基于SVM及电容层析成像的两相流流型识别

两相流测量中,流型的准确识别是流动参数准确测量的基础.电容层析成像(ECT)技术是自20世纪80年代发展起来的新型检测技术,可用于两相流/多相流流型识别及固相浓度测量.支持向量机是一种基于统计学习理论的机器学习算法,即使在小样本情况下也能得到很好的分类效果.应用ECT系统测量的包含流型信息的电容测量数据,采用支持向量机算法进行流型识别,对4种典型空气-油两相流流型识别分别进行了仿真和静态实验.结果表明,该方法辨识速度快,可准确地识别典型的流型.     

用电阻层析成像技术实现两相流流型识别

介绍了应用天津大学开发的TERT-Ⅱ系统样机,在不同实验装置上进行气/液两相流和油/水两相流的实验研究.提出了应用ERT系统进行流型识别的基本方法;并通过重建图像和分析测量数据实现了两相流流型的识别;并证明了应用ERT系统实现两相流流型和参数测量的可能性.     

基于学习矢量量化神经网络的气固两相流流型识别

管流截面流型作为描述气固两相流的重要参数之一，极大地影响着两相流动压力损失和传热传质等特性，同时还影响着其他参数(如流量、分相含率等)的准确测量以及流动系统的运行特性。传统的检测方法由于难以获得能真正反映流型的管道截面局部分布的实时信息，在工业应用中受到了限制。有鉴于此，在光学层析成像技术的基础上，提出了一种基于学习矢量量化神经网络的气固两相流流型识别方法，详细介绍了这种网络的结构、学习算法、训练样本集的确定等。通过计算机仿真，实验结果表明此方法对于气固两相流的8种流型能有较好的识别能力，为两相流参数检测提供了一种新的思路与方法。     

基于神经网络的水平管中气/液两相流流型识别

以ERT技术为基础,采用神经网络中的感知器和线性神经网络对ERT系统的测量数据进行特征提取和分析,初步实现对几种典型水平管流的识别.     

基于自组织神经网络的油气两相流管截面流型辨识

为了探索两相流流型辨识方法,本文在12电极电容检测系统的基础上,针对油气两相流,提出了基于神经网络的流型辨识方法,即采用自组织神经网络根据电容采样数据实现相应管截面流型的识别.研究表明,该方法是有效的.     

气液两相流型在线识别系统的开发和应用

在统计理论和小波理论的基础上,把神经网络应用到流型识别中,采用VB语言编程完成了软件的开发.按照信号测量、信号特征提取和流型识别3个实现流型客观自动识别的步骤,建立了一套气液两相流流型在线识别系统,在气液两相流实验系统上采集了180组仿真样本进行仿真实验.结果显示:流型正确识别率为91.7%,此系统应用于流型识别,不仅具有较高的识别率,而且很好地实现了识别结果的实时显示,运行速度快,达到了流型在线识别的目的.     

水平管道气液两相流的特征提取与流型识别

流型识别在工业工程中有着极其重要的意义，一直是两相流研究中的重要课题。本文以水平管道中的气／液两相流为研究对象．应用电阻屡析成像系统获取不同流型条件下的测量数据，经预处理后提取小波尺度能量百分比与突起段时间作为特征值．并采用统计学习理论中支持向量机（SVM）方法对特征向量进行分析，识别流型。实验数据的分析结果表明．该方法具有较好的识别率。     

一种基于经验模式分解的气液两相流流型识别方法

提出了一种基于经验模式分解的气液两相流流型识别方法.该方法首先对压差波动信号进行经验模式分解,将其分解为多个平稳的固有模式函数之和,再选取若干个包含主要流型信息的IMF分量,并从中提取时域特征指标-峭度系数作为LVQ神经网络的输入参数,从而实现流型的智能识别.对水平管内空气-水两相流流型的识别结果表明:以EMD为预处理器提取峭度系数的LVQ网络识别方法具有更高的识别率,可以准确、有效地识别流型.     

