Vapor-liquid two-phase analysis method, device, medium and equipment

Abstract

The invention discloses a vapor-liquid two-phase analysis method, device, medium and equipment, and relates to the technical field of vapor-liquid two-phase analysis. Comprising the following steps: acquiring a grayscale image of vapor-liquid two-phase flow in a flow channel in a nuclear reaction rod submerging and cooling process; the grayscale image is divided into a liquid phase main body area and a vapor phase area according to the grayscale gradient change, and the vapor phase area comprises a complete transmission area and an incomplete transmission area; calculating the initial area of the liquid-phase main body region, and correcting the initial area of the liquid-phase main body region according to the surface tension limitation of the incomplete transmission region to obtain the corrected area of the liquid-phase main body region; a dispersion liquid area dispersed in the vapor phase area is extracted from the grayscale image based on a neural network model, the area of the dispersion liquid area is calculated and corrected, and the corrected area of the dispersion liquid area is obtained; adding the liquid phase main body area correction area and the dispersion liquid area correction area to obtain the total liquid phase area, and taking the ratio of the total liquid phase area to the flow channel area as the liquid content.

Description

Technical Field

[0001] The present invention relates to the technical field of vapor-liquid two-phase analysis, and particularly to a vapor-liquid two-phase analysis method, device, medium and equipment. Background Art

[0002] During the operation of a nuclear reactor, flooded cooling is a safety measure to remove heat by introducing cooling water to prevent the reactor from overheating. During flooded cooling, the flow channel will successively experience multiple two-phase flow regimes (dispersion flow, inverse annular flow, slug flow, bubbly flow, etc.), that is, liquid water and water vapor coexist. The liquid holdup (gas holdup), as a key characteristic parameter in different heat transfer stages and different two-phase flow regimes, is an important input for heat transfer capacity and flow pressure drop prediction models in each stage. Therefore, an intelligent analysis method is needed to monitor and evaluate the liquid holdup to ensure that the thermodynamic state of the reactor is within a safe range. However, due to refraction caused by the uneven distribution of the liquid film phase interface, it affects the calculation of the liquid holdup in traditional visualization data processing. Summary of the Invention

[0003] Based on this, in order to solve the technical problems in the prior art, the present invention provides a vapor-liquid two-phase analysis method, device, medium and equipment.

[0004] The present invention provides a vapor-liquid two-phase analysis method, including:

[0005] Obtaining a grayscale image of the vapor-liquid two-phase flow in the flow channel during the flooded cooling of the nuclear reaction rod;

[0006] Dividing the grayscale image into a liquid-phase main region and a vapor phase region according to the change of the grayscale gradient, wherein the vapor phase region includes a completely transmissive region and an incompletely transmissive region; calculating the initial area of the liquid-phase main region, and correcting the initial area of the liquid-phase main region according to the surface tension limit of the incompletely transmissive region to obtain the corrected area of the liquid-phase main region;

[0007] Extracting the dispersed liquid regions in the vapor phase region from the grayscale image based on a neural network model, calculating the area of the dispersed liquid regions and correcting it to obtain the corrected area of the dispersed liquid regions;

[0008] Adding the corrected area of the liquid-phase main region and the corrected area of the dispersed liquid regions to obtain the total liquid area, and taking the ratio of the total liquid area to the area of the flow channel in the grayscale image as the liquid holdup.

[0009] Further, the correction of the initial area of the liquid-phase main region according to the surface tension limit of the incompletely transmissive region is achieved by the following formula:

[0010]

[0011] where y is the length of the narrow side of the flow channel, x1 and x 4 are the left and right dimensions of the completely transmissive region in the flow channel, x 2 and x 3 are the left and right dimensions between the incomplete transmissive region located between the completely transmissive region and the nuclear reaction rod in the flow channel; A 2 is the corrected area of the liquid-phase main body region; the meaning of the above formula is to correct the initial area of the liquid-phase main body region using an elliptical structure model according to the surface tension limiting effect of the incomplete transmissive region.

