Adrián García-Gutiérrez
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Adrián García-Gutiérrez has a total of 36 co-authors. Here are the top co-authors by number of shared publications:
- Diego Domínguez
University of Leon
13 shared publications - Jesús Gonzalo
Universidad de León
12 shared publications - Adrián Delgado
Aerospace Engineering Department, I4 Institute, Universidad de León, 24071 León, Spain
10 shared publications - Deibi López
Aerospace Engineering Area, Universidad de León, 24071 León, Spain
10 shared publications - Carlos Rubio
Aerospace Engineering Department, Universidad de León, León, Spain
7 shared publications - Fernando Parra
Aerospace Engineering Department, I4 Institute, Universidad de León, 24071 León, Spain
4 shared publications
Latest Publications
- Open Access
- 13 October 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (20) , 3507
Physics-Informed Neural Networks for Multi-Location ABL Wind Profiling with Minimal Instrumentation
The vertical wind profile within the Atmospheric Boundary Layer (ABL) is a critical variable for a wide range of applications, including wind energy, air traffic and Unmanned Aerial System (UAS) management, pollutant dispersion, and urban air mobility. However, its accurate characterization typically requires expensive vertical-profiling LiDAR instruments, which limits the number of locations that can be simultaneously monitored. Building upon previous work in which machine-learning algorithms were trained on LiDAR data from the University of León (Spain) to reconstruct full ABL wind profiles from a single near-surface measurement, this study replaces the traditional data-driven neural network approach with Physics-Informed Neural Networks (PINNs), embedding a physical constraint directly into the loss function as evaluated by automatic differentiation, and tests two such constraints under architectural control: the Ekman boundary-layer momentum balance, closed with a Monin–Obukhov eddy viscosity, and a shear-constrained formulation based on a power-law profile with an inferred exponent. With dense supervision at the target heights, neither constraint improves on an architecture-matched network without a physical term (RMSE 1.35–1.36 m/s, direction MAE 20.1–20.4° across all three, Kruskal–Wallis ). In a vertical-extrapolation protocol, training with gates up to 111 m, selecting models by their skill at 140 m, and testing blind on 170–300 m, the Ekman constraint shows no detectable advantage over its own architectural control. In contrast, the shear-constrained model does, reaching m/s speed RMSE and direction MAE against m/s and for the Ekman-constrained arm and m/s and for the physics-free arm, which are statistically indistinguishable from one another; the shear-constrained gain is confirmed against its own architectural control ( m/s, , matching the physics-free arm), ruling out the additional output as the explanation, and converts an erratic extrapolator into a stable one. The results are consistent with the benefit of embedding physics into the loss being therefore specific to the extrapolation regime and to the formulation tested, not a generic property of physics-informed learning. These findings are based on data from a single suburban site, and multi-site validation remains necessary before they can be generalized. By eliminating the need for continuous LiDAR operation after the training phase, the proposed methodology aims to enable cost-effective monitoring of multiple locations with a single instrument.- 13 October 2026
- Published by MDPI AG in Remote Sensing
- Vol. 18 (20) , 3507
Open AccessPhysics-Informed Neural Networks for Multi-Location ABL Wind Profiling with Minimal Instrumentation
The vertical wind profile within the Atmospheric Boundary Layer (ABL) is a critical variable for a wide range of applications, including wind energy, air traffic and Unmanned Aerial System (UAS) management, pollutant dispersion, and urban air mobility. However, its accurate characterization typically requires expensive vertical-profiling LiDAR instruments, which limits the number of locations that can be simultaneously monitored. Building upon previous work in which machine-learning algorithms were trained on LiDAR data from the University of León (Spain) to reconstruct full ABL wind profiles from a single near-surface measurement, this study replaces the traditional data-driven neural network approach with Physics-Informed Neural Networks (PINNs), embedding a