Litcius/Paper detail

Differentiable turbulence: Closure as a partial differential equation constrained optimization

Varun Shankar, Dibyajyoti Chakraborty, Venkatasubramanian Viswanathan, Romit Maulik

2025Physical Review Fluids16 citationsDOI

Abstract

Improved turbulence closure models for large eddy simulations (LES) have the potential to impact a large variety of societal applications. This work introduces differentiable turbulence, where deep learning is embedded within a differentiable LES solver to enhance closure models given sparse observations of the true flow state. By leveraging physics-informed neural network architectures and solver-in-the-loop optimization, we put forth a technique that allows for the learning of novel closures without the use of high-fidelity numerical simulations - opening a pathway to the development and identification of LES closures in a multifidelity setting.

Topics & Concepts

Closure (psychology)TurbulencePartial differential equationDifferentiable functionFirst-order partial differential equationApplied mathematicsK-omega turbulence modelMathematicsK-epsilon turbulence modelMathematical analysisPhysicsMechanicsEconomicsMarket economyFluid Dynamics and Turbulent FlowsFluid Dynamics and Vibration AnalysisComputational Fluid Dynamics and Aerodynamics