Control of Two-way Coupled Fluid Systems with Differentiable Solvers
This addresses the challenge of controlling complex nonlinear fluid-structure interactions, which is incremental as it builds on differentiable simulation methods.
The paper tackles the problem of controlling a rigid body in fluid using deep neural networks trained with a differentiable simulator, achieving reliable generalization to unseen inflow conditions and outperforming classical and learned alternatives.
We investigate the use of deep neural networks to control complex nonlinear dynamical systems, specifically the movement of a rigid body immersed in a fluid. We solve the Navier Stokes equations with two way coupling, which gives rise to nonlinear perturbations that make the control task very challenging. Neural networks are trained in an unsupervised way to act as controllers with desired characteristics through a process of learning from a differentiable simulator. Here we introduce a set of physically interpretable loss terms to let the networks learn robust and stable interactions. We demonstrate that controllers trained in a canonical setting with quiescent initial conditions reliably generalize to varied and challenging environments such as previously unseen inflow conditions and forcing, although they do not have any fluid information as input. Further, we show that controllers trained with our approach outperform a variety of classical and learned alternatives in terms of evaluation metrics and generalization capabilities.