Reducing Complex Physico-Chemical Model Systems

An important and fundamental challenge in science and engineering is the need to understand the system output variables (model predictions, measured observables) and their relation to the system¡¦s input variables (model inputs, specified experimental conditions). This project proposes to address this problem through the development of a family of numerical tools for implementing High Dimensional Model Representation (HDMR) to characterize the multi-parameter input-output (IO) behavior of complex physiochemical systems. These tools are intended to provide a robust suite of analysis programs to support both theoretical and experimental technology development research efforts. The starting point for this work is the computational HDMR algorithms and software which have been developed at Princeton University over the past several years. This Phase I and II project will: (a) build on that foundation by developing formulations for highly nonlinear systems, (b) apply HDMR to new applications, and (c) develop user friendly, commercial grade data analysis software for applications which are important to practicing scientists and engineers in a wide variety of scientific areas. The proposed Phase I research begins this work with the demonstration of HDMR techniques for thermal RDX decomposition.