A method for measuring bubble cluster characteristics in gas-liquid two-phase flow based on laser cross-section imaging
Through the laser cross-sectional imaging method, the measurement error problem caused by bubble adhesion and overlap in the gas-liquid two-phase flow is solved, and the accurate measurement of the characteristic parameters of the multi-dimensional bubble group is achieved, which is suitable for the safety analysis of nuclear power plant cores.
Patent Information
- Application Number
- CN202210281790.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-21
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2042-03-21
AI Technical Summary
In the prior art, when measuring the characteristic parameters of bubble groups in the gas-liquid two-phase flow, there are measurement errors and inaccuracy caused by bubble adhesion and overlap, making it difficult to achieve multi-dimensional synchronous measurement.
Using laser cross-section imaging method, by emitting laser beams to transparent containers with two-phase flow of gas and liquid, obtaining bubble cross-section images, and classifying and segmenting the image, calculating the gas content rate and bubble position distribution of liquid cross-sections, and extracting bubbles vertical motion speed and particle size.
It improves the accuracy and comprehensiveness of bubble group characteristics measurement, and can more effectively calculate the distribution and motion characteristics of bubbles in cross-section, which is suitable for the analysis of pressure distribution and heat transfer characteristics of nuclear power plant cores, providing guarantees for the safe operation of reactors.
Smart Images

Figure CN114594099B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of gas-liquid two-phase flow measurement, and in particular to a method for measuring characteristics of a gas-liquid two-phase flow bubble group based on laser cross-section imaging. Background Art
[0002] In a pressurized water reactor in a nuclear power plant, subcooled boiling is generally allowed to occur when the coolant flows through the fuel rod bundle channel, thereby greatly increasing the core outlet temperature and heat transfer coefficient. However, this also causes a large number of bubbles to be generated in the reactor, forming a gas-liquid two-phase flow. Since the heat transfer coefficient of the bubbles is extremely low, in order not to affect the heat transfer effect of the rod bundle heating surface, the gas content of the liquid phase cross section of the two-phase flow must be strictly controlled. Otherwise, once film boiling occurs, the surface of the fuel rod will be covered by a gas film for a long time, resulting in a strong pressure change in the core, which will then cause high temperature, vibration, noise, fuel rod erosion, and even lead to rod breakage and core melting accidents.
[0003] For the measurement of bubble parameters in gas-liquid two-phase flow, a Chinese patent with publication number CN105092684B discloses "A device for measuring the volume concentration of microbubbles", which includes a sampling tube, an ultrasonic device, a peristaltic pump, a conductivity measurement system, a measuring tube and analysis software; wherein one end of the sampling tube is placed at a sampling point of the gas-liquid two-phase flow, and the other end is connected to the measuring tube, a part of the sampling pipeline is placed in the ultrasonic device, a conductivity probe of the conductivity measurement system is inserted from the wall of the measuring tube to the center of the tube, one end of the measuring tube is connected to the peristaltic pump, and the analysis software estimates the volume concentration of microbubbles at the sampling point based on the voltage difference generated when the gas and liquid contact the conductivity probe.
[0004] The bubble volume concentration measuring device in the above-mentioned existing scheme measures parameters through the "probe method". The "probe method" refers to obtaining the size and speed parameters of the bubble by using the change in conductivity or reflected light intensity when the bubble contacts the probe. However, a single probe cannot obtain the distribution information of the bubble on the cross section in real time, and the introduction of multiple probes is likely to disrupt the inherent movement law of the bubbles inside the system. Therefore, the "probe method" does not have the ability to synchronously detect the characteristic parameters of the bubble group in multiple dimensions, that is, the comprehensiveness of the measurement of the characteristic parameters of the bubble group is not good. In addition, the existing image analysis method is a non-contact measurement method, which uses a camera placed on the side of the flow channel to shoot and obtain the movement characteristics of the bubble at high speed, and then obtain instantaneous full-field flow information. However, when the number of bubbles is large, the imaging method will produce serious measurement errors due to the adhesion and overlap of the bubble group image, so that it can generally only analyze the bubble sequence composed of several bubbles in the vertical direction, so there is a problem of low measurement accuracy. Therefore, how to design a method that can take into account the comprehensiveness and accuracy of the measurement of the characteristic parameters of the bubble group is a technical problem that needs to be solved urgently. Summary of the invention
[0005] In view of the above-mentioned deficiencies in the prior art, the technical problem to be solved by the present invention is: how to provide a method for measuring the characteristics of a gas-liquid two-phase flow bubble group based on laser cross-sectional imaging, so as to solve the cross-sectional calculation problem caused by bubble adhesion and overlap, and to realize the measurement of multiple and multi-dimensional bubble group characteristic parameters, thereby improving the comprehensiveness and accuracy of the measurement of the bubble group characteristic parameters.
