Brian Lockwood

My research involves sensitivity analysis for multiphysics simulations. Multiphysics is a branch of modeling which seeks to simulate physical systems in a comprehensive manner by including all the relevant physics affecting the performance of a system. Examples of multiphysics problems include helicopters, nuclear reactors, and hypersonic flows. For these types of problems, experimental data can be costly or dangerous to obtain and exhaustive computer simulations can be expensive both computationally and in real dollars. Additionally, these types of problems may include empirical correlations or modeling parameters which can drastically affect the results of the simulation. To address these problems, adjoint sensitivity analysis must be a vital part of any multiphysics simulation.

Adjoint sensitivity analysis is a powerful tool used to determine the sensitivity of objective functions to a large number of design variables. Example objective functions may include lift or drag in the case of aerodynamics, or thermal power in the case of nuclear reactors. In addition to providing valuable sensitivity data for design optimization, adjoint sensitivity can be used to estimate the error encountered during the solution process. This error information can then be used to drive the solution strategy, such as adapting the mesh, changing the time step or varying a parameter within the code.

My research will involve estimating the error encountered in coupling different types of physics, such as neutral particle transport, with computational fluid dynamics. With the error estimated, coupling strategies which minimize both the coupling error and computational time needed for numerical solutions will be designed. In addition to analyzing coupling errors, the uncertainty introduced into the solution by model parameters and empirical correlations will be estimated. By looking at the sensitivity of the solution to these parameters, the parameters can be optimized to fit experimental data or theoretical observations.

Lockwood, B.A. and Mavriplis, D.J., "Parameter Sensitivity Analysis for Hypersonic Viscous Flow using a Discrete Adjoint Approach," 48th AIAA Aerospace Sciences Meeting, Orlando, Fl, January 2010.

Ibarra L, Lockwood B, and de Oliveira CRE, "A One and Two Dimensional PCICE-Based Multiphysics Algorithm for Use in Reactor Core Simulations," Joint International Topical Meeting on Mathematics and Computation and Supercomputing in Nuclear Applications, Monterey, CA, April 2007.

Lockwood, B.A, A Two Dimensional Fluid Dynamics Solver for Use in Multiphysics Simulations of Gas Cooled Reactors, Master's thesis, Georgia Institute of Technology, 2007.