Method and device for acquiring vapor content of flowing boiling vapor-liquid two-phase flow and medium

Abstract

The invention discloses a method and device for acquiring the vapor content of a flowing boiling vapor-liquid two-phase flow and a medium, and the method comprises the steps: acquiring a bubble image which is a flowing boiling bubble image; converting the bubble image into a grey-scale map and carrying out binarization processing to obtain a binary image containing flow information; based on the binary image containing the flow information, extracting image features by adopting wavelet transform or a grey level histogram; a multi-layer deep BP neural network is constructed, the multi-layer deep BP neural network is trained based on the extracted image features, and a vapor-liquid two-phase flow pattern recognition model based on the BP neural network is obtained; a vapor-liquid two-phase flow pattern recognition model based on a BP neural network is adopted to carry out flow pattern recognition on the to-be-recognized bubble image, and a flow pattern recognition result is obtained; and according to different flow pattern recognition results, corresponding steam-containing rate calculation is carried out respectively, and the steam-containing rate value of the two-phase flow is obtained. The calculation accuracy is high, and the experimental data processing efficiency is effectively improved.

Description

Technical Field

[0001] The invention relates to the field of nuclear reactor flow boiling visualization test, and in particular to a method, a device and a medium for obtaining the vapor content of a flow boiling vapor-liquid two-phase flow. Background Art

[0002] The visualization test of nuclear reactor flow boiling is an important type of experiment in nuclear reactor thermal hydraulic test. The flow pattern and vapor fraction data obtained from the visualization test of nuclear reactor flow boiling are of great significance for determining the flow pattern characteristics and bubble dynamics characteristics in the reactor core channel.

[0003] However, the phenomenon of nuclear reactor flow boiling process is complex, with multiple flow patterns, and compared with ordinary gas-water two-phase flow, the bubble movement law is more complicated. The vapor-liquid interface is complex and easily blurred under many flow patterns. Moreover, the number of pictures obtained by visualization experiments ranges from a few thousand to hundreds of thousands. It takes a lot of manpower and time to analyze the vapor content of each picture. The existing vapor content calculation method mainly uses pressure, temperature, heating heat flux density and other methods for macroscopic calculation. The obtained vapor content data has low accuracy and cannot obtain the high-precision quantitative change law of the vapor content data over time.

[0004] In view of this, this application is hereby filed. Summary of the invention

[0005] The technical problem to be solved by the present invention is that the existing vapor content calculation method mainly uses pressure, temperature, heating heat flux density and other methods to perform macroscopic calculations, and the obtained vapor content data has low accuracy. The purpose of the present invention is to provide a method, device and medium for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow. First, a vapor-liquid two-phase flow pattern recognition model based on a BP neural network is used to divide the flow pattern of the vapor-liquid two-phase flow; secondly, the boundary marking method and the pixel intensity method are used to calculate and analyze the vapor content of different flow patterns. The calculation accuracy is high, which greatly reduces the quantitative analysis time of the boiling two-phase flow and effectively improves the experimental data processing efficiency. It is an important analysis method and tool for the visualization experiment of flow boiling in nuclear reactors.

[0006] The present invention is achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow, the method comprising:

[0008] Acquire a bubble image, wherein the bubble image is a bubble image of flowing boiling;

[0009] Converting the bubble image into a grayscale image and performing binarization processing to obtain a binary image containing flow information;

[0010] Based on the binary image containing flow information, wavelet transform or grayscale histogram is used to extract image features;

[0011] A multi-layer deep BP neural network is constructed and trained based on the extracted image features to obtain a gas-liquid two-phase flow pattern recognition model based on the BP neural network.

[0012] The flow pattern recognition model of vapor-liquid two-phase flow based on BP neural network is used to recognize the flow pattern of the bubble image to be recognized, and the flow pattern recognition result is obtained;

[0013] According to different flow pattern identification results, the corresponding vapor fraction calculation is performed to obtain the vapor fraction value of the two-phase flow.

[0014] Furthermore, the bubble image is obtained from a nuclear reactor flow boiling visualization experiment;

[0015] In the nuclear reactor flow boiling visualization test, the channel is heated, and the bubbles in the channel are generated by boiling heat transfer rather than by injecting air or other non-condensable gases from the channel inlet. The vapor-liquid interface is complex and easily blurred, and the bubble movement law is complex.