[0006] In order to solve the above technical problems, the present invention adopts the following technical solutions:
[0007] The method for measuring characteristics of bubble groups in gas-liquid two-phase flow based on laser cross-section imaging includes the following steps:
[0008] S1: emitting a laser beam toward a transparent container where a gas-liquid two-phase flow occurs, so as to form a laser detection cross section in the transparent container along the irradiation direction of the laser beam;
[0009] S2: Acquire the bubble cross-sectional image on the laser detection section;
[0010] S3: classifying bubbles on the bubble cross-section image to generate regular bubble images of single bubbles and irregular bubble images of adhered or overlapping bubbles;
[0011] S4: segmenting the adhered or overlapping bubbles in the irregular bubble image;
[0012] S5: Extract characteristic parameters based on the regular bubble image and the segmented irregular bubble image, and then calculate the corresponding liquid phase cross-section gas content and bubble position distribution.
[0013] Preferably, in step S1, a laser beam is emitted toward the transparent container at an inclination angle of 30° to 45° on a horizontal plane.
[0014] Preferably, in step S2, the bubble cross-sectional image is acquired at an angle perpendicular to the laser detection cross-sectional area.
[0015] Preferably, in step S2, the bubble cross-section image is subjected to image preprocessing, and the image preprocessing includes but is not limited to lens distortion correction, pixel scale calculation, interference spot elimination, binarization processing and noise suppression.
[0016] Preferably, the image preprocessing specifically includes the following steps:
[0017] S201: placing a checkerboard calibration plate at a position corresponding to the laser detection section to obtain a corresponding calibration image, and then performing lens distortion correction and pixel scale calculation on the bubble cross-section image by a polynomial coordinate transformation method;
[0018] S202: collecting a background image of the laser detection cross section in a bubble-free state, and then differentially eliminating interference spots in the bubble cross section image;
[0019] S203: converting the bubble cross-section image into a bubble cross-section grayscale image, and binarizing the bubble cross-section grayscale image by using a maximum inter-class variance method to obtain a cross-section bubble contour image;
[0020] S204: Suppressing scattered noise and random noise in the cross-sectional bubble contour image by median filtering and Gaussian filtering respectively.
[0021] Preferably, in step S3, the roundness value R of each bubble in the bubble cross-sectional image is calculated based on the connected domain feature. o ; Then based on the roundness value R of the bubble o The bubbles are divided into single bubbles and adhesion or overlapping bubbles; finally, a regular bubble image of the single bubble and an irregular bubble image of the adhesion or overlapping bubbles are generated;
[0022]
[0023] Where: L represents the perimeter of the connected domain of the bubble cross section; s represents the area of the connected domain of the bubble cross section; R o <0.98, the bubble is a single bubble; R o ≥0.98, the bubbles are adhesion or overlapping bubbles.
[0024] Preferably, step S4 specifically includes the following steps:
[0025] S401: Calculate the radial radius [R] of each edge pixel in the irregular bubble image based on radial symmetry transformation min ,R max ], and then convert the irregular bubble image into a local radially symmetric image that highlights the image gradient, and finally perform threshold segmentation to obtain the corresponding binary image;
[0026] S402: Calculate the centroid coordinates of the connected domain of the binary graph as the seed point; then calculate the correlation between each edge pixel and the seed point by using two indices, distance metric and gradient metric, to obtain the edge pixel set most correlated with each seed point;
[0027] S403: performing ellipse fitting on the edge pixel set of each seed point, inferring the missing part of the edge pixel set of the adhesion or overlapping object, so as to achieve the segmentation of the adhesion or overlapping bubbles;
[0028] S404: Determine the radial radius [R min ,R max ] is the optimal radial radius. If so, the segmented irregular bubble image is output; otherwise, the radial radius [R min ,R max ] and returns to step S401.
[0029] Preferably, in step S404, the radial radius [R min ,R max ] The condition for whether it is the optimal radial radius is that the overlap between the two output fitting ellipses is less than 0.7, and the ratio of the total area of the ellipse after removing the overlap to the area of the original undivided connected domain is greater than 0.95 and less than 1.05.
[0030] Preferably, in step S5, the corresponding liquid phase cross-section gas content is calculated by the following formula:
[0031] α=∑a n / A;
[0032] Where: α represents the gas content of the liquid phase cross section; a n represents the pixel area of each bubble in the laser detection cross section; A represents the total pixel area of the laser detection cross section.
[0033] Preferably, in step S5, the number of bubbles, bubble pixel area and bubble position coordinates of the regular bubble image and the segmented irregular bubble image are extracted by an ellipse fitting algorithm.
