Ensemble Gradient for Learning Turbulence Models from Indirect Observations
From MaRDI portal
Abstract: Training data-driven turbulence models with high fidelity Reynolds stress can be impractical and recently such models have been trained with velocity and pressure measurements. For gradient-based optimization, such as training deep learning models, this requires evaluating the sensitivities of the RANS equations. This paper explores the use of an ensemble approximation of the sensitivities of the RANS equations in training data-driven turbulence models with indirect observations. A deep neural network representing the turbulence model is trained using the network's gradients obtained by backpropagation and the ensemble approximation of the RANS sensitivities. Different ensemble approximations are explored and a method based on explicit projection onto the sample space is presented. As validation, the gradient approximations from the different methods are compared to that from the continuous adjoint equations. The ensemble approximation is then used to learn different turbulence models from velocity observations. In all cases, the learned model predicts improved velocities. However, it was observed that once the sensitivity of the velocity to the underlying model becomes small, the approximate nature of the ensemble gradient hinders further optimization of the underlying model. The benefits and limitations of the ensemble gradient approximation are discussed, in particular as compared to the adjoint equations.
Recommendations
- Ensemble Kalman method for learning turbulence models from indirect observation data
- Combining direct and indirect sparse data for learning generalizable turbulence models
- Ensemble averaging in turbulence modelling
- Forecasting turbulent modes with nonparametric diffusion models: learning from noisy data
- Learning nonlinear turbulent dynamics from partial observations via analytically solvable conditional statistics
- Learned turbulence modelling with differentiable fluid solvers: physics-based loss functions and optimisation horizons
- A perspective on machine learning methods in turbulence modeling
- Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
- Deep learning in turbulent convection networks
Cites work
- scientific article; zbMATH DE number 5060482 (Why is no real title available?)
- A continuous adjoint formulation for the computation of topological and surface sensitivities of ducted flows
- A more general effective-viscosity hypothesis
- Analysis of the ensemble Kalman filter for inverse problems
- Blackbox and derivative-free optimization: theory, algorithms and applications
- DAFI: An Open-Source Framework for Ensemble-Based Data Assimilation and Field Inversion
- Ensemble Kalman inversion: a derivative-free technique for machine learning tasks
- Ensemble Kalman methods for inverse problems
- Gradient-Based Aerodynamic Robust Optimization Using the Adjoint Method and Gaussian Processes
- Introduction to Derivative-Free Optimization
- Inverse Problem Theory and Methods for Model Parameter Estimation
- Multiple shooting shadowing for sensitivity analysis of chaotic dynamical systems
- Multivariate stochastic approximation using a simultaneous perturbation gradient approximation
- On turbulent secondary flows in pipes of noncircular cross-section
- Quantifying and reducing model-form uncertainties in Reynolds-averaged Navier-Stokes simulations: a data-driven, physics-informed Bayesian approach
- RANS turbulence model development using CFD-driven machine learning
- Reconstruction of unsteady viscous flows using data assimilation schemes
- Regularized ensemble Kalman methods for inverse problems
- Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
- Theoretical connections between optimization algorithms based on an approximate gradient
- Turbulence modeling in the age of data
- Uncertainty quantification of reservoir performance predictions using a stochastic optimization algorithm
Cited in
(7)- Optimal sensor placement for ensemble-based data assimilation using gradient-weighted class activation mapping
- Artificial neural network based response surface for data-driven dimensional analysis
- Large-eddy simulation-based shape optimization for mitigating turbulent wakes of a bluff body using the regularized ensemble Kalman method
- Combining direct and indirect sparse data for learning generalizable turbulence models
- Ensemble Kalman method for learning turbulence models from indirect observation data
- Variational assimilation of transonic shock-wave boundary-layer interaction using sparse pressure measurements
- A framework for learning symbolic turbulence models from indirect observation data via neural networks and feature importance analysis
This page was built for publication: Ensemble Gradient for Learning Turbulence Models from Indirect Observations
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5065144)