physical constraint directly into the loss function as evaluated by automatic differentiation, and tests two such constraints under architectural control: the Ekman boundary-layer momentum balance, closed with a Monin–Obukhov eddy viscosity, and a shear-constrained formulation based on a power-law profile with an inferred exponent. With dense supervision at the target heights, neither constraint improves on an architecture-matched network without a physical term (RMSE 1.35–1.36 m/s, direction MAE 20.1–20.4° across all three, Kruskal–Wallis ). In a vertical-extrapolation protocol, training with gates up to 111 m, selecting models by their skill at 140 m, and testing blind on 170–300 m, the Ekman constraint shows no detectable advantage over its own architectural control. In contrast, the shear-constrained model does, reaching m/s speed RMSE and direction MAE against m/s and for the Ekman-constrained arm and m/s and for the physics-free arm, which are statistically indistinguishable from one another; the shear-constrained gain is confirmed against its own architectural control ( m/s, , matching the physics-free arm), ruling out the additional output as the explanation, and converts an erratic extrapolator into a stable one. The results are consistent with the benefit of embedding physics into the loss being therefore specific to the extrapolation regime and to the formulation tested, not a generic property of physics-informed learning. These findings are based on data from a single suburban site, and multi-site validation remains necessary before they can be generalized. By eliminating the need for continuous LiDAR operation after the training phase, the proposed methodology aims to enable cost-effective monitoring of multiple locations with a single instrument. - Open Access
- 9 September 2026
- Published by MDPI AG in Preprints.org
Physics-Informed Neural Networks for Multi-Location ABL Wind Profiling with Minimal Instrumentation
The vertical wind profile within the Atmospheric Boundary Layer (ABL) is a critical variable 2 for a wide range of applications, including wind energy, air traffic and unmanned aerial 3 system (UAS) management, pollutant dispersion, and urban air mobility. However, its 4 accurate characterization typically requires expensive vertical-profiling LiDAR instruments, 5 which limits the number of locations that can be simultaneously monitored. Building upon 6 previous work in which machine-learning algorithms were trained on LiDAR data from the 7 University of León (Spain) to reconstruct full ABL wind profiles from a single near-surface 8 measurement, this study replaces the traditional data-driven neural network approach with 9 Physics-Informed Neural Networks (PINNs), embedding a physical constraint directly into 10 the loss function as evaluated by automatic differentiation, and tests two such constraints 11 under architectural control: the Ekman boundary-layer momentum balance, closed with a 12 Monin-Obukhov eddy viscosity, and a shear-constrained formulation based on a power-law 13 profile with an inferred exponent. With dense supervision at the target heights, neither 14 constraint improves on an architecture-matched network without a physical term (RMSE 15 1.35-1.36 m/s, direction MAE 20.1-20.4◦ across all three, Kruskal-Wallis p = 0.0125), and 16 in a vertical-extrapolation protocol, training with gates up to 111 m, selecting models 17 by their skill at 140 m, and testing blind on 170-300 m, the Ekman constraint shows 18 no advantage over its own architectural control, but the shear-constrained model does, 19 reaching 1.80 ± 0.01 m/s speed RMSE and 27.0◦ ± 0.3◦ direction MAE against 2.66 ± 0.10 20 m/s and 34.2◦ ± 1.8◦ for the Ekman-constrained arm and 2.61 ± 0.12 m/s and 33.9◦ ± 21 2.0◦ for the physics-free arm, which are statistically indistinguishable from one another; 22 the shear-constrained gain is confirmed against its own architectural control (2.64 ± 0.08 23 m/s, 33.0◦ ± 1.8◦, matching the physics-free arm), ruling out the additional output as 24 the explanation, and converts an erratic extrapolator into a stable one. The benefit of 25 embedding physics into the loss is therefore specific to the extrapolation regime and to the 26 formulation tested, not a generic property of physics-informed learning. By eliminating the 27 need for continuous LiDAR operation after the training phase, the proposed methodology 28 aims to enable cost-effective monitoring of multiple locations with a single instrument.
