Reasoning about Fluid Motion I: Finding Structures

With the increasing role of high performance computing in attacking complex physical problems, there is an urgent need for the development of advanced computational technology to provide scientists with high-level assistance in the analysis, interpretation, and modeling of a massive amount of quantitative data. A critical area where this need is quite evident is the problem of turbulence . The overall research goal is to develop a computational environment to help scientists efficiently make observations, test hypotheses, and make conceptual models of turbulence data sets . This paper presents the intellectual challenges and progress of this project. My approach is based on two key ideas: (1) Representing fluid behavior as local interactions amongsemi-persistent flexible objects like vortices enables automatic high-level qualitative interpretation of turbulence data, and (2) Abstracting from the particular features of fluid dynamcal reasoning, I propose five core operations aggregation, classification, re-description, spatial inference, and configuration change as part of a general theory of imagistic reasoning. A new vortex-finding algorithm is also presented . "Feymann said . . . when Einstein stopped creating it was because he stopped thinking in concrete physical images and became a manipulator of equations . "