Particle image velocimetry for liquid phase flow based on deep learning

被引：0
|
作者：
Bi X. [1 ]
He M. [1 ]
Yu C. [2 ]
Fan Y. [3 ]
机构：
[1] School of Information Engineering, Minzu University of China, Beijing
[2] College of Information and Communication Engineering, Harbin Engineering University, Harbin
[3] College of Shipbuilding Engineering, Harbin Engineering University, Harbin
来源：
Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University | 2023年 / 44卷 / 04期
关键词：
convolutional neural network (CNN); deep learning; motion estimation; object entering water; optical flow; particle image velocimetry (PIV); single-phase flow; two-phase flow;
D O I：
10.11990/jheu.202108017
中图分类号：
学科分类号：
摘要：
In order to improve the accuracy and the generalization performance of the particle image velocimetry (PIV) algorithm in different fluid scenarios, this paper presents an improved deep learning model called PIV-RAFT-2P, which could be used for the estimation of the liquid phase velocity field of single-phase flow and two-phase flow of an object entering water at the same time. According to different characteristics of particle image and optical flow dataset image data, the optical flow model recurrent all-pairs field transforms (RAFT) is improved in a targeted manner; the corresponding single-phase flow and two-phase flow PIV datasets are autonomously constructed for training and optimization of model parameters. Finally, the proposed method is verified on the synthetic and real particle images. Experimental results show that the proposed model has obtained high-precision estimation results in single-phase flow and two-phase flow of an object entering water, having strong generalization. Meanwhile, the model has the advantages of computing efficiency with a small number of parameters and fast reasoning speed. © 2023 Editorial Board of Journal of Harbin Engineering. All rights reserved.
引用
收藏
页码：622 / 630
页数：8
相关论文
共 25 条
  • [1] RAFFEL M, WILLERT C E, WERELEY S T, Et al., Particle Image Velocimetry, (2007)
  • [2] WESTERWEEL J., Digital particle image velocimetry: Theory and application [ D ], (1993)
  • [3] HORN B, SCHUNCK B G., Determining optical flow [C], Techniques and Applications of Image Understanding, (1981)
  • [4] SCARANO F., Iterative image deformation methods in PIV, Measurement science and technology, 13, 1, pp. R1-R19, (2002)
  • [5] RUHNAU P, KOHLBERGER T, SCHNORR C, Et al., Variational optical flow estimation for particle image velocimetry, Experiments in fluids, 38, 1, pp. 21-32, (2005)
  • [6] HEITZ D, MEMIN E, SCHNORR C., Variational fluid flow measurements from image sequences: synopsis and perspectives [ J ], Experiments in fluids, 48, 3, pp. 369-393, (2010)
  • [7] SHE Wenxuan, ZHOU Guangli, WU Tiecheng, Et al., Experimental investigation of ship stern flow field through Stereo-PIV, Journal of Harbin Engineering University, 40, 1, pp. 26-33, (2019)
  • [8] YU Pengyao, WANG Tianlin, ZHAO Yong, Et al., Water entry of 2-D wedge by hydroelastic theory, Journal of Harbin Engineering University, 40, 4, pp. 689-695, (2019)
  • [9] VENNEMANN B, ROSGEN T., A dynamic masking technique for particle image velocimetry using convolutional autoencoders, Experiments in fluids, 61, 7, pp. 1-11, (2020)
  • [10] LEE Yong, YANG Hua, YIN Zhouping, PIV-DCNN: cascaded deep convolutional neural networks for particle image velocimetry, Experiments in fluids, 58, 12, pp. 1-10, (2017)
123