- 9 September 2026
- Published by MDPI AG in Preprints.org
Open AccessPhysics-Informed Neural Networks for Multi-Location ABL Wind Profiling with Minimal Instrumentation
The vertical wind profile within the Atmospheric Boundary Layer (ABL) is a critical variable 2 for a wide range of applications, including wind energy, air traffic and unmanned aerial 3 system (UAS) management, pollutant dispersion, and urban air mobility. However, its 4 accurate characterization typically requires expensive vertical-profiling LiDAR instruments, 5 which limits the number of locations that can be simultaneously monitored. Building upon 6 previous work in which machine-learning algorithms were trained on LiDAR data from the 7 University of León (Spain) to reconstruct full ABL wind profiles from a single near-surface 8 measurement, this study replaces the traditional data-driven neural network approach with 9 Physics-Informed Neural Networks (PINNs), embedding a physical constraint directly into 10 the loss function as evaluated by automatic differentiation, and tests two such constraints 11 under architectural control: the Ekman boundary-layer momentum balance, closed with a 12 Monin-Obukhov eddy viscosity, and a shear-constrained formulation based on a power-law 13 profile with an inferred exponent. With dense supervision at the target heights, neither 14 constraint improves on an architecture-matched network without a physical term (RMSE 15 1.35-1.36 m/s, direction MAE 20.1-20.4◦ across all three, Kruskal-Wallis p = 0.0125), and 16 in a vertical-extrapolation protocol, training with gates up to 111 m, selecting models 17 by their skill at 140 m, and testing blind on 170-300 m, the Ekman constraint shows 18 no advantage over its own architectural control, but the shear-constrained model does, 19 reaching 1.80 ± 0.01 m/s speed RMSE and 27.0◦ ± 0.3◦ direction MAE against 2.66 ± 0.10 20 m/s and 34.2◦ ± 1.8◦ for the Ekman-constrained arm and 2.61 ± 0.12 m/s and 33.9◦ ± 21 2.0◦ for the physics-free arm, which are statistically indistinguishable from one another; 22 the shear-constrained gain is confirmed against its own architectural control (2.64 ± 0.08 23 m/s, 33.0◦ ± 1.8◦, matching the physics-free arm), ruling out the additional output as 24 the explanation, and converts an erratic extrapolator into a stable one. The benefit of 25 embedding physics into the loss is therefore specific to the extrapolation regime and to the 26 formulation tested, not a generic property of physics-informed learning. By eliminating the 27 need for continuous LiDAR operation after the training phase, the proposed methodology 28 aims to enable cost-effective monitoring of multiple locations with a single instrument.
- Open Access
- 9 July 2026
- Published by Elsevier in SSRN Electronic Journal
Analysis of the accuracy and computational efficiency of transition–turbulence models for propeller performance prediction at low and high Reynolds Numbers
The emergence of new aerospace vehicle concepts has renewed interest inoptimizing propeller design. Among them, stratospheric vehicles or eVTOLaircraft have in common that their propellers operate at low Reynolds num-bers. To determine propeller performance parameters, computational fluiddynamics is a widely used technique. Hence, this research assesses the accu-racy and efficiency of different transition-turbulence models to determine theconvenience —or not— of using a more complex turbulence model. The pre-diction of thrust and power coefficients for propellers operating at both highand low Reynolds numbers are compared using different turbulence models.From the analysis conducted in this work, at high Reynolds number operat-ing conditions, transition-turbulence models provide no advantage, whereasat low Reynolds numbers, γ - Reθ SST model is preferred. Additionally,at moderate Reynolds number operating conditions γ SST transition modeloffers the best compromise between accuracy and computational efficiency.These findings provide valuable insights for selecting turbulence models inpropeller aerodynamic design and help to reduce its computational cost.
