Gregory Duthé

Last Name

Duthé

First Name

Gregory

ORCID

0000-0002-0895-6766

Organisational unit

03890 - Chatzi, Eleni / Chatzi, Eleni

Show all metadata

Search Results

Publications 1 - 10 of 23

Flexible multi-fidelity framework for load estimation of wind farms through graph neural networks and transfer learning
Duthé, Gregory; de N Santos, Francisco; Abdallah, Imad; et al. (2024)
Data-Centric Engineering
With global wind energy capacity ramping up, accurately predicting damage equivalent loads (DELs) and fatigue across wind turbine populations is critical, not only for ensuring the longevity of existing wind farms but also for the design of new farms. However, the estimation of such quantities of interests is hampered by the inherent complexity in modeling critical underlying processes, such as the aerodynamic wake interactions between turbines that increase mechanical stress and reduce useful lifetime. While high-fidelity computational fluid dynamics and aeroelastic models can capture these effects, their computational requirements limits real-world usage. Recently, fast machine learning-based surrogates which emulate more complex simulations have emerged as a promising solution. Yet, most surrogates are task-specific and lack flexibility for varying turbine layouts and types. This study explores the use of graph neural networks (GNNs) to create a robust, generalizable flow and DEL prediction platform. By conceptualizing wind turbine populations as graphs, GNNs effectively capture farm layout-dependent relational data, allowing extrapolation to novel configurations. We train a GNN surrogate on a large database of PyWake simulations of random wind farm layouts to learn basic wake physics, then fine-Tune the model on limited data for a specific unseen layout simulated in HAWC2Farm for accurate adapted predictions. This transfer learning approach circumvents data scarcity limitations and leverages fundamental physics knowledge from the source low-resolution data. The proposed platform aims to match simulator accuracy, while enabling efficient adaptation to new higher-fidelity domains, providing a flexible blueprint for wake load forecasting across varying farm configurations.
Graph Neural Networks for Aerodynamic Flow Reconstruction from Sparse Sensing
Duthé, Gregory; Abdallah, Imad; Barber, Sarah; et al. (2023)
arXiv
Sensing the fluid flow around an arbitrary geometry entails extrapolating from the physical quantities perceived at its surface in order to reconstruct the features of the surrounding fluid. This is a challenging inverse problem, yet one that if solved could have a significant impact on many engineering applications. The exploitation of such an inverse logic has gained interest in recent years with the advent of widely available cheap but capable MEMS-based sensors. When combined with novel data-driven methods, these sensors may allow for flow reconstruction around immersed structures, benefiting applications such as unmanned airborne/underwater vehicle path planning or control and structural health monitoring of wind turbine blades. In this work, we train deep reversible Graph Neural Networks (GNNs) to perform flow sensing (flow reconstruction) around two-dimensional aerodynamic shapes: airfoils. Motivated by recent work, which has shown that GNNs can be powerful alternatives to mesh-based forward physics simulators, we implement a Message-Passing Neural Network to simultaneously reconstruct both the pressure and velocity fields surrounding simulated airfoils based on their surface pressure distributions, whilst additionally gathering useful farfield properties in the form of context vectors. We generate a unique dataset of Computational Fluid Dynamics simulations by simulating random, yet meaningful combinations of input boundary conditions and airfoil shapes. We show that despite the challenges associated with reconstructing the flow around arbitrary airfoil geometries in high Reynolds turbulent inflow conditions, our framework is able to generalize well to unseen cases.
Identifying evolving leading edge erosion by tracking clusters of lift coefficients
Abdallah, Imad; Duthé, Gregory; Barber, Sarah; et al. (2022)
Journal of Physics: Conference Series
This work proposes an approach to identify Leading Edge Erosion (LEE) of a wind turbine blade by tracking evolving and emerging clusters of lift coefficients CL time-series signals under uncertain inflow conditions. Most diagnostic techniques today rely on direct visual inspection, image processing, and statistical analysis, e.g. data mining or regression on SCADA output signals. We claim that probabilistic multivariate spatio-temporal techniques could play an eminent role in the diagnostics of LEE specifically leveraging CL time-series signals form multiple sections along the span of the blade. The proposed method extracts clusters' features based on Variational Bayesian Gaussian Mixture Models (VBGMM) and tracks their spatial and temporal changes, as well as interpret the evolution of the clusters through prior physics-based assumptions. The parameters of the VBGMM are the mean, the eigenvalues and eigenvectors of the covariance matrix, and the angle of orientation of the eigenvectors. We show that the distribution of the CL data may not show statistically separable clusters, however, the parameters of the VBGMM clusters fitted to the CL data, allows to discriminate moving clusters primarily due to varying inflow and operating conditions, versus emerging clusters primarily due to evolving severity of the blade LEE.
Multivariate prediction on wake-affected wind turbines using graph neural networks
