Computational Science and Engineering
Changing the Face of Scientific Inquiry
Rensselaer researchers are developing methods that let scientists and engineers create and use computer models of complex multiscale phenomena.
Such tools are increasingly necessary to solve many of the open biomedical and engineering problems in which the behaviors of interest are dictated by interactions at multiple length and time scales.
Understanding the human cardiovascular system, for example, requires prediction of the flow through the arteries and veins. The endothelial cells that line the arteries respond to the varied flow (pressure and shearing force). These changes, in turn, can be traced to genetic actions within the cells.
Biomedical researchers hoping to use a computer model to learn more about causes and potential cures of cardiovascular diseases, therefore, need a model that simulates the behavior and interaction of all these size levels over time. Similarly, materials scientists working with new nanophase materials need models that simulate macro-level behaviors which are dictated by properties that must be determined via atomic level calculations that are coupled to the macro-level fields.
Members of Rensselaer’s Scientific Computation Research Center (SCOREC) have built an international reputation for their creation of adaptive finite element modeling and other techniques that make it possible to solve such complex engineering problems.
Now they are working with collaborators across the campus and the country to extend these methods to complex multiscale models that track interactions at many lengths and time scales.
The models they are building will look, for example, at all of the interactions in a human knee: a macro-model of the engineering stresses on the joint, another model to look at the tissue, a third to consider what is happening within individual cells, and a fourth to simulate the genetic changes within the cells.
The goal is automated tools that have enough speed, cost-efficiency, user friendliness, and accuracy to make them useful for doctors, engineers, and scientists who are not themselves experts in computer modeling.