[0034] Compared with the prior art, the gas-liquid two-phase flow bubble characteristic measurement method of the present invention has the following beneficial effects:
[0035] The present invention emits a laser beam to a transparent container where a gas-liquid two-phase flow occurs, so as to form a laser detection cross section in the transparent container, so that when the bubble passes through the laser detection cross section, it can induce strong light scattering to form a bubble cross-sectional image. On the one hand, the present invention generates a regular bubble image of a single bubble and an irregular bubble image of a sticky or overlapping bubble by introducing a laser detection cross section and classifying the bubble cross-sectional image, and then segmenting the sticky or overlapping bubbles in the irregular bubble image, which can effectively solve the cross-sectional calculation problem caused by bubble adhesion and overlap, thereby improving the accuracy of bubble group feature measurement. On the other hand, the present invention can extract feature parameters based on the regular bubble image and the segmented irregular bubble image, and then calculate (measure) the gas content of the liquid phase cross section and the bubble position distribution, and can further measure the vertical movement speed and particle size of the bubble, and can realize the measurement of multiple and multi-dimensional bubble group feature parameters. Compared with the probe method, it will not disturb the internal flow field of the gas-liquid two-phase flow, and can obtain higher-dimensional bubble group features, thereby improving the comprehensiveness of bubble group feature measurement, and can be applied to measure the pressure distribution and heat transfer characteristics during the operation of the core, providing a strong safety guarantee for the reliable operation of the subsequent reactor. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] In order to make the purpose, technical solution and advantages of the invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, in which:
[0037] Figure 1 is a schematic diagram of a bubble image acquisition device;
[0038] Figure 2 It is a logic block diagram of the method for measuring the characteristics of bubble groups in gas-liquid two-phase flow based on laser cross-section imaging;
[0039] Figure 3 This is a flow chart of a method for measuring characteristics of bubble groups in gas-liquid two-phase flow based on laser cross-section imaging;
[0040] Figure 4 (a) is the shooting method of the calibration image, and (b) is the acquired calibration image;
[0041] Figure 5 (a) is the collected bubble cross-section image, (b) is the image after binarization processing, and (c) is the image after filtering;
[0042] Figure 6 (a) is the bubble cross-section image before classification, (b) is the regular bubble image, and (c) is the irregular bubble image;
[0043] Figure 7 A flow chart for segmenting adhered or overlapping bubbles in irregular bubble images;
[0044] Figure 8 This is a schematic diagram of the results of segmenting adhesion or overlapping bubbles;
[0045] Fig. 9 (a) is the instantaneous bubble number statistics, (b) is the instantaneous gas content curve;
[0046] Fig.10 (a) is the curve of gas content changing with mass gas content, (b) is the error analysis between the measured value and theoretical value of gas content in the cross section;
[0047] Fig.11 Thermal map for imaging the mean gas content in the cross section;
[0048] Fig.12 (a) is the target bubble tracking image, (b) is the bubble ellipsoid reconstruction;
[0049] Fig.13 is the histogram of the bubble group size distribution. DETAILED DESCRIPTION
[0050] The following is a further detailed description through specific implementation methods:
[0051] Example:
[0052] This embodiment discloses a method for measuring characteristics of a gas-liquid two-phase flow bubble group based on laser cross-sectional imaging.
[0053] This embodiment discloses Figure 1 The bubble image acquisition device shown in the figure is explained with the core model at normal temperature and pressure as the experimental system. The bubble image acquisition device consists of a core simulation pool (transparent container), a bubble generator, a bubble imaging laser (emitting laser beam), a high-speed camera (collecting bubble cross-sectional images), and a conductivity probe method bubble velocity measurement device. The core simulation pool is a 40mm×40mm×500mm transparent rectangular structure. Water is filled into the pool and 9 hollow transparent sealed tube bundles with an outer diameter of 9.5mm are inserted to form a 3×3 simulated fuel rod array. Acrylic materials are used for the pool body and tube bundles, and quartz materials can also be used as a substitute under high-temperature measurement conditions. In order to reduce the optical distortion introduced by the different refractive indices of the tube bundle material and water, the tube body adopts a 0.5mm thin-wall design and water is filled into the tube. The bubble generator consists of a peristaltic pump and a microporous ceramic block to produce an upward bubble group with a precisely controllable gas flow rate.
[0054] The bubble imaging laser is a 40mW, 650nm laser, combined with a Powell prism array to output a straight fan beam with a thickness of about 0.7mm, which illuminates the middle section of the core simulation pool at a horizontal angle of 30° (here, the purpose of the laser's tilted illumination is to make the detection section perpendicular to the camera, thereby reducing image distortion) to form a laser detection section. When a group of bubbles passes through the laser detection section, it will cause strong light scattering and form an image signal. At this time, a Hikvision MV-CA004-10UM high-speed camera is placed on the opposite side of the laser, perpendicular to the detection section, and a bubble morphology image sequence, i.e., a bubble cross-section image, is collected in a darkroom environment at a frame rate of 800FPS, and the characteristic parameters of the bubble group are further calculated.
[0055] like Figure 2 and Figure 3 As shown, the method for measuring characteristics of bubble groups in gas-liquid two-phase flow based on laser cross-section imaging includes:
[0056] S1: emit a laser beam toward a transparent container where gas-liquid two-phase flow occurs, so as to form a laser detection cross section in the transparent container along the irradiation direction of the laser beam; in this embodiment, the laser beam is emitted toward the transparent container at an inclination angle of 30° to 45° on the horizontal plane. The purpose of the inclined irradiation of the laser is to make the detection cross section perpendicular to the camera, thereby reducing image distortion.
