billmartin-160x225
734-764-5534
Methodologies: Data, Statistics and Stochastic Methods, High Performance Computing, Numerical Analysis and Methodologies

William R. Martin

Professor, Nuclear Engineering and Radiological Sciences

Affiliation(s):

Michigan Energy Institute

Martin’s research involves development of advanced methods for high fidelity analysis of nuclear reactors, with both Monte Carlo methods and deterministic methods. The methods utilize a high-dimensional phase space (7 independent variables) with large data structures that depend nonlinearly on the solution and are big enough to require domain decomposition. Hybrid techniques combining Monte Carlo and deterministic methods yield huge sparse matrices that require innovative storage and inversion algorithms.

Fission source eigenmodes for a 2D full-core, pressurized water reactor using Monte Carlo to estimate the entries of a 2500×2500 fission matrix, a theoretically full but practically sparse matrix of spatial transition rates.

Fission source eigenmodes for a 2D full-core, pressurized water reactor using Monte Carlo to estimate the entries of a 2500×2500 fission matrix, a theoretically full but practically sparse matrix of spatial transition rates.