Presidential Young Investigator Award

点击翻译按钮获取中文摘要

This grant is to provide research support for Dr. Lyle Ungar under the National Science Foundation's Presidential Young Investigator Awards program. Dr. Ungar's research is in two areas: (1) the application of artificial intelligence techniques to automate reasoning about chemical plants and (2) the modelling of transport in systems with moving boundaries and phase transformations. In (1), the area of design and control, he is writing knowledge-based expert systems which use causal models and first principles to gain a "deep" understanding of plant behavior which is applicable to a broad class of processes and problems. The underlying reason for this work is that to be as useful and reliable as possible, expert systems should be able to work from descriptions of processes such as flow sheets, and not need extensive customization for each new process. As part of this project the PI is developing an expert multivariable controller which diagnoses process malfunctions and makes appropriate structural changes to the controller structure, control algorithms and control parameters in order to acheive as tight control of the process as possible. He is working on isolating as much plant-independent knowledge about equipment and control algorithms as possible and on formalizing notions of causality used by engineers i~ design and trouble shooting. In (2), the area of transport through multiphase media, he is using analytical and finite element based perturbation methods to characterize the effect of solidification conditions on melt/solid interface shape and solute segregation and, in particular, to map out the transitions between interfaces of different morphology. A detailed understanding of the solidification process is important for a variety of emerging techniques in electronic materials processing. Pulsed laser annealing is one of these techniques. Pulsed lasers can be used to rapidly melt thin layers (1 micron) of semiconductors and metals which resolidify at rates of up to several meters per second. At these velocities - some 10 times normal crystal growth rates, the melt/solid interface is no longer at equilibrium and dopant segregation is radically altered. Defects can be removed and nonequilibrium (metastable) alloys formed. However, nonuniform solute segregation often limits the usefulness of the alloys. He is incorporating results from atomic scale modelling of rapid solidification into macroscopic transport models, and studying the resulting growth morphologies.