[0057] S2: Acquire a bubble cross-sectional image on the laser detection cross-sectional area; in this embodiment, the bubble cross-sectional image is acquired at an angle perpendicular to the laser detection cross-sectional area.
[0058] S3: classifying bubbles on the bubble cross-section image to generate regular bubble images of single bubbles and irregular bubble images of adhered or overlapping bubbles;
[0059] S4: segmenting the adhered or overlapping bubbles in the irregular bubble image;
[0060] S5: Extract characteristic parameters based on the regular bubble image and the segmented irregular bubble image, and then calculate the corresponding liquid phase cross-section gas content and bubble position distribution.
[0061] In the specific implementation process, the characteristic parameters such as the number of bubbles, bubble pixel area, and bubble coordinate position in the regular bubble image and the segmented irregular bubble image are extracted by the ellipse fitting algorithm. Among them, for the regular bubble image, the least squares ellipse fitting method can be directly used to extract the contour, and then the ellipse fitting algorithm is used to realize the characteristic analysis and statistics of each bubble.
[0062] It should be noted that the ellipse fitting algorithm is an existing conventional method, and its basic idea is: for a set of sample points on a given plane, find an ellipse that is as close to these sample points as possible. In other words, a set of data in the image is fitted with an ellipse equation as a model, so that a certain ellipse equation can satisfy these data as much as possible, and the various parameters of the ellipse equation are calculated. The center of the best ellipse finally determined is the bull's eye to be searched.
[0063] The corresponding liquid phase cross-section gas content is calculated by the following formula:
[0064] α=∑a n / A;
[0065] Where: α represents the gas content of the liquid phase cross section; a n represents the pixel area of each bubble in the laser detection section; A represents the total pixel area of the laser detection section (or the flow channel of gas-liquid two-phase flow).
[0066] At the same time, the bubble pixel area within a period of time can be integrated according to the flow channel position to measure the bubble position distribution (also a cross-sectional gas content distribution), that is, Fig.11 The thermodynamic map is a two-dimensional distribution map. Compared with the probe measurement system, the dimension of the cross-sectional gas content distribution is higher. However, the ordinary camera cannot obtain the cross-sectional gas content distribution by acquiring the bubble motion image of the flow channel measurement.
[0067] Furthermore, the vertical velocity of the bubble can be calculated by a mathematical model, and then a three-dimensional ellipsoid fitting can be performed based on the velocity and the coordinate information of the bubble cross-sectional profile (e.g. Fig.12 (b) shows that the size of the sphere equivalent to the fitted ellipsoid represents the particle size of the bubble, that is, the present invention can also measure the vertical velocity and particle size of the bubble.
[0068] In general, the method of the present invention can analyze and measure the gas content of the liquid phase cross section, the bubble position distribution, and the vertical movement speed and particle size of the bubbles in the gas-liquid two-phase flow.
[0069] It should be noted that the method for measuring characteristics of bubble groups in gas-liquid two-phase flow based on laser cross-sectional imaging in the present invention can generate corresponding software codes or software services through program programming, and can then be run and implemented on servers and computers.
[0070] The present invention emits a laser beam to a transparent container where a gas-liquid two-phase flow occurs, so as to form a laser detection cross section in the transparent container, so that when the bubble passes through the laser detection cross section, it can induce strong light scattering to form a bubble cross-sectional image. On the one hand, the present invention generates a regular bubble image of a single bubble and an irregular bubble image of a sticky or overlapping bubble by introducing a laser detection cross section and classifying the bubble cross-sectional image, and then segmenting the sticky or overlapping bubbles in the irregular bubble image, which can effectively solve the cross-sectional calculation problem caused by bubble adhesion and overlap, thereby improving the accuracy of bubble group feature measurement. On the other hand, the present invention can extract feature parameters based on the regular bubble image and the segmented irregular bubble image, and then calculate (measure) the gas content of the liquid phase cross section and the bubble position distribution, and can further measure the vertical movement speed and particle size of the bubble, and can realize the measurement of multiple and multi-dimensional bubble group feature parameters. Compared with the probe method, it will not disturb the internal flow field of the gas-liquid two-phase flow, and can obtain higher-dimensional bubble group features, thereby improving the comprehensiveness of bubble group feature measurement, and can be applied to measure the pressure distribution and heat transfer characteristics during the operation of the core, providing a strong safety guarantee for the reliable operation of the subsequent reactor.
[0071] In the specific implementation process, the bubble cross-section image is preprocessed, and the image preprocessing includes but is not limited to lens distortion correction, pixel scale calculation, interference spot elimination, binarization processing and noise suppression. Specifically including:
[0072] S201: Place the set (5mm×5mm) checkerboard calibration plate at a position that coincides with the laser detection cross section (such as Figure 4 (a) to obtain the corresponding calibration image (as shown in Figure 4 (b)), and then the bubble cross-section image is corrected for lens distortion and pixel scale calculated by polynomial coordinate transformation method;
[0073] In this embodiment, the area of each pixel is about 0.0204 mm 2 Since the acquired image is not perpendicular to the flow channel, the image needs to be transformed into a three-dimensional coordinate system. According to the coordinate transformation rules of three-dimensional homogeneous coordinates, the image coordinate system is rotated 30° to 45° counterclockwise around the X-axis.