[0016] Furthermore, the multi-layer deep BP neural network includes an input layer, an intermediate layer and an output layer;

[0017] The input layer is used to input a feature vector formed by effective feature parameters extracted by wavelet transform, or a feature vector formed by identification parameters extracted by probability distribution of gray levels in gray histogram;

[0018] The output layer is used to output flow type recognition results.

[0019] Furthermore, the flow pattern identification results include diffuse bubbly flow, bubbly flow, plug flow, annular flow and swirling flow.

[0020] Furthermore, according to different flow pattern identification results, corresponding steam content calculations are performed respectively, including:

[0021] When the flow pattern identification results are diffuse bubbly flow, bubbly flow, plug flow and annular flow, the boundary marking method is used to calculate the vapor fraction; the boundary marking method is to mark different areas separated by the phase interface as gas phase or liquid phase, that is, to calculate the vapor fraction value of the two-phase flow according to the area of ​​different phases;

[0022] When the flow pattern identification result is a churning flow, the vapor fraction is calculated using a pixel intensity method; the pixel intensity method is to calculate the vapor fraction value of the two-phase flow based on image pixels.

[0023] Furthermore, the boundary marking method is used to calculate the steam content, including:

[0024] After grayscale processing of the bubble image, the canny operator is used to detect the edge of the grayscale processed image to obtain the phase interface information;

[0025] At the same time, the grayscale processed image is closed, i.e., the operation of dilation followed by erosion, to obtain prominent interface information;

[0026] Based on the prominent interface information, the corresponding image is filled so that the gas phase is marked as white and the vapor content is 1; the liquid phase is marked as black and the vapor content is 0;

[0027] The vapor fraction corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0028] Furthermore, the pixel intensity method is used to calculate the vapor content, including:

[0029] After grayscale processing of the image, the pixel intensity of the grayscale processed image is divided into multiple different levels, and each level corresponds to a vapor content value;

[0030] The vapor fraction value corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0031] In a second aspect, the present invention further provides a device for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow, the device comprising:

[0032] An acquisition unit, used for acquiring a bubble image, wherein the bubble image is a bubble image of flowing boiling;

[0033] A grayscale and binarization processing unit, used for converting the bubble image into a grayscale image and performing binarization processing to obtain a binary image containing flow information;

[0034] A feature extraction unit, used for extracting image features by using a wavelet transform method or a grayscale histogram method based on a binary image containing flow information;

[0035] A model building and training unit is used to build a multi-layer deep BP neural network, train the multi-layer deep BP neural network based on the extracted image features, and obtain a gas-liquid two-phase flow pattern recognition model based on the BP neural network;

[0036] A flow pattern recognition unit is used to perform flow pattern recognition on the bubble image to be recognized by using a vapor-liquid two-phase flow pattern recognition model based on a BP neural network to obtain a flow pattern recognition result;

[0037] The vapor content calculation unit is used to perform corresponding vapor content calculations according to different flow pattern identification results to obtain the vapor content value of the two-phase flow.

[0038] Furthermore, the steam content calculation unit includes:

[0039] The first vapor content calculation subunit is used to calculate the vapor content by using a boundary marking method when the flow pattern recognition result is a diffuse bubbly flow, a bubbly flow, a plug flow, and annular flow; the boundary marking method is to mark different areas separated by the phase interface as gas phase or liquid phase, that is, to calculate the vapor content value of the two-phase flow according to the area of ​​different phases;

[0040] The second vapor fraction calculation subunit is used to calculate the vapor fraction using a pixel intensity method when the flow pattern recognition result is a churning flow; the pixel intensity method is to calculate the vapor fraction value of the two-phase flow based on image pixels.

[0041] Furthermore, the execution process of the first steam content fraction calculation subunit is as follows:

[0042] After grayscale processing of the bubble image, the canny operator is used to detect the edge of the grayscale processed image to obtain the phase interface information;

[0043] At the same time, the grayscale processed image is closed, i.e., the operation of dilation followed by erosion, to obtain prominent interface information;

[0044] Based on the prominent interface information, the corresponding image is filled so that the gas phase is marked as white and the vapor content is 1; the liquid phase is marked as black and the vapor content is 0;

[0045] The vapor fraction corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0046] Furthermore, the execution process of the second steam content fraction calculation subunit is as follows:

[0047] After grayscale processing of the image, the pixel intensity of the grayscale processed image is divided into multiple different levels, each level corresponds to a vapor content value;

[0048] The vapor fraction value corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0049] In a third aspect, the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and wherein the computer program, when executed by a processor, implements the above-mentioned method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow.

