基于图像多特征融合和支持向量机的气液两相流流型识别

The knowledge of flow regime is very important for quantifying the pressure drop, the stability and safety of two-phase flow systems. Based on image multi-feature fusion and support vector machine, a new method to identify flow regime in two-phase flow was presented. Firstly, gas-liquid two-phase flow images including bubbly flow, plug flow, slug flow, stratified flow, wavy flow, annular flow and mist flow were captured by digital high speed video systems in the horizontal tube. The image moment invariants and gray level co-occurrence matrix texture features were extracted using image processing techniques. To improve the performance of a multiple classifier system, the rough sets theory was used for reducing the inessential factors. Furthermore, the support vector machine was trained by using these eigenvectors to reduce the dimension as flow regime samples, and the flow regime intel-ligent identification was realized. The test results showed that image features which were reduced with the rough sets theory could excellently reflect the difference between seven typical flow regimes, and successful training the support vector machine could quickly and accurately identify seven typical flow regimes of gas-liquid two-phase flow in the horizontal tube. Image multi-feature fusion method provided a new way to identify the gas-liquid two-phase flow, and achieved higher identification ability than that of single characteristic. The overall identifica-tion accuracy was 100%, and an estimate of the image processing time was 8 ms for online flow regime identifica-tion.     

上升管中严重段塞流的流型和压力波动特性

During the exploitation of offshore oil and gas, it is easy to form severe slugging which can cause great harm in the riser connecting wellheads and offshore platform preprocessing system. The flow pattern and pressure fluctuation of severe slugging were studied in an experimental simulation system with inner diameter of 0.051 m. It is found that severe slugging can be divided into three severe slugging regimes: regime I at low gas and liquid flow rates with large pressure fluctuation, intermittent flow of liquid and gas in the riser, and apparent cutoff of liquid phase, regime II at high gas flow rate with non-periodic fluctuation and discontinuous liquid outflow and no gas cutoff, regime III at high liquid flow rate with degenerative pressure fluctuation in form of relatively stable bubbly or plug flow. The results indicate that severe slugging still occurs when the declination angle of pipeline is 0˚, and there are mainly two kinds of regimes: regime I and regime II. As the angle increases, the formation ranges of regime I and regime III increase slightly while that of regime II is not affected. With the increase of gas superficial velocity and liquid superficial velocity, the pressure fluctuation at the bottom of riser increases initially and then decreases. The maximum value of pressure fluctuation occurs at the transition boundary of regimes I and II.     

气水流动中气泡生成体积计算模型的改进

In order to address the bubble formation and movement in air-water two-phase flow,single bubble rising in stagnant water is experimentally studied by digital image processing.Bubbles are released individually from the submerged orifices with different diameters(1.81 mm,2.07 mm,2.98 mm,3.92 mm)at different detachment frequency.Images are recorded by a high-speed video camera and processed by digital image processing technique. The factors impacting the formed volume of bubble are discussed.The experimental results showed that a threshold of gas flow rate(400 mm 3 ·s- 1)divides the bubble formation into two regimes:the constant volume regime and the growing volume regime.Especially for the growing volume regime,the surface tension is taken into account.The bubble volume is consisted of two parts:the surface tension impacting part and the gas volume flow rate impacting part.An improved correlation for bubble volume prediction is developed for the two regimes and better coincidence with the experiment data than the previous models is obtained.     

气升式内环流反应器内局部气含率径向分布

The local gas holdup profiles in an internal-loop airlift reactor were studied experimentally by using dual electrical conductivity probe under different conditions,including superficial velocity,surface tension and liquid viscosity.The results showed that the radial gas holdup profile has a parabolic shape,which was consistent with the empirical model of Luo.Local gas holdup distribution parameters were obviously influenced by flow regime and almost remained unchanged in the same flow regime.In the gas distributor region,the profiles were steeper in the homogenous flow regime than in the heterogeneous flow regime.However,in the stable region,there was an inverse change trend in two flow regimes.The increase of surface tension,superficial velocity and liquid viscosity made the profile of local gas holdup steeper in two flow regimes.     