[0074] S202: collecting a background image of the laser detection cross section in a bubble-free state, and then differentially eliminating interference spots in the bubble cross section image;
[0075] S203: converting the bubble cross-section image into a bubble cross-section grayscale image, and binarizing the bubble cross-section grayscale image by using a maximum inter-class variance method to obtain a cross-section bubble contour image;
[0076] In this embodiment, since the scattered light of the bubble on the cross section will be secondary scattered when it contacts other bubbles outside the cross section, artifacts will be formed in the image, such as Figure 5 In order to avoid misjudgment in the subsequent algorithm, the maximum inter-class variance method is used to binarize the original grayscale image according to the characteristic that the secondary scattered light intensity is much lower than the primary scattered light intensity. Figure 5 (b) The cross-sectional bubble profile image shown.
[0077] S204: Suppress the scattered noise and random noise in the cross-sectional bubble contour image by 3×3 pixel median filtering and Gaussian filtering, thereby obtaining a preprocessed bubble cross-sectional image. Figure 5 As shown in (c), the contour contrast can be greatly improved. It should be noted that the filtering process may filter out some microbubble information, but due to its small volume, it has little effect on the subsequent calculation of the gas content of the liquid phase section.
[0078] The present invention realizes lens distortion correction and pixel scale calculation of bubble cross-section image by combining a checkerboard calibration plate with a polynomial coordinate transformation method, eliminates interference spots in the bubble cross-section image by differentially collecting a background image of a laser detection cross-section in a bubble-free state, binarizes the bubble cross-section grayscale image by a maximum between-class variance method to eliminate image artifacts, and suppresses scattered noise and random noise in the cross-section bubble contour image by median filtering and Gaussian filtering, respectively, so that image noise can be effectively suppressed and contour contrast can be greatly improved, which can provide an excellent image basis for subsequent bubble classification and segmentation, thereby further improving the accuracy of bubble group feature measurement.
[0079] In the specific implementation process, the roundness value R of each bubble in the bubble cross-section image is calculated based on the connected domain feature. o ; Then based on the roundness value R of the bubble o The bubbles are divided into single bubbles and adhesion or overlapping bubbles; finally, a regular bubble image of the single bubble and an irregular bubble image of the adhesion or overlapping bubbles are generated;
[0080]
[0081] Where: L represents the perimeter of the connected domain of the bubble cross section; s represents the area of the connected domain of the bubble cross section; R o <0.98, the bubble is a single bubble; R o ≥0.98, the bubbles are adhesion or overlapping bubbles.
[0082] In this embodiment, the bubble classification results are as follows: Figure 6 As shown. Among them, Figure 6 (a) shows the original bubble cross-section image before roundness value classification. Figure 6 (b) represents a regular bubble image, Figure 6 (c) shows the irregular bubble image.
[0083] The present invention classifies the roundness values of bubble images through connected domain features, so that single bubbles and adhered or overlapping bubbles can be effectively distinguished and regular bubble images of single bubbles and irregular bubble images of adhered or overlapping bubbles can be generated. Different methods can then be used to respectively analyze the regular bubble images and the irregular bubble images, thereby improving the computational analysis efficiency of bubble group feature measurement.
[0084] In the specific implementation process, Figure 7 As shown, step S4 includes the following steps:
[0085] S401: Calculate the radial radius [R] of each edge pixel in the irregular bubble image based on Fast Radial Symmetry (FRS) min ,R max ], and then convert the irregular bubble image into a local radially symmetric image that highlights the image gradient, and finally perform threshold segmentation to obtain the corresponding binary image;
[0086] S402: Calculate the centroid coordinates of the connected domain of the binary graph as the seed point; then calculate the correlation between each edge pixel and the seed point by using two indices, distance metric and gradient metric, to obtain the edge pixel set most correlated with each seed point;
[0087] S403: performing ellipse fitting on the edge pixel set of each seed point, inferring the missing part of the edge pixel set of the adhesion or overlapping object, so as to achieve the segmentation of the adhesion or overlapping bubbles;
[0088] S404: Determine the radial radius [R min ,R max ] is the optimal radial radius. If so, the segmented irregular bubble image is output; otherwise, the radial radius [R min ,R max ], and returns to step S401. In this embodiment, the radial radius [R min ,R max ] The condition for whether it is the optimal radial radius is that the overlap between the two output fitting ellipses is less than 0.7, and the ratio of the total area of the ellipse after removing the overlap to the area of the original undivided connected domain is greater than 0.95 and less than 1.05.
[0089] Among them, the change of radial radius [R min,R max ]: First, define the equivalent radius r of the smallest bubble and the largest bubble in the target image min and r max As the upper and lower limits of the search; then the ellipse fitting degree is used as the optimization target for cyclic iteration.