基于DNA遗传技术的气液两相流型辨识研究

利用DNA分子的复制、储存和传递信息的功能,采用基于DNA编码技术的遗传算法优化BP神经网络,建立了具有更高辨识能力的气液两相流流型辨识模型,并就若干参数对辨识模型计算效率的影响规律进行了研究。研究结果表明,基于DNA编码遗传技术对BP神经网络进行优化处理而形成的气液两相流型辨识模型比已有其它辨识模型具有更高的辨识能力;通过轴承腔气液两相流动数值仿真流型模式的辨识分析,表明这一辨识模型可应用于航空发动机轴承腔气液两相流的研究。     

基于遗传算法/神经网络组合技术的气液两相流型辨识

通过采用遗传算法训练BP神经网络、优化网络权值的技术,对气液两相流的流型进行了辨识研究,在此基础上建立了基于遗传算法/神经网络组合技术的气液两相流流型的预测模型,从而为发动机轴承腔内润滑油气液两相流流型识别提供了技术支持,也为考虑轴承腔气液两相流的相关设计和实验工作提供了技术条件。     

Theoretical study of vertical slug flow measurement by data fusion from electromagnetic flowmeter and electrical resistance tomography

Electrical resistance tomography (ERT) can be used to obtain the conductivity distribution or the phase distribution of gas/liquid flows (e.g. slug flow). Using proper parameter models and flow regime identification models, the measurement of phase size, void fraction, and pattern recognition can be realized. Electromagnetic flowmeters have been used to measure conductive single-phase liquid flows. However, neither ERT nor electromagnetic flowmeters (EMF) can provide accurate measurement of gas/liquid two-phase flows. This paper presents an approach to fuse the information from ERT and an electromagnetic flowmeter. A model for the measurement signal from the electromagnetic flowmeter has been developed based on the flow pattern and the phase distributions, which are obtained from the reconstructed images of ERT, aiming to reduce the measurement error of the electromagnetic flowmeter and enhance the measurement accuracy. Through the simulation research of virtual current density distribution, the feasibility of fusion of electromagnetic flowmeter and ERT to measure gas/liquid two-phase vertical slug flow is verified. By theoretical analysis, the relationship between the output of electromagnetic flowmeter and flow parameters is established. The electrical potential difference of the electromagnetic flowmeter, average velocity, volume flow rate and gas void fraction between the bubble size and location are also investigated. The fusion approach can be used to measure vertical slug flows.     

Application of dual-plane ERT system and cross-correlation technique to measure gas–liquid flows in vertical upward pipe

The progress of process tomography provides a new method for two-phase flow measurement. The dual-plane electrical resistance tomography (ERT) is combined with the correlation measurement technique to carry out the two-phase flow measurement in which the continuous phase is conductive liquid. The method of the estimation of void fraction and the disperse phase velocity by extracting the eigenvalue of the dual-plane ERT boundary measured data is presented. This method is applied to the transient flow-rate measurement of the air–water two-phase flow in vertical pipe. The information of disperse phase void fraction and distribution variation with time change can be considered adequately, and the estimated value of disperse phase void fraction and velocity can be gained fairly accurately in this method, which provides the data for the calculation of the transient flow-rate. The experiment results indicate that this kind of measurement method is valid when the distance between the upstream and downstream measured cross section is short enough.     

Independent component analysis of interface fluctuation of gas/liquid two-phase flows — experimental study

It is important to understand the behaviour of two-phase flows in industry. This paper presents a study of the interface fluctuation between gas/liquid two-phase flows in a horizontal pipeline. Having obtained the data of a gas/liquid flow by electrical resistance tomography (ERT), an independent component analysis (ICA) method can be applied not only to extract the flow regime information but also the interface fluctuation of the flow. The efficiency of ICA with the ERT data has been assessed by experiment. The independent components have been interpreted by comparing the obtained independent components with the reconstructed images by ERT, showing that ICA is not only effective in extracting flow regime information, but also provides the fluctuation of each individual phase and the interface between the two phases. Without modelling the forward problem, this method can be applied to other electrical tomography modalities.     