[0012] Furthermore, calculating the area of the dispersed liquid region and performing correction specifically includes:

[0013] Calculating the initial area A of the dispersed liquid region according to the cv2.coutourArea function 1-1 ;

[0014] Based on the three-dimensional model of spherical liquid droplets, correcting the initial area A of the dispersed liquid region 1-1 The correction formula is:

[0015]

[0016] where R is the radius of the dispersed liquid droplet; y is the narrow side length of the flow channel; n is the total number of dispersed liquid droplets; A 1 is the corrected area of the dispersed liquid region.

[0017] Furthermore, extracting the liquid-phase main body region from the grayscale image according to the change of the gray level gradient specifically includes:

[0018] Using any one of the gradient-based Suzuki and Abe contour tracking algorithm or the gradient-based canny edge algorithm to extract all the phase interface contours between the liquid-phase main body region and the vapor-phase region from the grayscale image;

[0019] Sorting all the phase interface contours according to the perimeter or area of the phase interface contours, and extracting the top N phase interface contours with the sorting of N;

[0020] Calculating the average pixel gray level within the N phase interface contours respectively, and comparing the average pixel gray level within the N phase interface contours with a preset threshold: taking the phase interface contour region greater than the preset threshold as the liquid-phase main body region and setting its pixel value to 0; taking the remaining region as the vapor-phase region and setting its pixel value to 255.

[0021] Furthermore, extracting the dispersed liquid region dispersed in the vapor-phase region from the grayscale image based on the neural network model specifically includes:

[0022] Perform background elimination and erosion-dilation operations on the grayscale image to eliminate and fill the tiny holes and gaps in the grayscale image, obtaining an initial diffuse grayscale image;

[0023] Extract the droplet positions and droplet sizes in the initial diffuse grayscale image through a pre-trained neural network model;

[0024] According to the droplet positions and droplet sizes, use the gradient-based canny edge algorithm to identify the droplet contours;

[0025] Based on a preset grayscale threshold, fill the droplet contours identified by the canny edge algorithm to obtain a diffuse liquid region dispersed in the vapor phase region.

[0026] Further, the calculation of the initial area of the main liquid phase region is implemented using the cv2.contourArea function.

[0027] The present invention provides a vapor-liquid two-phase analysis device, including:

[0028] A data acquisition module for acquiring grayscale images of the vapor-liquid two-phase flow in the flow channel during the flooding cooling process of the nuclear reaction rod;

[0029] A main liquid phase region area acquisition module for dividing the grayscale image into a main liquid phase region and a vapor phase region according to the grayscale gradient change, where the vapor phase region includes a fully transmitted region and a partially transmitted region; calculating the initial area of the main liquid phase region, and correcting the initial area of the main liquid phase region according to the surface tension limit of the partially transmitted region to obtain the corrected area of the main liquid phase region;

[0030] A diffuse liquid phase region area acquisition module for extracting a diffuse liquid region dispersed in the vapor phase region from the grayscale image based on a neural network model, calculating and correcting the area of the diffuse liquid region to obtain the corrected area of the diffuse liquid region;

[0031] A liquid holdup acquisition module for adding the corrected area of the main liquid phase region and the corrected area of the diffuse liquid region to obtain the total liquid phase area, and taking the ratio of the total liquid phase area to the flow channel area in the grayscale image as the liquid holdup

[0032] The present invention provides a computer-readable storage medium, where the storage medium stores a computer program, and when the computer program is executed by a processor, the above vapor-liquid two-phase analysis method is implemented.

[0033] The present invention provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the above vapor-liquid two-phase analysis method is implemented.

[0034] The above at least one technical solution adopted by the present invention can achieve the following beneficial effects:

[0035] In the gas-liquid two-phase analysis method provided by the present invention, area liquid-phase correction is performed on the refraction region: by quantifying the influence of the surface tension of the incomplete transmission region on the liquid film morphology through an elliptical structure model, the geometric characteristics of the incomplete transmission region are incorporated into the correction calculation, solving the area measurement error caused by uneven liquid film distribution, and significantly improving the calculation accuracy of the liquid holdup based on planar images.