de Nolasco Santos, Francisco; Duthé, Gregory; Abdallah, Imad; et al. (2024)
Journal of Physics: Conference Series ~ XII International Conference on Structural Dynamics
Modern wind turbines are large and slender dynamical structures with a fatigue loading profile of complex nature. The guarantee of their structural integrity is paramount for materializing cost efficient and more reliable wind energy. The measurement of the global dynamic response and loads of wind turbines is fundamental for achieving this goal. However, an industry-wide, cost-effective direct sensing framework is yet to arise. Moreover, deploying physical sensors and measurement systems on every structural component of interest of a wind turbine induces prohibitive costs in deployment, maintenance and data management. Considering that direct fluid-structure interaction simulations on a farm level are not computationally feasible, the preferred path for structural response estimation on wind farms has been surrogate modelling. Within this landscape, new model architectures have risen in recent years which are able to take into account graph structured data (i.e. non-euclidean data). Wind turbines positioned in a farm, where there is a layout- and topology-dependant interplay of aerodynamic wake affecting the loading profile and power production, lend themselves perfectly to this paradigm. Thus, in this contribution, we introduce the use of graph neural networks (GNN) for layout agnostic saptio-temporal joint modelling of fatigue loads effects, rotor-averaged wind speed and power production on individual turbines of wind farms. To this end, we generate stochastic dependent samples of inflow conditions for wind speed, wind direction, wind shear and nacelle yaw angles. Additionally, wind farm layouts are randomly generated based on different geometric shapes (rectangle, triangle, ellipse and sparse circles) with random parametrization (varying orientations, length/width ratio) for different numbers of turbines and minimal distance (based on the rotor diameter). Both the arbitrary layouts and the random inflow conditions are used as inputs for PyWake, a wind farm simulation tool capable of calculating wind farm flow fields, power and fatigue loads. In our analysis, we develop and compare the performance of the GENeralized Aggregation Networks (GEN), the Graph Attention Networks (GAT) and the Graph Isomorphism Network with Edges (GINE) in their accuracy and ability to generalize their joint predictions for unseen layouts, uncertain inflow conditions and fatigue load estimation on the blade root, tower top and tower base of any wind turbine in the farm. Our results indicate that the GEN layer yields the best performance, followed by GINE, while the GAT layer under-performs and is unable to differentiate between different wake conditions. We further observe that the GAT layer causes a latent space collapse, due to the coupled effect of the manner in which we initialise node features and the way in which its messages are computed.
Development of a wireless, non-intrusive, MEMS-based pressure and acoustic measurement system for large-scale operating wind turbine blades
Barber, Sarah; Deparday, Julien; Marykovskiy, Yuriy; et al. (2022)
Wind Energy Science
As the wind energy industry is maturing and wind turbines are becoming larger, there is an increasing need for cost-effective monitoring and data analysis solutions to understand the complex aerodynamic and acoustic behaviour of the flexible blades. Published measurements on operating rotor blades in real conditions are very scarce due to the complexity of the installation and use of measurement systems. However, recent developments in electronics, wireless communication and MEMS (micro-electromechanical systems) sensors are making it possible to acquire data in a cost-effective and energy-efficient way. In this work, therefore, a costeffective MEMS-based aerodynamic and acoustic wireless measurement system that is thin, non-intrusive, easy to install, low power and self-sustaining is designed and tested in a wind tunnel. The measurement system does not require an electrical connection to the wind turbine and can be mounted and removed without damaging the blade.The results show that the system is capable of delivering relevant results continuously, although work needs to be done on calibrating and correcting the pressure signals as well as on refining the concept for the attachment sleeve for weather protection in the field. Finally, two methods for using the measurements to provide added value to the wind energy industry are developed and demonstrated: (1) inferring the local angle of attack via stagnation point detection using differential pressure sensors near the leading edge and (2) detecting and classifying leading edge erosion using instantaneous snapshots of the measured pressure fields. Ongoing work involves field tests on a 6 kW operating wind turbine in Switzerland.
Local flow and loads estimation on wake-affected wind turbines using graph neural networks and PyWake
Duthé, Gregory; de Nolasco Santos, Francisco; Abdallah, Imad; et al. (2023)
Journal of Physics: Conference Series
