Tim Hoogenraad

Tim Hoogenraad is a graduate student in mechanical engineering automating mesh design for large-eddy simulation (LES) through data-driven grid adaptation. Traditional LES relies heavily on expert intuition and trial-and-error to distribute grid resolution, restricting scalability and preventing broader adoption in complex flows. His approach replaces heuristic mesh design with a predictive pipeline that learns how local resolution affects global quantities of interest. Tim introduces a partitioned quantity-of-interest framework decomposing global quantities into localized contributions that retain correlation while providing spatial resolution for adaptation. As a MathWorks Fellow, Tim will construct a data-driven model predicting how local resolution changes affect partitioned error by training on paired direct numerical simulation (DNS) and LES datasets. In MATLAB, he will implement and validate this framework across canonical turbulent flows including isotropic turbulence and channel flows. Tim will then integrate the framework into production solvers such as CharLES and MURPHY to demonstrate solver-independent performance. His work could establish a generalizable adaptive LES pipeline that eliminates manual mesh tuning and extends high-fidelity simulation to complex flow regimes.