[0036] Figure description

[0037] The figures described herein are used to provide a further understanding of the present invention and form a part of the present invention. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the figures:

[0038] Figure 1 is a schematic diagram of the defects of the prior art provided by the present invention;

[0039] Figure 2 is a schematic flow diagram of the two-phase identification and analysis method provided by the present invention;

[0040] Figure 3 is a schematic diagram of the image processing example of the main liquid-phase region provided by the present invention, Figure 3 where (a) is the original image, Figure 3 where (b) is a schematic diagram of the contour extraction result, Figure 3 where (c) is a schematic diagram of the maximum contour recognition result, Figure 3 where (d) is a schematic diagram of the region filling result;

[0041] Figure 4 is the black region area A provided by the present invention 2 schematic diagram of the correction principle;

[0042] Figure 5 is a schematic flow diagram of the background elimination provided by the present invention;

[0043] Figure 6 is the recognition result of the dispersed droplet image provided by the present invention;

[0044] Figure 7 is for the present invention Figure 6 schematic diagram of the result obtained by performing the filling operation. Detailed implementation manners

[0045] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding figures. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the specification, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0046] Existing two-phase intelligent identification methods generally use invasive equipment (probes, wire meshes, conductivity probes, etc.) in the original data acquisition method (phase refers to the homogeneous material part with the same composition and the same physical and chemical properties in a system, and there is a clear separable interface between the phases. Two-phase flow or multiphase flow refers to the flow of two or more different phases of matter at the same time). According to the difference in electrical conductivity between the vapor and liquid phases, different potential differences and retention times are generated to judge the size of the liquid or vapor phase. However, invasive equipment is not easy to install in narrow channels, and secondly, it will destroy the flow morphology of the downstream flow field in narrow flow channels. There are differences in the excitation fluorescence intensity of fluorescent particles under high temperature conditions, which affects the two-phase identification. High-speed visualization, as a non-invasive observation method, not only does not destroy the internal flow field but also completely preserves the two-phase behavior at each time sequence. Existing two-phase image processing methods such as "Zhou Yunlong, Shang Qiuhua, Fan Zhenru, et al. Image detection method for volumetric gas content of gas-liquid two-phase flow [J]. Thermal Energy and Power Engineering, 2008, (05): 507-511+556" use high-speed photography and image processing methods to process bubbles and obtain gas content. However, this article only realizes the two-phase identification and processing of bubbly flow; "Zhang Dalin, Zhou Jiancheng, Chen Guo, et al. A method for measuring gas content in a narrow channel with a large aspect ratio based on image processing [P]. Shaanxi Province: CN201910238272.8, 2020-07-28." realizes the measurement and processing of bubbly and elastic bubbles in a narrow channel with a large aspect ratio. However, the image processing process does not take into account the refraction problem caused by the non-uniform distribution of the liquid phase in the channel. The phase interface is rough, such as Figure 1 As shown, this leads to improper treatment of the vapor-liquid phase during the filling operation, which is only applicable to the two-phase treatment analysis under the condition that the vapor-liquid phase completely fills the channel and ignores the situation where multiple flow types exist at the same time.

[0047] Based on this, the present invention provides a vapor-liquid two-phase analysis method, which aims to provide an intelligent identification and analysis method for the full flow pattern of vapor-liquid two-phase in a narrow rectangular channel. The method can identify the full two-phase flow pattern in the submerged cooling process, including the dispersed flow stage with small liquid phase size and the anti-annular flow stage with violent phase interface fluctuations due to high-temperature phase change, and obtain the liquid content in different heat exchange stages and different two-phase flow states.

[0048] Two-phase intelligent identification and analysis through Figure 2Implementation of the method process. The right branch of the process schematic diagram is the extraction of the main contour recognition based on the flow pattern characteristics of the submerged cooling process. To compensate for the deviation in the calculation of the liquid holdup caused by improper image processing (forgetting the small contour of droplets) during the extraction of the main contour recognition, the left branch of the process schematic diagram is added. The characteristics of dispersed droplets are recognized and extracted through deep learning and image processing to accurately obtain the liquid holdup of spherical droplets in the two-phase flow pattern.

[0049] The following combines Figure 2 to describe in detail the technical solutions provided by each embodiment of the present application, specifically including the following steps:

[0050] S1: Obtain the grayscale image of the vapor-liquid two-phase flow in the flow channel during the submerged cooling process of the nuclear reactor.

[0051] After inputting the original images of the two-phase flow at each stage, first convert the original images into grayscale images.

