief summaries of MIT research are drawn from several
sources and are issued throughout the year. More information on any
of these stories can be obtained by
contacting
the MIT News Office. In some cases, photos may be
available for news organizations.
Piston Slap. Few
technologies have received more engineering attention than the
internal combustion engine. Yet engine designers continue to be
troubled by a phenomenon known as "piston slap." As a piston moves up
and down inside its cylinder it also shifts from side to side,
bumping first one side and then the other--a behavior that wastes
fuel, wears out engines and makes an annoying bang. A computer model
developed by MIT researchers can disentangle the factors that lead to
piston slap, helping engineers make design decisions that will reduce
its intensity. Given a description of the operating conditions and
design of an engine, the model can describe the pathway the piston
follows inside the cylinder, the force with which it hits the wall,
and even how its shape changes due to the impact. In parallel work,
the researchers have validated the model using an operating
experimental engine. The team was led by Dr. Victor Wong, a principal
research scientist in the
Energy
Laboratory and lecturer in the Department of Mechanical
Engineering. The work was funded by Nissan Motor Company.
Improving Wavelets. Today, complex information such as an image or a video signal is put into digital form, transmitted over communications channels, and reconstructed in close to its original form. The conventional technique for selecting the least information needed to reconstruct the original signal involves describing the signal in terms of cosine waves that theoretically continue forever. Enter a new technique based on a mathematical concept known as wavelets. Wavelets have a finite length, and for many applications they work better than cosines. While others develop new applications for wavelets, MIT mathematicians are improving the analytical tool itself. For example, they are using combinations of several different wavelets to represent the signal. The ideas are described by Professor Gilbert Strang of the Department of Mathematics in his new text for the MIT course "Wavelets and Filter Banks." The results of compressing and reconstructing the signal using the new ideas are promising.
Mapping a Diabetes Gene. Non-insulin-dependent diabetes mellitus (NIDDM) affects more than 100 million people worldwide. Screening more than 4,000 individuals from an isolated region in Finland, an international group of scientists has located a gene that may be involved in a significant fraction of adult-onset diabetes tied to low insulin secretion. The strategy used to find this gene has important implications for the future of complex disease research, as well as our understanding of the multiple causes of human diabetes. The scientists were led by Professor Eric Lander of the Whitehead Institute for Biomedical Research and the MIT Department of Biology, and Dr. Leif Groop of Lund University, Sweden. The work is reported in the September 1996 Nature Genetics. It was supported in part by the Whitehead Institute, the NIH, the Juselius Foundation, the Finnish Diabetes Research Foundation, Millennium Pharmaceuticals, Inc., the Albert Pahlsson Foundation, the Novo-Nordisk Foundation, the Medical Research Councils of Sweden, Canada and the UK, and the Medical Society of Finland.
Hypercholesterolemia. Familial hypercholesterolemia, a genetic disease characterized by high levels of cholesterol and early mortality, is caused by defects in the receptor for the low-density lipoprotein (LDL)--the bad cholesterol. Now, scientists from MIT, the Whitehead Institute for Biomedical Research, and Brigham and Women's Hospital have found that this occurs because mutations in the LDL receptor prevent the protein from folding into its normal shape. This in turn impedes the protein's ability to bind to bad cholesterol and remove it from the bloodstream, causing the hypercholesterolemia. All proteins are strings of amino acids that must fold into unique shapes to perform their functions. Knowing that a protein-folding defect is at the root of familial hyper-cholesterolemia may allow scientists to design better therapies. Professor Peter Kim of Whitehead and the MIT Department of Biology is lead author of a paper on the work in the September 1996 issue of Nature Structural Biology. The work was supported by the NIH's National Heart, Lung, and Blood Institute.
Groundwater Pollutants. In work that could lead to better ways of cleaning up certain pollutants from groundwater, MIT researchers are looking at the interactions between those pollutants and the soils and water they move through. Groundwater contaminants like carbon tetrachloride that are in the form of nonaqueous phase liquids (NAPLs) are extremely challenging to clean up, because the behavior of NAPLs in the subsurface is very complex. Researchers led by Professor Patricia Culligan-Hensley of the Department of Civil and Environmental Engineering are exploring that behavior via several different projects. In one, they are investigating how dense NAPLs migrate in bedrock systems. When such an NAPL enters the subsurface and falls under its own weight to an impermeable barrier such as bedrock, it will have enough force to actually push itself into fractures in the rock, where it becomes difficult to detect and to predict where it might move next. Using a special centrifuge to model the problem, the researchers are looking at how the NAPL moves into the fracture system under different gravitational forces. The work is sponsored by the EPA's Northeast Hazardous Substance Research Center.
Ocean Disposal of CO2? The ability to capture and dispose of carbon dioxide (CO2) emitted by electric power plants could make continued use of fossil fuels possible even if global warming proves a serious threat. Technology exists for capturing CO2, but disposing of it remains a problem. In DOE-supported work, MIT researchers have been examining one option for large-scale CO2 disposal: injection into the ocean. Using computer models, they have calculated how injected CO2 would change the acidity of the seawater around the injection point and how those changes would affect marine organisms nearby. Their analyses show that the quantity of CO2 injected is critical: twice as much CO2 generally kills more than twice as many organisms. However, certain methods of injection make the CO2 sufficiently dilute that the biological impacts are negligible. The analytical technique should also be of interest for assessing the biological impacts of other discharges and could provide a new method of setting emission standards. The researchers are led by Howard Herzog, a principal research engineer in the Energy Laboratory, and E. Eric Adams, a senior research engineer in the Department of Civil and Environmental Engineering. From September 9-11, 1996, the Energy Lab will host the Third International Conference on CO2 Removal.
Artificial Artist. The "artificial artist" is a computer program that creates sculpture in the style of Alexander Calder's abstract kinetic mobiles. Developed by Matthew Brand, a lecturer in media arts and sciences at the MIT Media Lab, the program does Calder one better by designing sculptures that are actually kinetic portraits. The program looks at an animal or face and identifies interesting parts, such as centers of muscular power, using computer vision and artificial intelligence techniques. Simulated force-fields then reshape these parts and link them into an attractive and well-engineered mobile, ensuring balanced counterweights and strong curve harmonies. Like their real-life counterparts, these virtual mobiles twist and turn on the computer screen, taking on a life of their own. Brand's work explores ways of making machines visually fluent so they can assist humans in creative activities such as design, assembly and play. Beyond sculpture, Brand's computing techniques provide new ways of solving complex spatial design problems such as making lightweight furniture that fits the contours and motions of our bodies. The work is funded by the Media Lab.