- 9 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessAnalysis of the accuracy and computational efficiency of transition–turbulence models for propeller performance prediction at low and high Reynolds Numbers
The emergence of new aerospace vehicle concepts has renewed interest inoptimizing propeller design. Among them, stratospheric vehicles or eVTOLaircraft have in common that their propellers operate at low Reynolds num-bers. To determine propeller performance parameters, computational fluiddynamics is a widely used technique. Hence, this research assesses the accu-racy and efficiency of different transition-turbulence models to determine theconvenience —or not— of using a more complex turbulence model. The pre-diction of thrust and power coefficients for propellers operating at both highand low Reynolds numbers are compared using different turbulence models.From the analysis conducted in this work, at high Reynolds number operat-ing conditions, transition-turbulence models provide no advantage, whereasat low Reynolds numbers, γ - Reθ SST model is preferred. Additionally,at moderate Reynolds number operating conditions γ SST transition modeloffers the best compromise between accuracy and computational efficiency.These findings provide valuable insights for selecting turbulence models inpropeller aerodynamic design and help to reduce its computational cost.
- 1 June 2026
- Published by IATED Academy in EDULEARN Proceedings
- Vol. 1, 1304
ULESAT, A SATELLITE FOR EDUCATIONAL INNOVATION: HANDS‑ON SUBSYSTEM DEVELOPMENT
- 1 June 2026
- Published by IATED Academy in EDULEARN Proceedings
- Vol. 1, 1304
ULESAT, A SATELLITE FOR EDUCATIONAL INNOVATION: HANDS‑ON SUBSYSTEM DEVELOPMENT
- 1 June 2026
- Published by IATED Academy in EDULEARN Proceedings
- Vol. 1, 1307
FROM HANGAR TO CLASSROOM: AUGMENTED REALITY STATIONS ON A TRAINING GLIDER FOR AN AERODYNAMICS COURSE
- 1 June 2026
- Published by IATED Academy in EDULEARN Proceedings
- Vol. 1, 1307
FROM HANGAR TO CLASSROOM: AUGMENTED REALITY STATIONS ON A TRAINING GLIDER FOR AN AERODYNAMICS COURSE
- 1 January 2026
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 36 (1)
Chaos meets stochasticity: A variance-based method for Lyapunov exponent estimation
The Largest Lyapunov Exponent (LLE) is a fundamental diagnostic of chaotic behavior in nonlinear dynamical systems, quantifying the exponential divergence of nearby trajectories. Classical computational approaches, such as Wolf’s algorithm, track individual particle trajectories to estimate the LLE, but these techniques face challenges related to noise sensitivity, computational efficiency, and scalability to high-dimensional systems. This work introduces a novel variance-based methodology for computing the LLE using intrusive polynomial chaos (IPC), an uncertainty quantification technique that evolves the probability distribution of initial conditions under deterministic dynamics rather than tracking discrete trajectories. The key innovation is extracting the LLE from the exponential growth rate of ensemble variance, which connects deterministic chaos with probabilistic descriptions. Validation against the classical trajectory-based algorithm is performed on three benchmark chaotic systems: the three-dimensional Lorenz and Rössler attractors, and a six-dimensional system from Al-Azzawi and Al-Obeidi, demonstrating that the IPC approach achieves comparable accuracy and convergence rates while offering the distinct advantage of directly computing the full statistical structure of ensemble dynamics. Comparison of convergence histories, probability density functions of instantaneous Lyapunov exponents, and statistical error measures confirms excellent agreement between the proposed IPC-based methodology and established algorithms. The results indicate that variance-based LLE estimation via polynomial chaos is a robust and viable alternative to trajectory-based methods.