[0052] S2: Divide the grayscale image into the liquid-phase main region and the vapor phase region according to the change of the grayscale gradient. Among them, the vapor phase region includes the completely transmissive region and the incompletely transmissive region; calculate the initial area of the liquid-phase main region, and correct the initial area of the liquid-phase main region according to the surface tension limit of the incompletely transmissive region to obtain the corrected area of the liquid-phase main region.

[0053] Use any one of the Suzuki and Abe contour tracing algorithms based on gradients or the canny edge algorithm based on gradients to extract all the phase interface contours between the liquid-phase main region and the vapor phase region from the grayscale image; sort all the phase interface contours according to the perimeter or area of the phase interface contours, and extract the top N phase interface contours with the sorting of N; calculate the average pixel grayscale values within the N phase interface contours respectively, and compare the average pixel grayscale values within the N phase interface contours with a preset threshold: regard the phase interface contour region greater than the preset threshold as the liquid-phase main region and set its pixel value to 0; regard the remaining region as the vapor phase region and set its pixel value to 255.

[0054] Extract the phase contour interface according to the grayscale gradient of the image, as shown in (b) of Figure 3 , Figure 3 where (a) in Figure 3 is the original image. At this time, due to the existence of discrete droplets and transmissive regions, there are a large number of phase interface contours. Based on the understanding of the vapor-liquid phase distribution in the annular flow (counter-annular flow) pattern, obtain the largest N contours according to the perimeter / area (implemented by the cv2.coutourArea function). The specific value of N needs to be considered in combination with the flow channel size and flow channel structure, as shown in (c) of Figure 3As shown in (d) therein, a phase distribution with a clear interface is obtained.

[0055] For the filled area, the actual liquid holdup cannot be directly calculated based on the area of the black area. The area of the black area needs to be corrected as follows:

[0056]

[0057] where y is the length of the narrow side of the flow channel, and x 1 and x 4 are the left and right dimensions of the completely transmissive area in the flow channel recorded by the camera, and x 2 and x 3 are the left and right dimensions between the completely transmissive area and the incomplete transmissive area (where a black area appears due to refraction) between the nuclear reaction rod in the flow channel recorded by the camera; A 2 is the corrected area of the liquid-phase main body area; the meaning of the above formula is to correct the initial area of the liquid-phase main body area using an elliptical structure model according to the surface tension limiting effect of the incomplete transmissive area. The liquid film wave height is approximated as an elliptical structure, as Figure 4 shown.

[0058] S3: Based on the neural network model, extract the dispersed liquid area in the vapor phase region from the grayscale image, calculate the area of the dispersed liquid area and correct it to obtain the corrected area of the dispersed liquid area.

[0059] Perform background elimination operation and erosion and dilation operation on the grayscale image to eliminate and fill the tiny holes and gaps in the grayscale image to obtain an initial dispersed grayscale image: Compare the input original image to eliminate the noise in the background image. Based on a kernel of size (2, 2), perform erosion and dilation (sliding window mechanism) on the image after background elimination, so that the tiny holes and gaps are eliminated and filled. The result of background elimination is as Figure 5 shown.

[0060] After background elimination, perform droplet annotation on the image, establish a training set and a test set for droplet image recognition, and obtain the droplet position and droplet size. Traverse the droplets (the number is n) in the image in sequence. According to the droplet position and droplet size, use the canny edge algorithm based on the gradient for contour recognition, and fill the droplet contours recognized by the canny edge algorithm based on a preset grayscale threshold to obtain the dispersed liquid area in the vapor phase region. The results of droplet image recognition and image processing are as Figure 6 and Figure 7 shown.

[0061] For the filled dispersed liquid droplet area, the actual liquid holdup cannot be directly calculated based on the area of the black area. The area of the black area needs to be corrected as follows:

[0062]

[0063] wherein, R is the radius of the dispersed droplets; y is the length of the narrow side of the flow channel; n is the total number of dispersed droplets; A 1 is the corrected area of the dispersed liquid region.

[0064] S4: Add the area of the main liquid phase region and the area of the dispersed liquid phase region to obtain the total liquid phase area, and take the ratio of the total liquid phase area to the area of the flow channel in the grayscale image as the liquid holdup.

