基于改进MobileViT的气固两相流检测方法

仝卫国; 王用沛; 林礼榕; 陈波帆; 钟海增

doi:10.19805/j.cnki.jcspe.2026.250063

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于改进MobileViT的气固两相流检测方法

数字化与智能化 | 更新时间：2026-10-10

- 基于改进MobileViT的气固两相流检测方法
- Gas-Solid Two-phase Flow Detection Method Based on Improved MobileViT
- 动力工程学报 2026年46卷第5期页码：154-164
- 作者机构：
  
  1. 华北电力大学自动化系,河北,保定,071003
  2. 华北电力大学河北省发电过程仿真与优化控制技术创新中心,河北,保定,071003
- 作者简介：
  
  仝卫国(1967—),男,河北保定人,副教授,博士研究生,主要从事先进检测技术的研究。王用沛(通信作者),男,硕士研究生,E-mail:15866049603@163.com。
- 基金信息：
  
  河北省省级科技计划资助项目(22567643H)
- DOI：10.19805/j.cnki.jcspe.2026.250063
  中图分类号： TB937
- 纸质出版：2026
- 稿件说明：
移动端阅览
仝卫国, 王用沛, 林礼榕, 陈波帆, 钟海增. 基于改进MobileViT的气固两相流检测方法[J]. 动力工程学报, 2026, 46(5): 154-164.
仝卫国, 王用沛, 林礼榕, 陈波帆, 钟海增. 基于改进MobileViT的气固两相流检测方法[J]. 动力工程学报, 2026, 46(5): 154-164. DOI： 10.19805/j.cnki.jcspe.2026.250063.

DOI： 10.19805/j.cnki.jcspe.2026.250063

摘要

针对气固两相流流动特性复杂导致流动参数难以准确检测的问题

结合神经网络技术对其流动过程中产生的音频信号进行分析

提出一种基于改进MobileViT的气固两相流检测方法。以轻量化网络MobileViT作为基础模型

首先通过谱减法对采集到的音频信号进行去噪

并提取其梅尔频谱图作为模型的输入;其次

在模型中引入双向空洞空间金字塔池化模块

提取特征图在水平和垂直维度上的多尺度特征信息;然后

在模型中引入全局局部空间注意力机制

增强模型对特征图关键区域特征信息的捕捉和表达能力;最后

使用细节增强卷积替换MobileViT block中的3×3标准卷积

以提取出更为丰富的特征图局部细节特征信息。结果表明:改进后的模型对6种流量条件下的气固两相流具有较好的识别效果

准确率为98.833%

较原模型提高3.166%。

Abstract

To address issues of accurately detecting flow parameters due to the complex flow characteristics of gas-solid two-phase flow

a detection method based on the improved MobileViT model was proposed

which combined neural network technology to analyze audio signals generated during the flow process of gas-solid two-phase flow. Lightweight MobileViT network was used as the base model. First

spectral subtraction was applied to denoise the collected audio signals

and their Mel spectrogram was extracted as model input. Next

a bidirectional dilated spatial pyramid pooling module was introduced into the model to extract multi-scale feature information in both horizontal and vertical dimensions. Then

a global-local spatial attention mechanism was incorporated to enhance the model's ability to capture and express the key feature information in feature map. Finally

a detail-enhanced convolution replaced the standard 3×3 convolution in the MobileViT block to extract richer local detail features from the feature map. Results show that the improved model achieves good recognition performance for gas-solid two-phase flow under six different flow conditions

with accuracy of 98.833% and 3.166% improvement over the original model.

关键词

Keywords

references

谢心喻, 王晓放, 郝祎琛, 等. 基于深度学习的全三维气固两相流时空耦合智能预测[J]. 工程热物理学报, 2024, 45(2): 446-452.XIE Xinyu, WANG Xiaofang, HAO Yichen, et al. Deep learning-based intelligent spatio-temporal prediction of full 3D gas-solid two-phase flow[J]. Journal of Engineering Thermophysics, 2024, 45(2): 446-452.

FAN J H, WANG F, CUI H B, et al. Concentration determination in a cylinder-simulated gas-solid two phase flow using ultrasonic backscattering method[J]. Applied Acoustics, 2023, 203: 109212.

田好雨, 高峰, 李炳辰, 等. 锅炉一次风煤粉流量测量技术研究综述[J]. 工业计量, 2024, 34(2): 26-30. TIAN Haoyu, GAO Feng, LI Bingchen, et al. A review of pulverized coal flow measurement in the primary air pipe of boilers[J]. Industrial Measurement, 2024, 34(2): 26-30.

薛飞宇, 韩健民, 梁双印. 电站锅炉双可调煤粉分配器气固两相流动特性研究[J]. 动力工程学报, 2021, 41(12): 1033-1039, 1108.XUE Feiyu, HAN Jianmin, LIANG Shuangyin. Research on gas-solid two-phase flow characteristics of double-adjustable pulverized coal distributors in power station boilers[J]. Journal of Chinese Society of Power Engineering, 2021, 41(12): 1033-1039, 1108.