Flow Regime Identification of Gas-liquid Two-phase Flow Based on HHT

A new method to identify flow regime in two-phase flow was presented, based on signal processing of differential pressure using Hilbert Huang transform （HHT）. Signals obtained from a Venturi meter were decomposed into different intrinsic mode functions （IMFs） with HHT, then the energy fraction of each intrinsic mode and the mean value of residual function were calculated, from which the rules of flow regime identification were summarized. Experiments were carried out on two-phase flow in the horizontal tubes with 50mm and 40mm inner diameter, while water flowrate was in the range of 1.3m^3.h^-1 to 10.5m^3.h^-1, oil flowrate was from 4.2m^3.h^-1 to 7.0m^3.h^-1 and gas flowrate from 0 to 15m^3.h^-1. The results show that the proposed rules have high precision for single phase, bubbly, and slug, plug flow regirne identification, which are independent of not only properties of two-phase fluid. In addition, the method can meet the need of industrial application because of its simple calculation.     

水平安装文丘里管在低压湿气流动下的特性

The performance of a Venturi tube used in wet gas flow have been explored mainly under higher-pressure condition, but very often, low-pressure test exists in some oil and gas fields in Tianjin Dagang Oil and Gas Field in China. In this study, the performance of horizontally mounted Venturi meters in low-pressure wet gas flow is discussed. Three 50 mm Venturi meters were tested systematically, with fl values of 0.4048, 0.55 and 0.70, the opera- tion pressure of 0.15 MPa, 0.20 MPa, 0.25 MPa, the gas densiometric Froude number from 0.6 to 2.0, the modified Lockhart-Maretinelli parameter from 0.0022 to 0.06, and the ratio of the gas liquid mass flow rate from 0.5 to 0.99. The effects of modified Lockhart-Maretinelli parameter, pressure, gas densiometric Froude number, diameter ratio, and gas-liquid mass flow rate ratio to the Venturi tube are analyzed with new independent data. Furthermore, low-pressure performance was compared with that under high pressure.     

Measurment of gas-liquid two-phase slug flow with a Venturi meter based on blind source separation☆

We propose a novel flow measurement method for gas–liquid two-phase slug flow by using the blind source sep-aration technique. The flow measurement model is established based on the fluctuation characteristics of differ-ential pressure (DP) signals measured from a Venturi meter. It is demonstrated that DP signals of two-phase flow are a linear mixture of DP signals of single phase fluids. The measurement model is a combination of throttle re-lationship and blind source separation model. In addition, we estimate the mixture matrix using the independent component analysis (ICA) technique. The mixture matrix could be described using the variances of two DP sig-nals acquired from two Venturi meters. The validity of the proposed model was tested in the gas–liquid two-phase flow loop facility. Experimental results showed that for most slug flow the relative error is within 10%. We also find that the mixture matrix is beneficial to investigate the flow mechanism of gas–liquid two-phase flow.     

Experimental study on the effects of big particles physical characteristics on the hydraulic transport inside a horizontal pipe

This paper presents an experimental study of the physical characteristic effects of large particles on hydraulic transport in a horizontal pipe. The particles are spherical and are large with respect to the diameter of the pipe (8%, 10%, 16%and 25%). Experiments were done to test the important parameters in solid transport (pressure, velocity, etc.). As a result, the relationship between the pressure gradient forces and the mixture velocity was sub-stantially different from the pure liquid flow. However, in a single-phase flow a monotonous behavior of the pres-sure drop curve is observed, and the curve of the solid particle flow attains its minimum at the critical velocity. The regimes are characterized with differential pressure measurements and visualizations.     