- 1 January 2026
- Published by AIP Publishing in Chaos: An Interdisciplinary Journal of Nonlinear Science
- Vol. 36 (1)
Chaos meets stochasticity: A variance-based method for Lyapunov exponent estimation
The Largest Lyapunov Exponent (LLE) is a fundamental diagnostic of chaotic behavior in nonlinear dynamical systems, quantifying the exponential divergence of nearby trajectories. Classical computational approaches, such as Wolf’s algorithm, track individual particle trajectories to estimate the LLE, but these techniques face challenges related to noise sensitivity, computational efficiency, and scalability to high-dimensional systems. This work introduces a novel variance-based methodology for computing the LLE using intrusive polynomial chaos (IPC), an uncertainty quantification technique that evolves the probability distribution of initial conditions under deterministic dynamics rather than tracking discrete trajectories. The key innovation is extracting the LLE from the exponential growth rate of ensemble variance, which connects deterministic chaos with probabilistic descriptions. Validation against the classical trajectory-based algorithm is performed on three benchmark chaotic systems: the three-dimensional Lorenz and Rössler attractors, and a six-dimensional system from Al-Azzawi and Al-Obeidi, demonstrating that the IPC approach achieves comparable accuracy and convergence rates while offering the distinct advantage of directly computing the full statistical structure of ensemble dynamics. Comparison of convergence histories, probability density functions of instantaneous Lyapunov exponents, and statistical error measures confirms excellent agreement between the proposed IPC-based methodology and established algorithms. The results indicate that variance-based LLE estimation via polynomial chaos is a robust and viable alternative to trajectory-based methods.
- Open Access
- 21 October 2025
- Published by Springer Nature in Theoretical and Computational Fluid Dynamics
- Vol. 39 (6) , 1-22
Dynamical response of a Newtonian fluid interface to mixed oscillatory deformations
This study presents a numerical investigation of the dynamic response of a Newtonian fluid interface subjected to mixed oscillatory deformations. A slender cylindrical probe, floating horizontally at a water–air interface and partially submerged by capillary forces, is driven sinusoidally in the direction perpendicular to its axis. The interface exhibits shear, dilatational, and extensional viscosities, and the system is modeled using the finite volume method within the OpenFOAM framework. The governing equations are nondimensionalized and solved for a wide range of Marangoni numbers ( $$Ma \in [1, 10^5]$$ ) and dilatational-to-shear viscosity ratios ( $$\Theta \in [1, 10^5]$$ ). The simulations enable the decomposition of the total interfacial force into its constituent components, revealing distinct regimes dominated by Marangoni, dilatational, or extensional stresses. The results demonstrate the feasibility of isolating these contributions under specific conditions, offering insights into the design and interpretation of interfacial rheometry experiments. The influence of geometric parameters and oscillation characteristics is also explored, confirming the robustness of the observed force dynamics.- 21 October 2025
- Published by Springer Nature in Theoretical and Computational Fluid Dynamics
- Vol. 39 (6) , 1-22
Open AccessDynamical response of a Newtonian fluid interface to mixed oscillatory deformations
This study presents a numerical investigation of the dynamic response of a Newtonian fluid interface subjected to mixed oscillatory deformations. A slender cylindrical probe, floating horizontally at a water–air interface and partially submerged by capillary forces, is driven sinusoidally in the direction perpendicular to its axis. The interface exhibits shear, dilatational, and extensional viscosities, and the system is modeled using the finite volume method within the OpenFOAM framework. The governing equations are nondimensionalized and solved for a wide range of Marangoni numbers ( $$Ma \in [1, 10^5]$$ ) and dilatational-to-shear viscosity ratios ( $$\Theta \in [1, 10^5]$$ ). The simulations enable the decomposition of the total interfacial force into its constituent components, revealing distinct regimes dominated by Marangoni, dilatational, or extensional stresses. The results demonstrate the feasibility of isolating these contributions under specific conditions, offering insights into the design and interpretation of interfacial rheometry experiments. The influence of geometric parameters and oscillation characteristics is also explored, confirming the robustness of the observed force dynamics. - 1 October 2025
- Published by Springer Nature
- p. 73-87
An Agile Inflatable-Structure Airship for Air Freight: The iRPAS Project
- 1 October 2025
- Published by Springer Nature
- p. 73-87
An Agile Inflatable-Structure Airship for Air Freight: The iRPAS Project
- 27 April 2025
- Published by Springer Nature
- p. 535-542
Design and Performance of Micro Wind Turbine Airfoils Under Uncertainty
- 27 April 2025
- Published by Springer Nature
- p. 535-542
Design and Performance of Micro Wind Turbine Airfoils Under Uncertainty