Application of electrical resistance tomography to two-phase pipe flow parameters measurement

This paper introduces the TERT-IV prototype developed by Tianjin University. The application of the TERT-IV system to measurement parameters of two-phase flow has been studied. The methods of analyzing measured data of ERT system are presented and applied to identify flow regimes and estimate void fraction. For the several typical flow regimes, the methods of principal component analysis and artificial neural network to identify the two-phase flow regimes is presented, and that is proved to have higher recognition rate by experimental test. For the different phase distribution on a pipe cross-section, the methods of relative changes summation and polynomial regression are used to estimate void fraction, and are proved to be possible by comparing the results of simulation calculation to the analytic results of experimental measured data.The research results show that the method is feasible using feature extraction and analysis data to measure the parameters of two-phase flow under the different flow conditions, and prove that it is possible to monitor on-line the transportation process of air/water two-phase flow using ERT system.     

Identification of gas/liquid two-phase flow regime through ERT-based measurement and feature extraction

Gas/liquid two-phase flow is of great importance in various industrial processes. As the most important characteristic of a two-phase flow, the flow regime not only characterizes the flow condition in an explicit way, but also determines the measurement model in each measuring method. Based on the application of Electrical Resistance Tomography (ERT) to a gas/liquid two-phase flow on a vertical pipe, features reflecting the characteristics of gas/liquid two-phase flow are extracted directly from the measured data without reconstruction of the cross-sectional images. The statistical features are derived through time series statistical analysis. Meanwhile features in the wavelet-scale domain are derived through both one-dimensional and two-dimensional wavelet transform. All extracted features are considered as the input of a Support Vector Machine (SVM) algorithm to recognize the flow regime. The preliminary results show that the feature extraction methods of multi-feature fusion and high-dimensional wavelet transform are suitable for the identification of gas/liquid two-phase flow regimes.     

Data fusion in dual-mode tomography for imaging oil–gas two-phase flow

Process tomography (PT) techniques have been developed rapidly for visualizing the internal behavior of industrial processes, e.g. multi-phase flow measurement. Most of tomography systems employ a single measurement technique, such as computerized tomography (CT), optical tomography (OT), electrical resistance tomography (ERT) or electrical capacitance tomography (ECT). It is now possible to fit two or more tomographic systems to an industrial process. Detailed information from different modalities can be gained by inspection of separate tomographs, and the advantage of the strongest features provided by each unit can be taken. A combined tomogram can be produced of superior quality to any of the separate tomograms. To maximize the information available from the combined tomographic system, data fusion is the better option. In this paper, a dual-mode tomography system based on capacitance sensor and gamma sensor was developed to capture oil–gas two-phase flow. The two modalities can work at the same time. Two fusion methods, namely image fusion method and data fusion method, are proposed. Both simulation and static experiments for oil–gas two-phase flow were conducted. The reconstruction results of different fusion methods and modalities were compared and discussed.     

基于神经网络的航空发动机轴承腔气液两相流流型辨识方法研究

基于航空发动机轴承腔润滑中所存在的气液两相流问题，采用基于神经网络的理论方法建立预测模型，以便进行轴承腔内气液两相流流型的识别。研究以管道气液两相流为原型，采用3种典型的神经网络对流型进行模式识别，通过考察3种网络的辨识率，发现BP网络的识别方法具有较高的准确性。     

Measurement of vertical oil-in-water two-phase flow using dual-modality ERT–EMF system

Oil-in-water two-phase flows are often encountered in the upstream petroleum industry. The measurement of phase flow rates is of particular importance for managing oil production and water disposal and/or water reinjection. The complexity of oil-in-water flow structures creates a challenge to flow measurement. This paper proposes a new method of two-phase flow metering, which is based on the use of dual-modality system and multidimensional data fusion. The Electrical Resistance Tomography system (ERT) is used in combination with a commercial off-the-shelf Electromagnetic Flow meter (EMF) to measure the volumetric flow rate of each constituent phase. The water flow rate is determined from the EMF with an input of the mean oil-fraction measured by the ERT. The dispersed oil-phase flow rate is determined from the mean oil-fraction and the mean oil velocity measured by the ERT cross-correlation velocity profiling. Experiments were carried out on a vertical upward oil-in-water pipe flow, 50 mm inner-diameter test section, at different total liquid flow rates covering the range of 8–16 m³/hr. The oil and water flow rate measurements obtained from the ERT and the EMF are compared to their respective references. The accuracy of these measurements is discussed and the capability of the measurement system is assessed.