Luis J. Trujillo Corona
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Luis J. Trujillo Corona has a total of 4 co-authors. Here are the top co-authors by number of shared publications:
- Andreas Gross
Department of Mechanical and Aerospace Engineering , New Mexico State University , Las Cruces , NM , USA
1 shared publication David A. Espinoza Gutierrez
Department of Mechanical and Aerospace Engineering , New Mexico State University , Las Cruces , NM , USA
1 shared publication- Fangjun Shu
New Mexico State University
1 shared publication - Qiong Liu
Department of Mechanical and Aerospace Engineering , New Mexico State University , Las Cruces , NM , USA
1 shared publication
Latest Publications
- 12 August 2026
- Published by Walter de Gruyter GmbH in Transport Phenomena
- Vol. 1 (3)
Machine learning-based reconstruction of wake flow from surface pressure measurements
Off-body flow-field reconstruction from a limited number of surface sensor measurements is of significant practical value for applications where the flow field state is important for control but cannot be measured directly. In this work, we employ machine learning techniques to reconstruct the wake flow for a natural laminar flow airfoil NLF(1)-0115 at Re = 20, 000 and α = 5° from a limited number of pressure sensor measurements on the airfoil surface. The training data obtained from numerical simulations comprise the uncontrolled flow as well as controlled flows with steady and unsteady actuation. Two multilayer perceptron neural networks and a convolutional neural network were trained. A grid-search method was used to tune the hyperparameters of the neural networks. Hyperparameter tuning improved the convergence and reconstruction accuracy of the fully connected neural network models, which achieved excellent interpolation performance but showed limited cross-condition generalization. In contrast, the convolutional neural network demonstrated superior reconstruction performance and produced more accurate flow field reconstructions for cross-condition datasets. Among the test cases, all models achieved their best performance for the unsteady controlled flow. One possible explanation is that the controlled flow exhibits more organized and repeatable dynamics, which may facilitate reconstruction from sparse pressure measurements. These results suggest that incorporating both spatial and temporal flow information may improve cross-condition generalization and provide a promising direction for future flow-control applications.
- 12 August 2026
- Published by Walter de Gruyter GmbH in Transport Phenomena
- Vol. 1 (3)
Machine learning-based reconstruction of wake flow from surface pressure measurements
Off-body flow-field reconstruction from a limited number of surface sensor measurements is of significant practical value for applications where the flow field state is important for control but cannot be measured directly. In this work, we employ machine learning techniques to reconstruct the wake flow for a natural laminar flow airfoil NLF(1)-0115 at Re = 20, 000 and α = 5° from a limited number of pressure sensor measurements on the airfoil surface. The training data obtained from numerical simulations comprise the uncontrolled flow as well as controlled flows with steady and unsteady actuation. Two multilayer perceptron neural networks and a convolutional neural network were trained. A grid-search method was used to tune the hyperparameters of the neural networks. Hyperparameter tuning improved the convergence and reconstruction accuracy of the fully connected neural network models, which achieved excellent interpolation performance but showed limited cross-condition generalization. In contrast, the convolutional neural network demonstrated superior reconstruction performance and produced more accurate flow field reconstructions for cross-condition datasets. Among the test cases, all models achieved their best performance for the unsteady controlled flow. One possible explanation is that the controlled flow exhibits more organized and repeatable dynamics, which may facilitate reconstruction from sparse pressure measurements. These results suggest that incorporating both spatial and temporal flow information may improve cross-condition generalization and provide a promising direction for future flow-control applications.