College of Engineering -- University of Wisconsin-Madison
ord vehicle owners may be in for a smoother ride and increased fuel economy. Using an innovative observer system designed by recent PhD graduate Roy Davis (left), researchers can estimate instantaneous crankshaft torque and its ripple in an internal combustion engine. They use this estimate to form a disturbance input decoupling control signal, and command an electric motor mounted on the crankshaft to generate the opposite of the ripple torque and smooth the vehicle's vibration.
Initially Davis researched the best ways to model the engine system that produces the torque, then produced a simplified model that could be implemented in real time. He also researched and modified existing observer techniques to incorporate the effects of running the electric motor in conjunction with the combustion engine.
A Ford Fellow for two years, Davis conducted his research both at Ford Research Laboratory and with assistance from Professor Robert Lorenz. Currently Davis's system is being used in a hybrid vehicle powertrain, but in the future, the crankshaft-mounted electric motors could replace the alternator in conventional vehicles, or replace the torque sensors in the newer, more fuel-efficient direct-injected spark-ignited engines.
Supercomputer speeds small-engine study
A new $1.4 million high-performance supercomputer will speed up study at the college's Engine Research Center. Installed in spring 1999, the Origin 2000, a 32-processor computer that supplements a five-year-old Cray Research, Inc. computer with eight processors, will increase computational power by 16-fold and allow nearly 40 faculty, staff and students to complete larger research projects with faster turnover.
Jay Martin, a mechanical engineering professor and ERC director, says the Orion 2000 will enhance research for the state's small-engine manufacturing industry. Through computer simulations, ERC researchers can learn more about engine performance and design cleaner-running, more fuel-efficient engines for everything from lawn mowers to boats.
The ERC purchased the supercomputer with help from a $315,000 grant from the state Department of Commerce Technology Development Fund and a $340,000 in-kind gift from its manufacturer, Mountain View, Calif.-based Silicon Graphics Inc. (SGI).
New chips on the block
With $1 million in new funding over three years from the U.S. Defense Advanced Research Projects Agency and continued funding from SEMATECH, a 10-member consortium of semiconductor manufacturers, Professor Roxann L. Engelstad can research advanced computer chip-making technology in her recently established Computational Mechanics Laboratory.
Computer chips and integrated circuits are the core of every modern electronic system, and new technology is necessary because current optical lithography cannot create the smaller feature sizes of chips needed to deliver more power, store more memory and perform more tasks. Engelstad's goal is to produce features to 0.10 microns and below.
Her group is testing alternative technologies, including lithographies such as X-ray, 157 nm optical, projection electron beam, projection ion beam and extreme ultraviolet. They use state-of-the-art equipment, such as a supercomputer, to conduct simulations to determine the accuracy and stability of the sensitive, fragile thin-film masks that are templates for transferring patterns to computer chips. Engelstad says the alternative technologies aren't available commercially because of the tremendous cost to make them.
Models improve mixing in motion
Results of Professor Tim Osswald's polymer-mixing computer simulations are so promising that manufacturers could design the screw in a single-screw extruder to mix as efficiently as that in a twin-screw extruder--but at one-tenth the cost.
Because of increased understanding of the physics behind the boundary element method, a surface-modeling method that's less complicated than finite element analysis, the computer simulations allow researchers to model the mixing phenomena of polymer blends inside complex three-dimensional mixing devices like extruders easily without the cost of physically designing them. "Previously people designed mixers by trial and error," says Osswald, who is collaborating with PhD student Antoine Rios. Osswald likens the most efficient method of mixing continuous-flow polymers to the stretching and folding involved in making bread. "Current extruders deform material by shear," he says.
Osswald's simulations may help manufacturers optimize mixing equipment for maximal blending, controlled viscous heating, reduced residence time and maximum throughput. In collaboration with Rauwendaal Extrusion, Los Altos, Calif., a Madison-based company, The Madison Group, has designed the CRD mixing section for a single-screw extruder, which takes advantage of the dispersive mixing capability found in a twin-screw extruder; and the dispersive/distributive static mixer (DDSM). Patents are pending for both.
Copyright © 1999 University System Board of Regents