Project Description Understanding Unsteady Bioflows through Simulation, Modeling, Visualization, Art, and Psychology

The purpose of computing is insight, not numbers – Hamming We propose to develop and evaluate computational modeling, simulation, visualization, and data-analysis tools and apply them to specific scientific areas involving time-varying flows near complex 3D boundary shapes. Tools will be developed in a highly multidisciplinary manner, incorporating insights and knowledge from the biological domains, engineering, applied mathematics, computer science, visual design, art education, and perceptual psychology. The research process will integrate all the disciplines through the following steps. Biologists will pose hypotheses about how the biological systems they study interact with fluids. Together with engineers, they will capture relevant 3D motion of the biological systems as well as 3D time-varying flow measurements from around the systems. Applied mathematicians will simulate 3D time-varying flows consistent with the captured motion. They will also create computational models of both the motion and the unsteady flows. Biologists will then work with computer scientists to use new visualization software to visualize the resulting complex motion and flow data to evaluate the original scientific hypotheses. These visualizations will be of the captured data, the simulated data, and computational models merging the two. The new visualization software will be developed with expert input from artists, designers, and perceptual psychologists. Computational modeling research will include both direct numerical simulation and modal analysis (proper orthogonal decomposition) to create parameterized flow solutions in which the parameters allow 'what if' questions to be posed interactively. With these and other parameterized solutions, we will be able to interpolate across relatively large gaps in experimental data, allowing more effective comparison with simulated results and permitting more thorough and effective 3D visual analysis. Simulation research will include developing methods for computing flows and flow-structure interactions as well as for propagating uncertainty through the simulation process. Without knowledge of uncertainty, we cannot know whether features we see are significant. Simulated uncertainty will also propagate to and be visualized in our visual-ization research. Visualization research is at a crossroad. There are many visualization methods that address relatively simple problems well, but complicated data that is time-varying, 3D, multi-valued, and uncertain are very difficult to explore with these tools. In a sense, today's methods comprise a set of visualization primitives. How to choose the right primitives, combine them effectively, identify and discover primitives that are missing, and then select the right colors, line weights, transfer functions, or other parameters for each primitive is a tremendous challenge. …