[0065] Finally, the liquid holdup can be obtained as:

[0066]

[0067] wherein, A_flow_channel is the area of the flow channel in the visualization image, which can be equivalent to the area of the non-black region in the background image.

[0068] Based on Figure 2 the vapor-liquid two-phase analysis method shown above, has the following effects:

[0069] 1. Effectively solves the deviation in liquid holdup calculation caused by non-fully filled liquid phase refraction, and at the same time considers the dispersed liquid phase in the complex mixed flow pattern, improving the accuracy of liquid holdup in image recognition.

[0070] 2. This method is based on non-invasive data acquisition means and image recognition and processing technology, and can provide solutions for the difficulty in identifying complex mixed flow patterns in the field of boiling two-phase flow.

[0071] 3. Can provide an image digital processing method for two-phase flow research; promote the construction of two-phase flow pattern maps and establish the quantitative relationship between flow and heat transfer parameters and vapor-liquid two-phase flow patterns; the results can provide accurate data input for flow and heat transfer relations.

[0072] 4. This method can be extended to the industrial production field. By real-time collecting flow parameters and signals and using fast feature extraction and recognition algorithms, it can realize the real-time identification and monitoring of two-phase flow patterns.

[0073] The above is the vapor-liquid two-phase analysis method provided by one or more embodiments of the present invention. Based on the same idea, the present invention also provides a corresponding vapor-liquid two-phase analysis device, including:

[0074] A data acquisition module, used to obtain the grayscale image of the vapor-liquid two-phase flow in the flow channel during the flooding cooling process of the nuclear reaction rod.

[0075] The main liquid-phase area acquisition module is used to divide the grayscale image into a liquid-phase main area and a vapor-phase area according to the change of grayscale gradient. Among them, the vapor-phase area includes a completely transmissive area and an incompletely transmissive area; calculate the initial area of the liquid-phase main area, and correct the initial area of the liquid-phase main area according to the surface tension limit of the incompletely transmissive area to obtain the corrected area of the liquid-phase main area.

[0076] The dispersed liquid-phase area acquisition module is used to extract the dispersed liquid areas in the vapor-phase area from the grayscale image based on a neural network model, calculate the area of the dispersed liquid areas and correct it to obtain the corrected area of the dispersed liquid areas.

[0077] The liquid holdup acquisition module is used to add the corrected area of the liquid-phase main area and the corrected area of the dispersed liquid areas to obtain the total liquid area, and use the ratio of the total liquid area to the flow channel area in the grayscale image as the liquid holdup.

[0078] For the specific limitations of the vapor-liquid two-phase analysis device, reference can be made to the limitations of the vapor-liquid two-phase analysis method in the above text, which will not be elaborated here. Each module in the above vapor-liquid two-phase analysis device can be implemented in whole or in part by software, hardware and their combination. The above modules can be embedded in the processor in the computer device in the form of hardware or be independent of the processor, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0079] The present invention also provides a computer-readable storage medium, which stores a computer program, and the computer program can be used to execute the above-provided vapor-liquid two-phase analysis method.

[0080] The present invention also provides the structure of a computer device. At the hardware level, the computer device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory. Of course, it may also include other hardware required for other services. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to implement the above-provided vapor-liquid two-phase analysis method.

[0081] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided by the present invention can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0082] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope recorded by the present invention.

Claims

1. A vapor-liquid two-phase analysis method, characterized in that: include: Obtaining grayscale images of the vapor-liquid two-phase flow in the flow channel during the submerged cooling process of the nuclear reactor rod; The grayscale image is divided into a liquid phase main region and a vapor phase region according to the grayscale gradient change, wherein the vapor phase region includes a completely transparent region and an incompletely transparent region; the initial area of ​​the liquid phase main region is calculated, and the initial area of ​​the liquid phase main region is corrected according to the surface tension limit of the incompletely transparent region to obtain a corrected area of ​​the liquid phase main region; Based on the neural network model, the dispersed liquid region dispersed in the vapor phase region is extracted from the grayscale image, the area of ​​the dispersed liquid region is calculated and corrected, and the corrected area of ​​the dispersed liquid region is obtained; The total liquid area is obtained by adding the corrected area of ​​the liquid main region and the corrected area of ​​the dispersed liquid region, and the ratio of the total liquid area to the flow channel area in the grayscale image is taken as the liquid content.