[0090] It should be noted that although the computational complexity of this iterative process is large, when continuously segmenting a multi-frame image sequence, since the bubble characteristics change very little over a period of time, the segmentation speed can be greatly improved by storing multiple sets of historical merits and making priority judgments on them. According to statistics, the computational time for segmenting 800 frames of images can be reduced from 35 minutes to about 5 minutes before and after the introduction of historical merit judgments.
[0091] Figure 8 The segmentation results of three bubble overlapping and adhesion images are shown in Figure 2. It can be seen that after optimizing the radial radius value, a better segmentation effect can be obtained.
[0092] The present invention realizes the segmentation of adhered or overlapping bubbles through the fast radial symmetry transformation method, and then completes the estimation of the bubble boundary contour and data expansion in combination with ellipse fitting, so that the adhered or overlapping bubbles in the irregular bubble image can be effectively segmented, and the cross-section calculation problem caused by bubble adhesion and overlap can be effectively solved, thereby further improving the accuracy of bubble group feature measurement.
[0093] In order to better illustrate the advantages of the present invention, this embodiment discloses the following experiments.
[0094] 1. Measurement of instantaneous cross-section gas content
[0095] This experiment is based on Figure 1 The bubble image acquisition device in the embodiment was implemented, and four groups of air intake volumes were set at 40ml / min, 60ml / min, 80ml / min and 100ml / min, and the instantaneous bubble number and gas content in the 50 consecutive bubble image sequences were counted. The results are shown in Fig. 9 The dotted line is the numerical average of the number of bubbles and the gas content.
[0096] from Fig. 9 It can be seen that the number of bubbles and the gas content fraction obtained by the method of the present invention are positively correlated with the air intake volume. However, since the volume and number of bubbles output by the bubble generator are highly random, the characteristic parameters of the bubble group between each frame image fluctuate greatly, while the mean gas content fraction shows an arithmetic variation law under the current low air intake volume condition.
[0097] It should be noted that the air intake and gas content do not satisfy a proportional relationship. Generally, it is necessary to use the mass gas content calculation model related to the flow channel characteristics to describe it. The mass gas content represents the proportion of the gas phase in the total mass of the two-phase fluid flowing through a certain section of the channel per unit time, which is converted from the air intake. Among them, Fig.10 (a) The curve of the 50-frame average gas content measured experimentally under different air inlet conditions versus mass gas content is shown. At the same time, the prediction curves of four mass gas content calculation models, namely, the Milopolsky sliding speed ratio model (from Bankoff S G. Variable density single fluid model for two-phase flow, especially steam-water flow), the Smith mixed phase-single phase model (from Hibiki T, Mao K, Ozaki T. Development of void fraction-quality correlation for two-phase flow in horizontal and vertical tubebundles), the Bankoff variable density model (from Smith S. Void fractions in two-phase flow: acorrelation based upon an equal velocity head model) and the homogeneous model (from Sun HJ, Feng Y. Evaluation of calculation method for cross-section vapor fraction of gas-liquid two-phase flow based on sliding speed ratio model) are plotted for comparison.
[0098] Since only the Milopolsky model considers the flow characteristics when multiple rod bundles are interpolated, the prediction results are in good agreement with the experimental data. However, when the mass gas fraction continues to increase, the two-phase flow pattern can change from bubbly flow to other flow patterns such as slug flow, in which excessive adhesion and overlapping bubble images will increase the measurement error of the cross-sectional gas fraction.
[0099] This experiment mainly considers the measurement of bubbly flow patterns, such as Fig.10 As shown in (b), when the current air intake volume is in the range of 0-180 ml / min, the error between the measured value and the theoretical value is within 5%. When the air flow rate is extended to 300 ml / min, the error is only about 10%. Compared with the traditional direct imaging method, it has a significantly wider gas content range and higher measurement accuracy.
[0100] 2. Cross-sectional gas content mean imaging
[0101] Since the method of the present invention can accurately measure the position and area of the bubble in the cross section, the two-dimensional integration of the bubble group image surface can be performed by multi-frame accumulation to achieve cross-sectional gas content mean imaging. The experiment set the air intake volume to 60ml / min and 100ml / min respectively, and processed 8000 frames of images taken within 10 seconds to obtain the imaging results as shown in the figure. Fig.11 shown.
[0102] like Fig.11 As shown, the position of each fuel rod is clearly displayed, and it can be observed that the sub-channels between the rod bundles have a higher gas content, indicating that the bubble group tends to gather in the center of the sub-channel, showing a typical "center peak distribution" (from Mizutani Y, Tomiyama A, Hosokawa S, et al. Two-phase flow patterns in a four by four rod bundle). In addition, it can be distinguished that the gas content on the right side of the channel is significantly higher than that on the left side. This is because the bubble generator is installed in the gap of the right rod bundle, and the bubbles are blocked by the fuel rods when they rise and are not evenly distributed.