The estimation of the flow conditions throughout a modern wind farm is a relevant albeit complex problem involving several different modelling scales. In this context, recent years have seen the rise of the demand - normally industry-driven - of quick, computationally inexpensive, wind farm flow field simulators upon which further calculations might be undertaken (e.g. power output, fatigue loads, etc.). Naturally, the trustworthiness of such models hinges upon their ability to accurately estimate wakes, under the evident trade-off between computational time and accuracy. One such simulator that has been recently gaining traction within the wind community is PyWake, a Python-based flow solver, with several possible add-ons and modularity options. Based on runs from PyWake, this contribution studies the applicability of graph neural networks (GNN) as a simultaneous wind turbine fleet, wind inflow and loads surrogate. As graphs are discrete mathematical sets of dependent objects, one can see how the layout-dependent interaction between wind turbines' aerodynamics, from which wake arises, lends itself to be modelled as a graph. Thus, we introduce the use of GNNs for layout-agnostic joint modelling of rotor-averaged wind speed, turbulence intensity, power production and damage equivalent loads (DEL) on individual turbines of wind farms. To this end, probabilistic samples of inflow conditions from a Weibull distribution for wind speed, uniform wind direction and conditional normal distributions of wind shear and nacelle yaw angles are generated as main inflow properties in the numerical simulations. Additionally, arbitrary wind farm layouts are created based on varied geometric shapes with random parametrization (varying orientations, length/width ratio). Both the arbitrary layouts and the random inflow conditions are then used as inputs for PyWake. PyWake's output, namely, rotor averaged wind speed, turbulence intensity, power production and DELs of each individual wind turbine in the wind farms, are used to train a general GNN model. We elect to implement an Encode-Process-Decode GNN as a graph learning model. We compare the performance of various approaches of the relational dependency representation (edge-forming techniques) amongst the nodes (wind turbines) on the graph, such as Delaunay triangulation, KNN, radius-based methods and fully connected scheme. In our analysis, we evaluate the accuracy of GNNs and its ability to generalize their joint predictions for unseen layouts and varying inflow conditions.
Graph Transformers for inverse physics: reconstructing flows around arbitrary 2D airfoils
Duthé, Gregory; Abdallah, Imad; Chatzi, Eleni (2025)
arXiv
We introduce a Graph Transformer framework that serves as a general inverse physics engine on meshes, demonstrated through the challenging task of reconstructing aerodynamic flow fields from sparse surface measurements. While deep learning has shown promising results in forward physics simulation, inverse problems remain particularly challenging due to their ill-posed nature and the difficulty of propagating information from limited boundary observations. Our approach addresses these challenges by combining the geometric expressiveness of message-passing neural networks with the global reasoning of Transformers, enabling efficient learning of inverse mappings from boundary conditions to complete states. We evaluate this framework on a comprehensive dataset of steady-state RANS simulations around diverse airfoil geometries, where the task is to reconstruct full pressure and velocity fields from surface pressure measurements alone. The architecture achieves high reconstruction accuracy while maintaining fast inference times. We conduct experiments and provide insights into the relative importance of local geometric processing and global attention mechanisms in mesh-based inverse problems. We also find that the framework is robust to reduced sensor coverage. These results suggest that Graph Transformers can serve as effective inverse physics engines across a broader range of applications where complete system states must be reconstructed from limited boundary observations.
Stochastic estimation of wake-induced loads in wind farms using conformalized graph neural networks
Duthé, Gregory; de Nolasco Santos, Francisco; Chatzi, Eleni (2024)
e-Journal of Nondestructive Testing ~ Proceedings of the 11th European Workshop on Structural Health Monitoring (EWSHM 2024)
The inclusion of lifetime estimation into a holistic optimization approach for wind farms remains challenging today. This difficulty can partly be attributed to the complex effects of intra-farm wind turbine wakes, which significantly affect power production and cause non-linear fatigue behavior. A precise assessment of wake-induced loads is therefore a prerequisite for accurately estimating the Remaining Useful Lifetime (RUL) of turbines. In an effort to address this challenge, we introduce a novel approach which combines Graph Neural Networks (GNNs) with Conformal Predictors to model wind farms and evaluate wake-induced loads while estimating uncertainty. GNNs are particularly adept at capturing complex interactions in a wind farm due to their flexibility and ability to model graph-structured data both quickly and accurately. This allows for a tractable yet flexible load estimation framework. Alongside the GNN, we employ Conformal Predictors for uncertainty estimation. Conformalizing the GNN allows us to obtain statistically valid prediction sets based on past data with only minimal assumptions, thus yielding reliable confidence intervals. This is crucial for understanding the confidence associated with predictions under given site conditions (e.g. for a given wind speed/direction range) and for identifying conditions under which the GNN model may be less reliable. We test our approach on unseen wind farm layouts with varying inflow conditions. The fusion of GNNs with Conformal Predictors offers a robust framework for predicting turbine loads, while reliably quantifying the uncertainty associated with these predictions.