Hydrodynamics and bubble behaviour in a three-phase two-stage internal loop airlift reactor

Local hydrodynamics of a gas–liquid–solid system,such as bubble circulation regime,gas holdup,liquid velocity and axial profile of solid concentration,are studied in a two-stage internal loop airlift reactor.Empirical correlations for gas holdup and liquid velocity are proposed to ease the reactor design and scale-up.Different bubble circulation regimes were displayed in the first(lower) and second(upper) stages.Increasing superficial gas velocity and solid loading can promote regime transition of the second stage,and the gas holdup of the second stage is higher than that of the lower stage.In addition,the effects of solid loading on bubble behaviour are experimentally investigated for each stage.It is found that bubble size in the downcomer decreases with the presence of solid particles,and bubble size distribution widens under higher superficial gas velocity and lower solid loading.     

垂直管内通气空泡掺气过程仿真分析简

A semi-empirical gas entrainment model was proposed for the ventilated cavity in vertical pipe, based on which, a complete numerical scheme was established by coupling with the Eulerian-Eulerian two-fluid model to predict the multiscale flow field created by ventilated cavity. Model predictions were validated against experimental measurements on void fraction and bubble size distributions. Simulations were carried out to explore the effect of ventilation rate and inlet turbulence intensity on the macroscale cavity shape and the bubbly flow downstream of the ventilated cavity. As the ventilation rate increasing, a reverse trend was observed for the void fraction and bub- ble size distributions. It is concluded that the average void fraction in the pipe flow region is determined by the volumetric ratio between liquid and gas. However, the bubble size evolution is dominated by the breakage effect induced by turbulence in the vortex region. Furthermore, simulations were conducted to analyze geometric scale effect based upon Froude similitude. The results imply that the velocity distributions were properly scaled. Slight scale effect was seen for the void fraction caused by faster dispersion of bubbles in the larger size model. The com- paratively greater bubble size was predicted in the smaller model, implying significant scale effects in terms of tur- bulence and surface tension effect. It reveals that empirical correlations valid in wide range are required for the ex- trapolation from small-size laboratory models.     

基于复杂度特征的气液二相流流型识别方法

针对气液二相流压差波动信号的非平稳和非线性特征,提出了一种基于经验模式分解(EMD)复杂度特征和支持向量机的流型识别方法。该方法首先对二相流压差波动信号进行经验模式分解,将其分解为若干个固有模态函数(IMF),然后对每一个IMF分量提取复杂度特征作为流型特征向量,并以此作为输入参数建立支持向量机分类器来识别流型。对水平管内空气-水二相流的实验结果表明,文中提出的方法能准确地识别流型,从而为流型识别提供了一种新的有效方法。     

基于灰度共生矩阵和支持向量机的气液两相流流型识别

根据灰度共生矩阵具有较好的纹理表达能力的特性，提出了一种基于图像灰度共生矩阵和支持向量机相结合的气液两相流流型识别的新方法。该方法利用高速摄影系统获取水平管道内气液两相流的流动图像，经过图像处理后，提取图像灰度共生矩阵的纹理特征，进而建立流型图像的灰度共生矩阵纹理特征向量，并以此特征向量作为流型样本对支持向量机进行训练，实现了对流动图像的流型智能化识别。实验结果表明，支持向量机能够快速准确地识别水平管道内的7种典型流型，整体识别率达到100%，每幅流型图像的判别时间约为1.7 s，为流型在线识别提供一种新方法。     

基于神经网络的两相流流型识别方法研究

采集了水平管内气水两相流动的差压信号,利用概率密度函数(PDF)对差压信号特征进行了分析;定义了PDF的四个特征参数,即PDF波峰个数K1、波峰峰值K2、波峰位置K3、以及PDF的方差K4来反映流型的特征。运用四个参数构成的特征向量对径向基函数(RBF)神经网络进行训练并识别流型,结果表明,该方法具有识别速度快、准确率高的特点,从而为两相流的流型识别提供了一种有效的手段。     

