Closed Loop Adaptive Refinement of Dynamical Models for Complex Chemical Reactions

Abstract

This research is concerned with the development of a systematic method for efficiently performing molecular dynamics (MD) simulations of complex chemical reactions and optimizing the underlying potential energy surfaces (PESs), ultimately using suitable laboratory data in a closed loop fashion. Two main objectives of the research are to (a) identify key parameters of each PES based on the global non-linear input-output Random Sampling High Dimensional Model Representation (RS-HDMR) mapping technique [1-7] and (b) use the RS-HDMR maps to efficiently capture the PES observable relationships [8-10]. The RS-HDMR analysis in turn provides essential information for subsequent full implementation of PES optimization within the proposed adaptive closed-loop learning algorithm in conjunction with laboratory feedback. In this project we have (1) formulated a fully equivalent operational model (FEOM) based on RS-HDMR, in place of the time-consuming Newton equations of motion for performing multi-dimensional MD simulations, and (2)performed detailed studies on intermolecular energy transfer for the model systems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 26, 2008
Accession Number
ADA499595

Entities

People

  • Herschel A. Rabitz

Organizations

  • Princeton University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Chemical Kinetics
  • Chemical Reactions
  • Department Of Defense
  • Dynamics
  • Energy
  • Energy Transfer
  • Engineering
  • Equations
  • Equations Of Motion
  • Molecular Dynamics
  • Potential Energy
  • Simulations
  • Statistical Analysis
  • Statistical Sampling
  • Students
  • Trajectories

Readers

  • Computational Modeling and Simulation
  • Quantum Chemistry
  • Robotics and Automation.