computational fluid dynamics
PulseAugur coverage of computational fluid dynamics — every cluster mentioning computational fluid dynamics across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
Neural fields win aerodynamic prediction challenge with ensemble learning
Researchers have developed a novel approach using neural field ensembles to predict aerodynamic surface properties, achieving first place in the ONERA CRM Wall Distribution Regression Challenge. This method models the p…
-
Physics-Informed Neural Networks applied to aortic aneurysm study
Researchers have developed a novel three-dimensional Physics-Informed Neural Network (PINN) framework to study blood flow dynamics within the human aorta. This model simulates pulsatile blood flow over a two-minute peri…
-
New Flow State Attention Network Enhances Aerodynamic Prediction Accuracy
Researchers have introduced the Flow State Attention Network (FSAN), a novel deep learning model designed to improve the accuracy and applicability of aerodynamic predictions. Traditional computational fluid dynamics (C…
-
VFNet: New neural network estimates void fraction from multi-view videos
Researchers have developed VFNet, a novel dual-branch spatio-temporal neural network designed to estimate void fraction in gas-liquid two-phase flow using synchronized multi-view videos. This method overcomes limitation…
-
New framework fuses CFD and wind-tunnel data to boost aerospace model accuracy
Researchers have developed a novel data fusion framework to improve the accuracy of aerospace surrogate models by integrating experimental wind-tunnel data with computational fluid dynamics (CFD) simulations. The framew…
-
Deep learning enhances 4D flow MRI for better blood flow assessment
Researchers have developed a deep learning framework to improve the resolution and reduce noise in 4D flow MRI data, a technique used for visualizing blood flow. The proposed model integrates multi-scale feature extract…
-
New VATO method enhances prediction of unsteady aerofoil flows
Researchers have developed VATO (Vortex-Force-Aware Transformer Operator), a novel approach that integrates the Vortex Force Map (VFM) method with a geometry-aware neural operator to improve the prediction of unsteady s…
-
New framework models constrained multi-output physical fields using row-wise PCA and GP
Researchers have developed a novel framework for simultaneously predicting multiple high-dimensional physical fields that are subject to linear equality constraints. This problem is common in physics and machine learnin…
-
New STCO framework enhances neural operator predictions for time-dependent PDEs
Researchers have introduced STCO, a novel conditional neural operator designed for time-dependent partial differential equations (PDEs). This new framework allows for predictions conditioned on prescribed inputs like bo…
-
Residual Learning Improves Unsteady Aerodynamic Load Prediction
Researchers have explored the use of residual learning with long short-term memory (LSTM) neural networks to enhance the prediction of unsteady aerodynamic loads for aeroelastic applications. The study utilized the NLR …
-
New Deep Learning Model Enhances AAA Hemodynamic Prediction
Researchers have developed a novel Modified Multi-Input Multi-Output Physics-Informed DeepONet (M3PI-DeepONet) architecture to more accurately predict complex 3D blood flow dynamics in Abdominal Aortic Aneurysms (AAAs).…
-
Neural ODEs forecast power transformer thermal behavior
Researchers have developed a novel framework using Neural Ordinary Differential Equations (Neural ODEs) to accurately model and forecast the thermal behavior of power transformers. This physics-aware approach integrates…
-
New ADEx-FNO framework unifies Fourier Neural Operators for complex geometries
Researchers have introduced ADEx-FNO, a novel framework designed to enhance Fourier Neural Operators (FNOs) for applications involving complex and varying geometries. This method embeds physical domains within a fixed a…
-
New PCINN model predicts SALD surface coverage with high accuracy
Researchers have developed a Physics-Chemistry-Informed Neural Network (PCINN) designed to predict surface coverage in spatial atomic layer deposition (SALD) with high accuracy and speed. This hybrid surrogate model ach…
-
Robostreet Flow architecture slashes freight costs with electric, automated convoys
A new research paper introduces "Robostreet Flow," an innovative architecture for freight transportation designed to significantly reduce costs. The system features a lightweight, ultra-low-drag electric tractor with a …
-
AI research accelerates PDE solvers with novel Newton and Transformer methods
Two new research papers propose novel methods for accelerating the solution of complex partial differential equations (PDEs) using machine learning techniques. The first paper introduces a two-stage Newton initial guess…
-
Label-free training method for neural surrogates in fluid dynamics
Researchers have developed a novel method for training neural surrogates for thermo-fluid field predictions, utilizing a label-free approach based on minimizing finite-volume method (FVM) residuals. This technique, appl…
-
New AI framework optimizes hydrogen leak detection sensors
Researchers have developed a novel computational framework to optimize sensor placement for detecting hydrogen leaks in enclosed infrastructure, such as vehicle parking facilities. This system integrates computational f…
-
CFD transforms marine engineering with advanced ship design simulations
Computational Fluid Dynamics (CFD) is revolutionizing marine engineering by enabling detailed simulations of fluid behavior before ship construction. This technology allows engineers to optimize ship design, predict per…
-
Fortran code gains automatic differentiation via LFortran and Enzyme
Researchers have developed a method to enable automatic differentiation for legacy Fortran code, allowing it to be integrated into modern machine learning frameworks like JAX and PyTorch. This approach uses LFortran to …