[0050] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0051] The present invention discloses a method, device and medium for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow. Firstly, a vapor-liquid two-phase flow pattern recognition model based on a BP neural network is adopted to analyze the two-phase flow image obtained by a flowing boiling visualization test, and the vapor-liquid two-phase flow pattern is automatically obtained. Secondly, a boundary marking method and a pixel intensity method are adopted to calculate and analyze the vapor content of different flow patterns respectively, with high calculation accuracy, greatly reducing the quantitative analysis time of the boiling two-phase flow, and effectively improving the experimental data processing efficiency. The method is an important analysis method and tool for a flowing boiling visualization test of a nuclear reactor. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0053] Figure 1 A flow chart of a method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to the present invention;

[0054] Figure 2 A detailed flow chart of a method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to the present invention;

[0055] Figure 3 Schematic diagram of wavelet transform of the present invention;

[0056] Figure 4 It is a schematic diagram of the boundary marking method of the present invention;

[0057] Figure 5 The present invention is a structural block diagram of a device for acquiring the vapor content of a flowing boiling vapor-liquid two-phase flow. DETAILED DESCRIPTION

[0058] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0059] The existing vapor content calculation methods mainly use pressure, temperature, heating heat flux density and other methods to perform macroscopic calculations, and the obtained vapor content data has low accuracy. The present invention designs a vapor content acquisition method, device and medium for flowing boiling vapor-liquid two-phase flow. Firstly, a vapor-liquid two-phase flow pattern recognition model based on BP neural network is adopted to analyze the two-phase flow images obtained from the flow boiling visualization test and automatically obtain the vapor-liquid two-phase flow pattern; secondly, the boundary marking method and pixel intensity method are used to calculate and analyze the vapor content of different flow patterns respectively, with high calculation accuracy, greatly reducing the quantitative analysis time of boiling two-phase flow, and effectively improving the experimental data processing efficiency. It is an important analysis method and tool for nuclear reactor flow boiling visualization test.

[0060] Example 1

[0061] like Figure 1 and Figure 2 As shown, the present invention provides a method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow, the method comprising:

[0062] Step 1, obtaining a bubble image, wherein the bubble image is a bubble image of flowing boiling;

[0063] In this embodiment, the bubble image is obtained from a nuclear reactor flow boiling visualization experiment;

[0064] In the nuclear reactor flow boiling visualization test, the channel is heated, and the bubbles in the channel are generated by boiling heat transfer rather than by injecting air or other non-condensable gases from the channel inlet. The vapor-liquid interface is complex and easily blurred, and the bubble movement law is complex.

[0065] Step 2, converting the bubble image into a grayscale image and performing binarization processing to obtain a binary image containing flow information; in this implementation, the bubble image is trimmed to remove the background and only retain the flow information inside the pipeline, the image is converted into a grayscale image, and then binarization processing is performed to obtain a binary image containing flow information.

[0066] Step 3, based on the binary image containing flow information, image features are extracted using wavelet transform or grayscale histogram;

[0067] In this embodiment, the specific steps of wavelet transform are:

[0068] The three-layer wavelet transform decomposition of the binary image containing flow information and the corresponding sub-images of each layer are encoded. The image is decomposed into low-frequency signal A1 and high-frequency signal D1 through a first-level wavelet transform. A1 is subjected to wavelet transform to obtain A2 and D2, and A2 is subjected to wavelet transform to obtain A3 and D3. The high-frequency signal of each level contains horizontal, vertical and diagonal information respectively, which is stored in the sub-image obtained by wavelet transform.

[0069] In this embodiment, the grayscale histogram is the probability statistics of the distribution of multiple pixels with different grayscales in the image window.

[0070] Step 4, constructing a multi-layer deep BP neural network, training the multi-layer deep BP neural network based on the extracted image features and image data set, and obtaining a vapor-liquid two-phase flow pattern recognition model based on the BP neural network;

[0071] In this embodiment, the multi-layer deep BP neural network includes an input layer, an intermediate layer and an output layer;

[0072] The input layer is used to input a feature vector formed by effective feature parameters extracted by wavelet transform, or a feature vector formed by identification parameters extracted by probability distribution of gray levels in gray histogram;

[0073] The output layer is used to output flow pattern recognition results, with numbers 1-5 representing different flow patterns, where 1 represents diffuse bubbly flow, 2 represents bubbly flow, 3 represents plug flow, 4 represents churned flow, and 5 represents annular flow.

