Data-Driven Aerodynamic Modeling Using the DLR SMARTy Toolbox

Bekemeyer, Philipp; Bertram, Anna; Chaves, Derrick A. Hines; Ribeiro, Mateus Dias; Garbo, Andrea; Kiener, Anna L.; Sabater, Christian; Stradtner, Mario; Wassing, Simon; Widhalm, Markus; Goertz, Stefan; Jaeckel, Florian; Hoppe, Robert; Hoffmann, Nils

doi:10.2514/6.2022-3899

Cited by 17 publications

(7 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Franz [23] uses radial basis functions as a surrogate model based on a thin plate spline (TPS) due to its good approximation capability. The linear, thin-plate-spline and the cubic method employed in this study are taken from the library SMARTy [37].…”

Section: Surrogate Model Of the Hollow Fibre Humidifiermentioning

confidence: 99%

Analysis of Surrogate Models for Vapour Transport and Distribution in a Hollow Fibre Membrane Humidifier

et al. 2023

Self Cite

View full text Add to dashboard Cite

To achieve high efficiency and low degradation of a polymer electrolyte fuel cell (PEMFC), it is necessary to maintain an appropriate level of humidification in the fuel cell membrane. Thus, membrane humidifiers are typically used in PEMFC systems. Parameter studies are important to evaluate membrane humidifiers under various operating conditions to reduce the amount of physical tests. However, simulative studies are computationally expensive when using detailed models. To reduce the computational cost, surrogate models are set up. In our study, a 3D computational fluid dynamics (CFD) model of a hollow fibre membrane humidifier is presented and validated using measurement data. Based on the results of the validated CFD model, a surrogate model of the humidifier is constructed using proper orthogonal decomposition (POD) in combination with different interpolation methods. To evaluate the surrogate models, their results are compared against reference solutions from the CFD model. Our results show that a Halton design combined with a thin-plate-spline interpolation results in the most accurate surrogate humidifier model. Its normalised mean absolute error for 18 test points when predicting the water mass fraction in the membrane humidifier is 0.58%. Furthermore, it is demonstrated that the solutions of the POD model can be used to initialise CFD calculations and thus accelerate the calculation of steady state CFD solutions.

show abstract

Section: Surrogate Model Of the Hollow Fibre Humidifiermentioning

confidence: 99%

Analysis of Surrogate Models for Vapour Transport and Distribution in a Hollow Fibre Membrane Humidifier

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…This tandem approach generates complementary datasets in terms of fidelity and cost, typically consisting of extensive data points. Multi-fidelity GPs are often used for mixing CFD simulations and wind tunnel experiments [7,12,[35][36][37], but they cannot be applied to large databases. Hence, from a theoretical point of view, a potential future work is to extend the sparse approximations discussed in this paper to multi-fidelity GPs.…”

Section: Discussionmentioning

confidence: 99%

A Python Toolbox for Data-Driven Aerodynamic Modeling Using Sparse Gaussian Processes

Valayer,

Bartoli,

Castaño-Aguirre

et al. 2024

Aerospace

View full text Add to dashboard Cite

In aerodynamics, characterizing the aerodynamic behavior of aircraft typically requires a large number of observation data points. Real experiments can generate thousands of data points with suitable accuracy, but they are time-consuming and resource-intensive. Consequently, conducting real experiments at new input configurations might be impractical. To address this challenge, data-driven surrogate models have emerged as a cost-effective and time-efficient alternative. They provide simplified mathematical representations that approximate the output of interest. Models based on Gaussian Processes (GPs) have gained popularity in aerodynamics due to their ability to provide accurate predictions and quantify uncertainty while maintaining tractable execution times. To handle large datasets, sparse approximations of GPs have been further investigated to reduce the computational complexity of exact inference. In this paper, we revisit and adapt two classic sparse methods for GPs to address the specific requirements frequently encountered in aerodynamic applications. We compare different strategies for choosing the inducing inputs, which significantly impact the complexity reduction. We formally integrate our implementations into the open-source Python toolbox SMT, enabling the use of sparse methods across the GP regression pipeline. We demonstrate the performance of our Sparse GP (SGP) developments in a comprehensive 1D analytic example as well as in a real wind tunnel application with thousands of training data points.

show abstract

“…DLR's SMARTy toolbox [44], which is a modular, object-oriented python package, is used as a framework for implementing the candidate function derived by GEP in TAU and extracting computational information .…”

Section: Methodsmentioning

confidence: 99%

Evolutionary Algorithm applied to Differential Reynolds Stress Model for Turbulent Boundary Layer subjected to an Adverse Pressure Gradient

Alaya

Grabe

Eisfeld

2022

AIAA AVIATION 2022 Forum

View full text Add to dashboard Cite

In this paper, an evolutionary algorithm is implemented for the purpose of performing symbolic regression in an attempt to improve Reynolds-Averaged-Navier-Stokes models predictions. In contrast to most machine learning algorithms, Gene Expression Programming generates a mathematical expression that can be directly interpreted and implemented into the Computational Fluid Dynamics solver. In this paper, the latter feature is exploited based on high-fidelity data to ascertain novel correlations for the pressure-strain correlation within a particular Differential Reynolds Stress Model, the Speziale-Sarkar-Gatski (SSG) model. The CFD-driven Gene Expression Programming is considered for the curved backward-facing step. Two models are obtained regarding the industrially relevant phenomenon of a turbulent boundary layer under adverse pressure gradient. The models are tested on a range of validation cases.

show abstract

Data-Driven Aerodynamic Modeling Using the DLR SMARTy Toolbox

Cited by 17 publications

References 69 publications

Analysis of Surrogate Models for Vapour Transport and Distribution in a Hollow Fibre Membrane Humidifier

Analysis of Surrogate Models for Vapour Transport and Distribution in a Hollow Fibre Membrane Humidifier

A Python Toolbox for Data-Driven Aerodynamic Modeling Using Sparse Gaussian Processes

Evolutionary Algorithm applied to Differential Reynolds Stress Model for Turbulent Boundary Layer subjected to an Adverse Pressure Gradient

Contact Info

Product

Resources

About