Abstract - IPAM

Abstract

Adjoint Direct Simulation Monte Carlo (DSMC) for Optimization and Control of Rarefied Flows

Yunan Yang

Cornell University

We derive adjoint equations for Direct Simulation Monte Carlo (DSMC) methods applied to the spatially inhomogeneous Boltzmann equation. The formulation incorporates spatial transport and a range of boundary conditions, including periodic boundaries, specular and thermal reflection, and prescribed inflow. The resulting adjoint system enables efficient gradient evaluation for quantities of interest defined through particle-based simulations. Numerical experiments validate the accuracy of the adjoint formulation and demonstrate its use in sensitivity analysis and gradient-based optimization. This work provides a foundation for integrating kinetic-scale particle methods into multi-fidelity optimization and control frameworks, with potential applications to rarefied plasma flows and edge physics in fusion devices.
No video available
Back to Workshop IV: Multi-Fidelity Methods to Enable Robust Optimization and Real-Time Control of Fusion Processes