Learning to diagnose leading edge erosion degradation on an airfoil via aerodynamic pressure coefficients
Duthé, Gregory; Abdallah, Imad; Marykovskiy, Yuriy; et al. (2021)
Identifying the extent of erosion on the leading edge of wind turbine blades is becoming increasingly important in the wind energy industry as blades become larger and more flexible. Leading edge erosion (LEE) can result from abrasive airborne particles or weather conditions, and can impact the Annual Energy Production of a MW-scale wind turbine on the order of 5% (Langel et al., 2015). Current methods for identifying LEE involve manual (Nielsen et al., 2020) or drone-based visual inspection (Shihavuddin et al., 2019), electrical signal analysis (He et al., 2020) or vibration monitoring (Skrimpas et al., 2016) , methods which either require the turbine to be shut down or are limited for continuous monitoring (Du et al., 2020). In this work, we propose a data-driven model to predict the state of degradation of the leading edge of a 2D airfoil via aerodynamic pressure coefficient learning, under the influence of various uncertain inputs and parameters. The learning-based algorithm is trained on a novel dataset, comprised of pressure coefficient distributions that are generated by 2D steady-state Computational Fluid Dynamics (CFD) simulations. Each individual simulation is executed with a different set of input parameters, constructed via probabilistic sampling, with distributions designed to mimic operational conditions of a wind turbine. The airfoil LEE status can be broadly sorted into four categories corresponding to the severity of the degradation (Sareen et al., 2014): (1) undamaged, (2) presence of pits, (3) presence of pits and gouges and (4) presence of pits, gouges, and delamination. This categorization forms the prediction goal of the learning algorithm. In our CFD setup, which is based on the k-Omega SST Reynolds-averaged Navier–Stokes turbulence model using OpenFOAM, the surface conditions of the leading edge are emulated via the use of rough wall-functions, based on previous work by (Knopp et al., 2009). By adjusting the sand-grain roughness and the spatial extent of the rough patch, we can simulate all degradation categories. We validate this approach by comparing our simulation results in terms of lift, drag and pressure coefficients to experimental data and CFD simulation data from the literature (Maniaci et al., 2016). We aim to train a model which is robust to uncertain conditions and which can generalize to any type of airfoil. This requires a dataset established across a wide range of airfoil shapes and flow conditions. To increase robustness to shape warping induced by blade deflections, we propose a novel generative learning method to produce unique, random yet coherent, airfoil geometries, which are subsequently used as the basis for the CFD simulation meshes. As a result, during training, the learning algorithm is never exposed to multiple samples with the exact same airfoil geometry which allows for better generalization. By using random probabilistic sampling to draw the airfoil geometry and the other CFD flow condition inputs, we create an ensemble of unique and diverse pressure coefficient curves The deep learning algorithm we propose admits as inputs the pressure coefficient distribution and geometry of an airfoil, and outputs one of four degradation labels. A neural network is chosen as the data-driven model, although we test and compare multiple different network types and architectures. We show classification scores for all models and discuss their suitability for the task at hand, as well as potential improvements.
Physics-aware deep learning for wind prediction over complex terrain
Duthé, Gregory (2019)
Rapidly computing the wind flow over complex terrain features is a challenging problem with many potential fields of application. Such domains include autonomous UAV path planning, naturally ventilated building design or even wind farm layout optimization. Traditionally, this requires performing Computational Fluid Dynamics (CFD) simulations, which are computationally expensive and time consuming. One recent approach has been to use deep convolutional neural networks as a data-driven simulation alternative which leverages CFD simulation results for training. Albeit fast and relatively accurate, this method does not, however, explicitly take into account flow properties. We believe that doing so could improve prediction accuracy, especially in previously under-predicted areas which usually exhibit strong shearing flows and large velocity magnitudes. In this thesis, we investigate training deep neural networks with physics based loss functions and loss weighting methods, constructed to accentuate desirable flow features and properties. Furthermore, we explore other potential solutions, such as a physics informed adversarial approach, or using the recently proposed SPADE network which reported good spatial expressivity. Comparisons to baseline results are provided for all tested solutions, allowing us to select the combination of methods which provides the most accurate wind flow predictions. We show that training a network with a loss weighted with the velocity gradient magnitude results in a 5% decrease in maximum prediction error while retaining a comparable mean prediction error, when compared to the baseline approach.

Publications 1 - 10 of 23

Gregory Duthé

Last Name

First Name

ORCID

Organisational unit

Filters

Search Results