2. The vapor-liquid two-phase analysis method according to claim 1, characterized in that: The correction of the initial area of ​​the liquid phase main region according to the surface tension limitation of the incomplete transmission region is achieved by the following formula: Wherein, y is the narrow side length of the flow channel, x1 and x4 are the left and right dimensions of the completely transmitted area in the flow channel, x2 and x3 are the left and right dimensions between the incompletely transmitted area between the completely transmitted area and the nuclear reactor rod in the flow channel; A2 is the corrected area of ​​the liquid phase main area; the above formula means that according to the surface tension limitation of the incompletely transmitted area, the initial area of ​​the liquid phase main area is corrected using an elliptical structure model.

3. The vapor-liquid two-phase analysis method according to claim 1, characterized in that: The calculating and correcting the area of ​​the dispersion region specifically includes: Calculate the initial area A of the dispersion region according to the cv2.coutourArea function 1-1 ; The initial area A of the dispersed liquid region based on the three-dimensional model of the spherical droplet 1-1 Make corrections, the correction formula is: Wherein, R is the radius of the dispersed droplet; y is the narrow side length of the flow channel; n is the total number of dispersed droplets; A1 is the corrected area of ​​the dispersed liquid region.

4. The vapor-liquid two-phase analysis method according to claim 1, characterized in that: The step of extracting the liquid phase main region from the grayscale image according to the grayscale gradient change specifically includes: Extract all phase interface contours between the liquid phase bulk region and the vapor phase region from the grayscale image using either the gradient-based Suzuki and Abe contour tracing algorithm or the gradient-based canny edge algorithm; All the phase interface contours are sorted according to the perimeter or area of ​​the phase interface contours, and the top N phase interface contours are extracted; Calculate the average grayscale values ​​of the pixels in the N phase interface contours respectively, and compare the average grayscale values ​​of the pixels in the N phase interface contours with a preset threshold: take the phase interface contour area greater than the preset threshold as the main liquid phase area, and set its pixel value to 0; take the remaining area as the vapor phase area, and set its pixel value to 255.

5. The vapor-liquid two-phase analysis method according to claim 1, characterized in that: The method of extracting the dispersed liquid region dispersed in the vapor phase region from the grayscale image based on the neural network model specifically includes: Perform background removal and erosion and dilation operations on the grayscale image to eliminate and fill tiny holes and gaps in the grayscale image to obtain an initial diffuse grayscale image; The droplet position and droplet size in the initial diffuse grayscale image are extracted through the pre-trained neural network model; According to the droplet position and droplet size, the droplet contour is identified using the gradient-based canny edge algorithm; The droplet contour identified by the canny edge algorithm is filled based on a preset grayscale threshold to obtain a dispersed liquid area dispersed in the vapor phase area.

6. The vapor-liquid two-phase analysis method according to claim 1, characterized in that: The calculation of the initial area of ​​the liquid phase main region is achieved using the cv2.coutourArea function.

7. A vapor-liquid two-phase analysis device, characterized in that: include: A data acquisition module, used to obtain a grayscale image of the vapor-liquid two-phase flow in the flow channel during the submerged cooling process of the nuclear reactor rod; The main liquid phase area acquisition module is used to divide the grayscale image into a liquid phase main area and a vapor phase area according to the grayscale gradient change, wherein the vapor phase area includes a completely transparent area and an incompletely transparent area; calculate the initial area of ​​the liquid phase main area, and correct the initial area of ​​the liquid phase main area according to the surface tension limit of the incompletely transparent area to obtain the corrected area of ​​the liquid phase main area; A dispersed liquid phase region area acquisition module is used to extract the dispersed liquid region dispersed in the vapor phase region from the grayscale image based on a neural network model, calculate the area of ​​the dispersed liquid region and perform correction to obtain the corrected area of ​​the dispersed liquid region; The liquid content acquisition module is used to add the corrected area of ​​the liquid phase main region and the corrected area of ​​the dispersed liquid region to obtain the total liquid phase area, and take the ratio of the total liquid phase area to the flow channel area in the grayscale image as the liquid content.

8. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method described in any one of claims 1 to 6 is implemented.

Images