Uncertainty Reduction in Aeroelastic Systems with Time-Domain Reduced-Order Models

Sarma, Rakesh; Dwight, Richard P.

doi:10.2514/1.j055527

Cited by 10 publications

(6 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We now introduce an additive random variable ε to account for noise in the measurements, specified as an unbiased normal distribution with a covariance matrix Σ, obtained from the experimental data (variance due to the moments reported for repeated runs or rotations of the turbine). Neglecting systematic modelling errors as in Sarma and Dwight, the statistical model can be written as follows:

{boldM}^{+} = bold \overset{bold^}{M} false(bolds, boldg false) + ε

ε \sim scriptN false(0, normalΣ false),

which gives the probability of observing M + given s and g , also known as the likelihood as follows:

double-struckP false({boldM}^{+} false| bolds, boldg false) : = {double-struckP}_{ε} false({boldM}^{+} - bold \overset{bold^}{M} false(bolds, boldg false) false) .

…”

Section: Uncertainty Reduction Of Experimental Systemmentioning

confidence: 99%

“…Neglecting systematic modelling errors as in Sarma and Dwight,37 the statistical model can be written as follows:…”

Section: Bayesian Framework For Uncertainty Reductionmentioning

confidence: 99%

“…In this section, we introduce a framework for construction of ROMs for predicting aeroelastic forces. The framework is based on an earlier study by the present authors applied to test case of the Goland wing; here, we discuss it concisely. It is based on building the ROM for prediciting the aerodynamics only, which is responsible for bulk of the computational time.…”

Section: Reduced Order Modelling Frameworkmentioning

confidence: 99%

“…The ROM is initially trained through aerodynamic data obtained from precribed forced motion to the structure. Two types of ROMs were developed in Sarma and Dwight: an autoregressive with exogenous (ARX) model and a linear parameter varying (LPV)‐ARX model. The ARX model is a time‐marching method that predicts the output of a system based on earlier observations and inputs, represented in the form of a recurrence relation.…”

Section: Reduced Order Modelling Frameworkmentioning

confidence: 99%

“…The framework is adopted from an earlier study by the present authors, applied to the case of aircraft wing. Initially, a high‐fidelity aeroelastic solver is developed with a modal structural solver and RANS‐based fluid solver.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Aeroelastic validation and Bayesian updating of a downwind wind turbine

2020

Self Cite

View full text Add to dashboard Cite

Summary Downwind wind turbine blades are subjected to tower wake forcing at every rotation, which can lead to structural fatigue. Accurate characterisation of the unsteady aeroelastic forces in the blade design phase requires detailed representation of the aerodynamics, leading to computationally expensive simulation codes, which lead to intractable uncertainty analysis and Bayesian updating. In this paper, a framework is developed to tackle this problem. Full, detailed aeroelastic model of an experimental wind turbine system based on 3‐D Reynolds‐averaged Navier‐Stokes is developed, considering all structural components including nacelle and tower. This model is validated against experimental measurements of rotating blades, and a detailed aeroelastic characterisation is presented. Aerodynamic forces from prescribed forced‐motion simulations are used to train a time‐domain autoregressive with exogenous input (ARX) model with a localised forcing term, which provides accurate and cheap aeroelastic forces. Employing ARX, prior uncertainties in the structural and rotational parameters of the wind turbine are introduced and propagated to obtain probabilistic estimates of the aeroelastic characteristics. Finally, the experimental validation data are used in a Bayesian framework to update the structural and rotational parameters of the system and thereby reduce uncertainty in the aeroelastic characteristics.

show abstract

{boldM}^{+} = bold \overset{bold^}{M} false(bolds, boldg false) + ε

ε \sim scriptN false(0, normalΣ false),

which gives the probability of observing M + given s and g , also known as the likelihood as follows:

double-struckP false({boldM}^{+} false| bolds, boldg false) : = {double-struckP}_{ε} false({boldM}^{+} - bold \overset{bold^}{M} false(bolds, boldg false) false) .

…”

Section: Uncertainty Reduction Of Experimental Systemmentioning

confidence: 99%

“…Neglecting systematic modelling errors as in Sarma and Dwight,37 the statistical model can be written as follows:…”

Section: Bayesian Framework For Uncertainty Reductionmentioning

confidence: 99%

Section: Reduced Order Modelling Frameworkmentioning

confidence: 99%

Section: Reduced Order Modelling Frameworkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Aeroelastic validation and Bayesian updating of a downwind wind turbine

2020

Self Cite

View full text Add to dashboard Cite

show abstract

Efficient multi-fidelity reduced-order modeling for nonlinear flutter prediction

Wang,

Song,

Peng

et al. 2024

Aerospace Science and Technology

View full text Add to dashboard Cite

Low-dimensional representation of intermittent geophysical turbulence with high-order statistics-informed neural networks (H-SiNN)

Foldes,

Camporeale,

Marino

2024

Physics of Fluids

View full text Add to dashboard Cite

We present a novel machine learning approach to reduce the dimensionality of state variables in stratified turbulent flows governed by the Navier–Stokes equations in the Boussinesq approximation. The aim of the new method is to perform an accurate reconstruction of the temperature and the three-dimensional velocity of geophysical turbulent flows developing non-homogeneities, starting from a low-dimensional representation in latent space, yet conserving important information about non-Gaussian structures captured by high-order moments of distributions. To achieve this goal, we modify the standard convolutional autoencoder (CAE) by implementing a customized loss function that enforces the accuracy of the reconstructed high-order statistical moments. We present results for compression coefficients up to 16, demonstrating how the proposed method is more efficient than a standard CAE in performing dimensionality reduction of simulations of stratified geophysical flows characterized by intermittent phenomena, as observed in the atmosphere and the oceans.

show abstract

Uncertainty Reduction in Aeroelastic Systems with Time-Domain Reduced-Order Models

Cited by 10 publications

References 34 publications

Aeroelastic validation and Bayesian updating of a downwind wind turbine

Aeroelastic validation and Bayesian updating of a downwind wind turbine

Efficient multi-fidelity reduced-order modeling for nonlinear flutter prediction

Low-dimensional representation of intermittent geophysical turbulence with high-order statistics-informed neural networks (H-SiNN)

Contact Info

Product

Resources

About