[0103] 3. Bubble 3D reconstruction and average particle size distribution measurement
[0104] Because the image of each bubble in the image sequence is generally connected up and down, the overlapping judgment of the previous and next frames can determine the multiple fault contours of the bubble when it passes through the detection section, such as Fig.12 (a). This process takes a very short time, during which the velocity can be considered to be basically unchanged. Then, when the instantaneous vertical velocity is known, the three-dimensional shape of the bubble can be calculated, and the three-dimensional reconstruction of the bubble group can be achieved as a whole. However, the method of the present invention cannot directly obtain the bubble velocity, and it is considered to solve it indirectly through morphological modeling.
[0105] Since the relative speed of bubbles and fluid is closely related to the three-dimensional shape of bubbles, that is, when the speed is zero, the bubbles are in a perfect circle due to the influence of surface tension; as the speed increases, the fluid resistance increases rapidly, and the bubbles will gradually become ellipsoidal. Large bubbles can also present other complex shapes such as straw hats when moving at high speeds. This process can be accurately described by the VOF (Volume of Fluid) model. Then, based on the constraint relationship of "fault contour-velocity-bubble shape", the vertical velocity of the bubbles can be solved. However, the computational overhead of the VOF model is considerable, and it is difficult to achieve rapid analysis when there are a large number of bubbles. In this experiment, without considering large bubbles (diameter > 10 mm), based on the bubble roundness value R oThe vertical velocity is calculated using the simplified relationship between the dimensionless fluid parameters We (Wellek number) and Re (Reynolds number) (from Ya HJ, Zhao GJ, Liu L, et al. Experimental Research on the Shape and Rising Law of Single Bubble in Still Water), as shown in the following formula. This can effectively reduce the amount of calculation required while meeting the calculation accuracy.
[0106]
[0107] Among them, R o The definition of is the same as the above formula, but the roundness value of the transverse section is changed to the maximum roundness value of the longitudinal section of the bubble. d is the equivalent circle diameter of the bubble ellipse fitting, ρ1 is the liquid density, σ is the surface tension, μ1 is the liquid viscosity, ν T is the bubble velocity. The water temperature in this experiment is 9°C, and the physical properties of water are shown in Table 1.
[0108] Table 1 Liquid phase physical properties
[0109]
[0110] The solution process is Fig.12 Taking the target bubble in (a) as an example, we first define ν T The initial value of the iterative search is 0.05 m / s, and the bubble contour set is obtained by using the tracked 15-frame image sequence, combined with the current ν T Take the value to perform ellipse fitting and calculate the corresponding R o and d, and then substitute the above formula to get ν T ′. Gradually increase ν T , until ν T With ν T ′ is approximately equal to v T Finally, the bubble velocity is calculated to be 0.159 m / s, and the ellipse fitting result is as follows: Fig.12 As shown in (b), the bubble size is 2.66 mm. According to the test, when the connected images of the bubble in the image sequence exceed 5 frames, a stable speed and shape solution can be obtained. Fig.12 The additional 9 bubbles in the L1 frame in (a) also meet this condition, and the solution results are shown in Table 2. The bubble diameter is the diameter of a sphere with the same volume.
[0111] Table 2 Analysis of particle size and velocity of multiple bubbles
[0112]
[0113] At 60ml / min and 100ml / min gas flow rates, 800 frames (1 second) of bubble image sets were collected for bubble group reconstruction. 1149 and 1487 bubbles were tracked respectively, and the particle size distribution was obtained as follows: Fig.13 As shown, the computational overhead is 27 and 34 minutes respectively, which is proportional to the number of bubbles. Fig.13 It can be seen that the bubble size of the current system is mainly distributed in the range of 1.6mm to 3.6mm. As the gas flow rate increases, the average bubble size increases slightly, but small-size bubbles are still the main ones, which is consistent with the characteristics of the microporous ceramic bubble generator and the low probability of bubble aggregation under bubbly flow conditions.
[0114] Since the conductivity probe method can only obtain the mean characteristics of the bubble group, the mean data obtained from the particle size analysis experiment are compared with it. The results are shown in Table 3, and the mutual deviation is less than 10%. Considering that the probe method measurement data also contains a certain degree of uncertainty, the deviation should be within a reasonable range.
[0115] Table 3 Bubble group characteristic mean statistics
[0116]
[0117] IV. Conclusion
[0118] The error between the instantaneous gas content measurement value and the theoretical calculation value of the sliding speed ratio model in the bubbling flow interval is less than 10%; the peak distribution characteristics of the wall bubble group can be distinguished in the gas content mean imaging measurement; the measured average velocity and average particle size are compared with the conductivity probe method measurement data, and the mutual deviation is less than 10%. In terms of range and accuracy, the method of the present invention is superior to the traditional longitudinal plane imaging method, and the measurement types of bubble characteristic parameters are richer; compared with the probe method, it will not disturb the internal flow field and can obtain higher-dimensional bubble distribution information, which has practical value for analyzing the influence of different working conditions on the bubble distribution characteristics in reactor thermal design.