[0074] In a specific implementation, the number of layers may be 6 for the input layer, 12 for the middle layer, and 4 for the output layer.

[0075] Specifically, the feature vector formed by the effective feature parameters extracted by wavelet transform includes:

[0076] The image texture analysis method based on wavelet transform performs multi-scale refinement analysis on functions or signals through the expansion and translation of the mother wavelet, extracts texture information using wavelet transform coefficients, independently extracts feature parameters at each scale, and combines effective feature parameters to form a feature vector. Images with different texture features correspond to different feature vectors, so texture images can be classified.

[0077] The signal can be divided into two parts by wavelet transform, one part is the projection of the signal through the low-frequency band filter H, and the other part is the projection of the signal through the high-frequency band filter G. The pyramid algorithm is used to decompose the flow image and extract the image texture structure features. The db4 orthogonal wavelet transform with finite support is used. This wavelet transform has regularity and continuous vanishing moment proportional to its support size (2N). The high smoothness and large vanishing moment indicate that it has a good frequency division effect as a bandpass filter. At the same time, as a low-pass filter, it can better maintain the low-frequency component of the original signal without producing a large blurring effect. From the perspective of filtering, it is hoped that the decomposition layer number N takes a larger value. However, in image decomposition, N is proportional to the boundary effect of the image, especially when the decomposition level is higher, its influence is greater, and the wavelet transform time is proportional to N2. After many attempts, it was found that the effect was better when N=3, so it was determined to decompose a flow image I into three layers for feature extraction.

[0078] Therefore, the three-layer wavelet transform decomposition of the flow image (taking bubbly flow as an example) and the corresponding sub-image encoding of each layer are as follows: Figure 3 As shown, the image is decomposed into low-frequency signal A1 and high-frequency signal D1 through a first-level wavelet transform. A1 is subjected to wavelet transform to obtain A2 and D2, and A2 is subjected to wavelet transform to obtain A3 and D3. The high-frequency signal of each level contains horizontal, vertical and diagonal information respectively, which is stored in the sub-image obtained by wavelet transform.

[0079] Calculate the texture energy feature E of each sub-image after three-layer wavelet transform decomposition iand L1 norm feature F i , which are defined as follows:

[0080]

[0081] Among them, g i (x, y) represents the wavelet decomposition coefficient of a point (x, y) in the image at scale i, x, y represent the row and column values ​​of the pixel, M×N represents the size of the sub-image after transformation. After obtaining the energy characteristics and norm characteristics of the sub-images at all levels, the recognition parameters can be constructed according to the following rules to form a feature vector:

[0082]

[0083] The above six parameters represent the ratio of the diagonal component of the texture to the sum of the horizontal and vertical components in each decomposition level (same resolution), forming a feature vector X = [x1, x2, x3, x4, x5, x6] which can be used for flow type recognition.

[0084] Specifically, the grayscale image observed in the visual system must correspond to a grayscale histogram with a certain probability distribution, so different images can be distinguished by comparing the similarity of the grayscale histograms of two images.

[0085] The feature vector formed by the recognition parameters extracted using the probability distribution of gray levels in the gray histogram includes:

[0086] According to the probability distribution of gray levels, six identification parameters, namely mean m, standard deviation σ, smoothness R, third-order moment μ3, consistency U and average entropy e, are extracted, and the feature vector is composed of these six identification parameters.

[0087] In this embodiment, the image data set uses existing flow boiling bubble images, selects a certain number of pictures (generally should be greater than 100 pictures), and marks them as a training set.

[0088] In this embodiment, the feature vector extracted by wavelet transform is used as the input layer of the multi-layer deep BP neural network. After learning and training, the output layer outputs the flow pattern recognition result.

[0089] The multi-layer deep BP neural network training process in the software sets the number of neurons in the middle layer to 12, the training learning rate to 0.5, the training batch size to 150, and the number of training rounds to 5000. The final model can achieve a classification accuracy of 97.5% for the validation set.

[0090] Step 5, using a vapor-liquid two-phase flow pattern recognition model based on a BP neural network to perform flow pattern recognition on the bubble image to be recognized, and obtaining a flow pattern recognition result;

[0091] In this embodiment, the flow patterns in the reactor core channel under flow boiling conditions, that is, the flow pattern identification results include diffuse bubbly flow, bubbly flow, plug flow, annular flow and churning flow.

