DeepPTV: Particle Tracking Velocimetry for Complex Flow Motion via Deep Neural Networks

Liang, Jiaming; Cai, Shengze; Xu, Chao; Chen, Tehuan; Chu, Jian

doi:10.1109/tim.2021.3120127

Cited by 12 publications

(10 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In two related papers, Liang et al (2022) and Liang et al (2023) present novel deep learning approaches to PTV for efficient and accurate flow field measurements. DeepPTV, the first method, uses a deep neural network to learn complex fluid flow motion from consecutive particle sets, aggregating local spatial geometry information from neighboring particles.…”

Section: Deep Learning For Pivmentioning

confidence: 99%

Real-time and On-site Aerodynamics using Stereoscopic PIV and Deep Optical Flow Learning

Elrefaie,

Hüttig,

Gladkova

et al. 2024

Preprint

View full text Add to dashboard Cite

We introduce Recurrent All-Pairs Field Transforms for Stereoscopic ParticleImage Velocimetry (RAFT-StereoPIV). Our approach leverages deep optical flowlearning to analyze time-resolved and double-frame particle images from on-sitemeasurements, particularly from the 'Ring of Fire,' as well as from wind tunnelmeasurements for real-time aerodynamic analysis. A multi-fidelity datasetcomprising both Reynolds-Averaged Navier-Stokes (RANS) and Direct NumericalSimulation (DNS) was used to train our model. RAFT-StereoPIV outperforms allPIV state-of-the-art deep learning models on benchmark datasets, with a 68% error reduction on the validation dataset, Problem Class 2, and a 47% errorreduction on the unseen test dataset, Problem Class 1, demonstrating itsrobustness and generalizability. In comparison to the most recent works in thefield of deep learning for PIV, where the main focus was the methodologydevelopment and the application was limited to either 2D flow cases or simpleexperimental data, we extend deep learning-based PIV for industrialapplications and 3D flow field estimation. As we apply the trained network tothree-dimensional highly turbulent PIV data, we are able to obtain flowestimates that maintain spatial resolution of the input image sequence. Incontrast, the traditional methods produce the flow field of~16 × lower resolution. We believe that this study brings the field of experimentalfluid dynamics one step closer to the long-term goal of having experimentalmeasurement systems that can be used for real-time flow field estimation.

show abstract

Section: Deep Learning For Pivmentioning

confidence: 99%

Real-time and On-site Aerodynamics using Stereoscopic PIV and Deep Optical Flow Learning

Elrefaie,

Hüttig,

Gladkova

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Liang et al [37] proposed a novel architecture called DeepPTV, which uses deep neural networks to learn complex fluid motion across different scales (Figure 7). The DeepPTV architecture features an enhanced multi-scale feature learner and a convection architecture.…”

Section: Deepptv: a Deep Neural Network-based Frameworkmentioning

confidence: 99%

“…The advent of data-driven PTV harnesses the power of machine learning to enhance analysis and interpretation. Examples of such advancements include PTV using shallow neural networks [36], DeepPTV [37], PINN-augmented PTV [38], LSTM-enhanced PTV [39], and stochastic particle advection velocimetry (SPAV) [40], among others. Each of these approaches offers unique advantages in terms of accuracy, processing speed, and the ability to handle complex flow scenarios.…”

Section: Introductionmentioning

confidence: 99%

Micro-Scale Particle Tracking: From Conventional to Data-Driven Methods

Wang,

Hong,

Chamorro

2024

Micromachines

View full text Add to dashboard Cite

Micro-scale positioning techniques have become essential in numerous engineering systems. In the field of fluid mechanics, particle tracking velocimetry (PTV) stands out as a key method for tracking individual particles and reconstructing flow fields. Here, we present an overview of the micro-scale particle tracking methodologies that are predominantly employed for particle detection and flow field reconstruction. It covers various methods, including conventional and data-driven techniques. The advanced techniques, which combine developments in microscopy, photography, image processing, computer vision, and artificial intelligence, are making significant strides and will greatly benefit a wide range of scientific and engineering fields.

show abstract

“…Architectures based on long short-term memory (LSTM) RNNs have been proposed [51,52] to predict particle positions given by the past trajectory. Liang et al [53] introduced an architecture based on a modified version of the end-to-end scene flow estimator FlowNet3D [54]. FlowNet3D contains three modules: hierarchical point feature learning based on PointNet++ [55], flow embedding based on two consecutive views of the point cloud, and upsampling and refinement.…”

Section: Extraction Of Velocity Fields Using Neural Network (Nns)mentioning

confidence: 99%

Machine learning for flow field measurements: a perspective

Discetti¹,

Liu²

2022

Meas. Sci. Technol.

View full text Add to dashboard Cite

Advancements in machine-learning (ML) techniques are driving a paradigm shift in image processing. Flow diagnostics with optical techniques is not an exception. Considering the existing and foreseeable disruptive developments in flow field measurement techniques, we elaborate this perspective, particularly focused to the field of particle image velocimetry. The driving forces for the advancements in ML methods for flow field measurements in recent years are reviewed in terms of image preprocessing, data treatment and conditioning. Finally, possible routes for further developments are highlighted.

show abstract

DeepPTV: Particle Tracking Velocimetry for Complex Flow Motion via Deep Neural Networks

Cited by 12 publications

References 47 publications

Real-time and On-site Aerodynamics using Stereoscopic PIV and Deep Optical Flow Learning

Real-time and On-site Aerodynamics using Stereoscopic PIV and Deep Optical Flow Learning

Micro-Scale Particle Tracking: From Conventional to Data-Driven Methods

Machine learning for flow field measurements: a perspective

Contact Info

Product

Resources

About