[0119] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit the technical solution. Those skilled in the art should understand that those modifications or equivalent substitutions of the technical solution of the present invention that do not depart from the purpose and scope of the technical solution should be included in the scope of the claims of the present invention.
Claims
1. A method for measuring characteristics of bubble groups in gas-liquid two-phase flow based on laser cross-section imaging, characterized in that: The following steps are involved: S1: emitting a laser beam toward a transparent container where a gas-liquid two-phase flow occurs, so as to form a laser detection cross section in the transparent container along the irradiation direction of the laser beam; S2: Acquire the bubble cross-sectional image on the laser detection section; In step S2, a bubble cross-sectional image is collected at an angle perpendicular to the laser detection cross-sectional area; In step S2, the bubble cross-section image is preprocessed, and the image preprocessing specifically includes the following steps: S201: placing a checkerboard calibration plate at a position corresponding to the laser detection section to obtain a corresponding calibration image, and then performing lens distortion correction and pixel scale calculation on the bubble cross-section image by a polynomial coordinate transformation method; S202: collecting a background image of the laser detection cross section in a bubble-free state, and then differentially eliminating interference spots in the bubble cross section image; S203: converting the bubble cross-section image into a bubble cross-section grayscale image, and binarizing the bubble cross-section grayscale image by using a maximum inter-class variance method to obtain a cross-section bubble contour image; S204: Suppressing scattered noise and random noise in the cross-sectional bubble contour image by median filtering and Gaussian filtering respectively; S3: classifying bubbles on the bubble cross-section image to generate regular bubble images of single bubbles and irregular bubble images of adhered or overlapping bubbles; In step S3, the roundness value R of each bubble in the bubble cross-section image is calculated based on the connected domain feature. o ; Then based on the roundness value R of the bubble o The bubbles are divided into single bubbles and adhesion or overlapping bubbles; finally, a regular bubble image of the single bubble and an irregular bubble image of the adhesion or overlapping bubbles are generated; Where: L represents the perimeter of the connected domain of the bubble cross section; s represents the area of the connected domain of the bubble cross section; R o <0.98, the bubble is a single bubble; R o ≥0.98, the bubbles are adhesion or overlapping bubbles; S4: segmenting the adhered or overlapping bubbles in the irregular bubble image; Step S4 specifically includes the following steps: S401: Calculate the radial radius [R] of each edge pixel in the irregular bubble image based on radial symmetry transformation min ,R max ], and then convert the irregular bubble image into a local radially symmetric image that highlights the image gradient, and finally perform threshold segmentation to obtain the corresponding binary image; S402: Calculate the centroid coordinates of the connected domain of the binary graph as the seed point; then calculate the correlation between each edge pixel and the seed point by using two indices, distance metric and gradient metric, to obtain the edge pixel set most correlated with each seed point; S403: performing ellipse fitting on the edge pixel set of each seed point, inferring the missing part of the edge pixel set of the adhesion or overlapping object, so as to achieve the segmentation of the adhesion or overlapping bubbles; S404: Determine the radial radius [R min ,R max ] is the optimal radial radius. If so, the segmented irregular bubble image is output; otherwise, the radial radius [R min ,R max ] and returns to step S401; In step S404, the radial radius [R min ,R max The condition for whether it is the optimal radial radius is that the overlap between the two output fitting ellipses is less than 0.7, and the ratio of the total area of the ellipse after removing the overlap to the area of the original undivided connected domain is greater than 0.95 and less than 1.05; S5: extracting characteristic parameters based on the regular bubble image and the segmented irregular bubble image, and then calculating the corresponding liquid phase cross-section gas content and bubble position distribution; In step S5, the number of bubbles, bubble pixel area and bubble position coordinates of the regular bubble image and the segmented irregular bubble image are extracted by an ellipse fitting algorithm, and then the corresponding liquid phase cross-section gas content and bubble position distribution are calculated.
2. The method for measuring characteristics of gas-liquid two-phase flow bubble groups based on laser cross-section imaging according to claim 1, characterized in that: In step S1, a laser beam is emitted toward the transparent container at an inclination angle of 30° to 45° on a horizontal plane.
3. The method for measuring characteristics of gas-liquid two-phase flow bubble groups based on laser cross-section imaging according to claim 1, characterized in that: In step S5, the corresponding liquid phase cross-section gas content is calculated by the following formula; α=∑a n / A; Where: α represents the gas content of the liquid phase cross section; a n represents the pixel area of each bubble in the laser detection cross section; A represents the total pixel area of the laser detection cross section.
Citation Information
Patent Citations
A device for measuring the volume concentration of microbubbles in a gas-liquid two-phase flow
CN105092684B
Method for acquiring and identifying structure images of foams flowing in porous medium in real time
CN102129548A
Device for measuring section void fraction of two-phase fluids
CN103529003A
Gas-liquid two-phase flow microbubble volume concentration measurement device
CN105092684A
Gas-liquid two-phase flow overlapped bubble image segmentation method
CN110415257A