石全成, 吕为智, 周文台, 等. 可调节煤粉分配器的风粉两相流动特性数值模拟[J]. 动力工程学报, 2022, 42(12): 1183-1190, 1205.SHI Quancheng, LÜ Weizhi, ZHOU Wentai, et al. Numerical simulation of two-phase flow characteristics of pulverized coal distributor[J]. Journal of Chinese Society of Power Engineering, 2022, 42(12): 1183-1190, 1205.

张立峰, 肖凯. 基于集成学习及电阻层析成像的两相流流型辨识[J]. 动力工程学报, 2023, 43(9): 1103-1110.ZHANG Lifeng, XIAO Kai. Identification of two-phase flow pattern based on ensemble learning and electrical resistance tomography[J]. Journal of Chinese Society of Power Engineering, 2023, 43(9): 1103-1110.

李慧君, 王业库, 鲁许鳌, 等. 基于EMMS-1M曳力模型的空气反应器气固流动及化学反应特性数值分析[J]. 动力工程学报, 2023, 43(7): 815-822, 834.LI Huijun, WANG Yeku, LU Xuao, et al. Numerical analysis of gas-solid flow and chemical reaction characteristics of air reactor based on EMMS-1M tracing force model[J]. Journal of Chinese Society of Power Engineering, 2023, 43(7): 815-822, 834.

华能平凉发电有限责任公司. 一种基于音频信号的煤粉细度测量方法及系统: 114910395A[P]. 2022-08-16.

LIU Z X, WANG H G, SUN S J, et al. Investigation of wetting and drying process in a spout-fluid bed using acoustic sensor and electrical capacitance tomography[J]. Chemical Engineering Science, 2023, 281:119160.

ZENG X X, YAN Y, QIAN X C, et al. Mass flow rate measurement of pneumatically conveyed solids in a square-shaped pipe through multisensor fusion and data-driven modeling[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 7508012.

WANG K, CHANG Z A, LI Y C, et al. The fusion of deep learning and acoustic emission response methods for identifying solid particles in annular multiphase flows[J]. Geoenergy Science and Engineering, 2023, 225: 211685.

王志春, 袁小健, 王月明, 等. 基于声发射的气固两相流质量流率及粒径的测量[J]. 仪表技术与传感器, 2017(7): 93-96.WANG Zhichun, YUAN Xiaojian, WANG Yueming, et al. Gas-solid two-phase mass flow rate and particle size measurement based on acoustic emission[J]. Instrument Technique and Sensor, 2017(7): 93-96.

曹翌佳, 王靖岱, 阳永荣. 声波信号多尺度分解与固体颗粒质量流率的测定[J]. 化工学报, 2007, 58(6): 1404-1410.CAO Yijia, WANG Jingdai, YANG Yongrong. Multi-scale analysis of acoustic emissions and measurement of particle mass flowrate in pipeline[J]. CIESC Journal, 2007, 58(6): 1404-1410.

HONG G, SUH D. Mel spectrogram-based advanced deep temporal clustering model with unsupervised data for fault diagnosis[J]. Expert Systems with Applications, 2023, 217: 119551.

仝卫国, 门国悦, 蔡天娇, 等. 基于音频信号的气固两相流检测方法研究[J]. 计量学报, 2025, 46(3): 383-390.TONG Weiguo, MEN Guoyue, CAI Tianjiao, et al. Research on gas-solid two-phase flow detection method based on audio signal[J]. Acta Metrologica Sinica, 2025, 46(3): 383-390.

NIU Z Y, ZHONG G Q, YU H. A review on the attention mechanism of deep learning[J]. Neurocomputing, 2021, 452: 48-62.

CHEN Z X, HE Z W, LU Z M. DEA-net: single image dehazing based on detail-enhanced convolution and content-guided attention[J]. IEEE Transactions on Image Processing, 2024, 33: 1002-1015.

司坤宇, 牛春晖. 基于混合差分卷积和高效视觉Transformer网络的三重多模态图像融合算法[J]. 红外与激光工程, 2024, 53(11): 20240305.SI Kunyu, NIU Chunhui. Triple multi-modal image fusion algorithm based on mixed difference convolution and efficient vision Transformer network[J]. Infrared and Laser Engineering, 2024, 53(11): 20240305.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于改进LSTM与注意力机制的SCR入口NOx生成浓度预测

基于IGRA和CNN-LSTM的垃圾焚烧炉主蒸汽温度预测

热电厂耦合蒸汽储能电热特性研究与优化

基于半马尔可夫的风电机组自触发变桨控制

基于WPHM和非支配性遗传算法的燃气轮机透平叶片多目标优化维护决策