基于距离评估的气液二相流流型识别方法

为了克服气液二相流特征融合后不相关特征过多的问题,提出了基于距离评估和支持向量机(SVM)的气液二相流流型识别方法。首先利用经验模式分解和小波包方法对原始的压差波动信号进行分解,分别提取原始信号和各分解信号的时域特征参数组成融合特征,然后采用距离评估方法对融合特征进行评估,根据距离评估因子的大小挑选出敏感特征作为SVM的输入,进而实现对流型的自动识别。水平管内空气-水二相流流型识别结果表明:该方法能够准确获取流型的敏感特征,减小运算规模,提高识别准确率。     

基于图像不变矩特征的气液二相流流型识别

气液二相流流型极大地影响气液二相流的流动和传热特性,准确识别流型对相关设备的设计和运行具有重要意义。根据不变矩能有效检测出具有平移、旋转、比例变化的图像特性,提出了一种基于图像不变矩和概率神经网络相结合的气液二相流流型识别的新方法。该方法利用高速摄影系统获取水平管道内气液二相流的流动图像,经过图像处理后提取图像不变矩特征向量,并以此特征向量作为流型样本对概率神经网络进行训练,实现了对流动图像的流型智能化识别。实验结果表明,训练成功的概率神经网络能够快速准确地识别水平管道内的7种典型流型,整体识别率达到99.3%,为流型在线识别提供一种新的有效方法。     

粉煤密相气力输送流型

基于电容层析成像技术对粉煤密相气力输送系统的流型进行了研究,获得了水平管和竖直上升管粉煤密相气力输送系统的典型流动形态。研究结果表明,水平管输送时的流型随时间复杂多变,典型流型有满管流、沉积层流、悬浮流等;统计分析发现,随表观气速不同,存在明显的主导流型;结合固相速度及管道压力信号分析,展示出其与流型之间存在一定的对应关系,进一步证实了密相气力输送系统的不稳定性特征。对竖直上升管的ECT测试结果表明,输送流型为环核结构。     

高含气率气液两相流差压信号时频特征分析

为了更有效地揭示高含气率气液两相流流动特征,研究了一种新的时频特征分析方法。首先,对差压波动信号进行小波包变换并由变换系数计算信号能量的时频分布;然后,应用统计方法对时频分布进行特征提取得到一组时频特征量。应用类可分性测量准则分析该组特征量区分不同流型的效果,并与以往小波包特征分析方法相比较,结果表明:该组特征量具有更强的流型特征表征能力。最后,以该组特征量为输入向量,构建了集成多类支持向量机分类器实现了流型识别,其流型正确识别率可达97%。     

多尺度信息熵特征的气液二相流流型识别方法

为研究垂直上升管中的气液二相流的流型,利用自制的多电导探针的测量系统采集了4种典型流型的电导波动信息。根据小波包变换能将电导波动信号按任意时频分辨率分解到不同频段的特性,对其进行了3层小波包分解后并计算了各个频段的信息熵特征向量,并作为特征参数输入到E lm an神经网络进行训练,实现了与神经网络相结合流型的智能识别。研究结果表明,该方法能够很准确地识别出4种流型,且提取特征比较方便,从而为流型识别提供了一种新的有效方法。     

Passive acoustic identification of bubble flow regime based on synchrosqueezing wavelet transform and deep learning

A passive acoustic method based on synchrosqueezing wavelet transform (SWT) and deep learning was proposed to automatically identify bubble flow regimes in process industries. The method was established on the bijection relationship between bubble flow regime and the time-frequency (TF) textures of its passive acoustic emission (PAE) signals. Specifically, the PAE signal of the bubble flow was first acquired by hydrophones, then converted into TF representations (TFRs) by SWT and finally used to train convolutional neural network (CNN) to identify bubble flow regimes automatically. The effects of TF analysis method and CNN architecture were studied: the SWT represented TF textures of PAE signals the best, and its combination with ResNet gave the highest identification accuracy of 99.67%. As its main contribution, the method validated the feasibility of PAE and CNN in bubble flow regime identification, and achieved automatic identification of bubble flow regime with high accuracy and barely no subjective interference.