[0092] Step 6: According to different flow pattern identification results, the corresponding vapor fraction is calculated to obtain the vapor fraction value of the two-phase flow.

[0093] The present invention uses the flow pattern recognition result of the model as the condition for measuring the vapor fraction. This is to accurately calculate the vapor fraction values ​​of different flow patterns. Two different image processing algorithms are designed, namely the boundary marking method and the pixel intensity method, which are used for calculating the vapor fraction of different flow patterns. The details are as follows:

[0094] First, when the flow pattern identification results are diffuse bubbly flow, bubbly flow, plug flow and annular flow, the boundary marking method is used to calculate the vapor fraction; the boundary marking method is to mark different areas separated by the phase interface as gas phase or liquid phase, that is, the vapor fraction value of the two-phase flow is calculated according to the area of ​​different phases;

[0095] The boundary marking method is suitable for dealing with diffuse bubbly flow, bubbly flow, plug flow and annular flow. These flow types have relatively obvious gas-liquid interfaces. The different areas separated by the phase interface are marked as gas phase or liquid phase, and the vapor fraction can be calculated based on the area of ​​different phases.

[0096] Specifically, Figure 4 As shown in the figure, the boundary marking method is used to calculate the steam content, including:

[0097] After grayscale processing of the bubble image, the canny operator is used to detect the edge of the grayscale processed image to obtain the phase interface information;

[0098] Since some boundaries may not be closed, it is necessary to perform a closing operation on the grayscale processed image at the same time, that is, an operation of dilation followed by erosion, to obtain prominent interface information;

[0099] Based on the prominent interface information, the corresponding image is filled so that the gas phase is marked as white and the vapor content is 1; the liquid phase is marked as black and the vapor content is 0;

[0100] The vapor fraction corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0101] In the specific implementation, there are some differences in the details of the algorithm processing process for the four flow types of diffuse bubbly flow, bubbly flow, plug flow and annular flow. In the closing operation, a basic structural unit is needed. The bubbles in the bubbly flow image are generally small. The use of a structural unit with a 2×2 matrix size can better handle small bubbles. The bubbles in the plug flow image are generally large. It is necessary to appropriately select a larger structural unit such as a 9×9 matrix size, otherwise the bubble boundary may still not be closed. There are generally no complete bubbles in the annular flow image. Before the filling operation, it is necessary to manually add upper and lower boundaries to the image to assist in identifying bubbles. In the image processed by the boundary marking method, white represents the gas phase with a vapor content of 1, and black represents the liquid phase with a vapor content of 0. The vapor content value corresponding to each pixel in the image is counted and then averaged to obtain the vapor content of the two-phase flow in the image.

[0102] Second, when the flow pattern recognition result is a churning flow, the vapor fraction is calculated using a pixel intensity method; the pixel intensity method is to calculate the vapor fraction value of the two-phase flow based on image pixels.

[0103] The pixel intensity method is mainly suitable for processing churned flows, in which liquid and gas are mixed with each other, and there is no obvious phase interface between the two phases. The mixed part appears as different pixel intensities in the image. The pixel intensity is divided into different levels corresponding to different vapor content values, and the vapor content can be calculated based on the image pixels.

[0104] Specifically, the pixel intensity method is used to calculate the vapor content, including:

[0105] After grayscale processing of the image, the pixel intensity of the grayscale processed image is divided into six different levels according to the rules in Table 1, and each level corresponds to a vapor content value;

[0106] The vapor fraction value corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0107] Table 1 Pixel intensity levels corresponding to vapor content

[0108] Phase Gas content Steam content value Pixel Intensity Liquid Phase 0 0<＝m<75 Gas-liquid mixing 0.2 75<＝m<100 Gas-liquid mixing 0.4 100<＝m<125 Gas-liquid mixing 0.6 125<＝m<155 Gas-liquid mixing 0.8 155<＝m<190 Gas Phase 1 190<＝m<＝235

[0109] The present invention improves the recognition rate and accuracy of the flow pattern of the vapor-liquid two-phase flow by introducing an intelligent algorithm, and uses the flow pattern recognition result as a condition for measuring the vapor fraction to make the vapor fraction calculation more accurate. It has broad application prospects in the field of multiphase flow industrial applications.

[0110] Example 2

[0111] like Figure 5 As shown, the difference between this embodiment and embodiment 1 is that this embodiment provides a device for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow, the device comprising:

[0112] An acquisition unit, used for acquiring a bubble image, wherein the bubble image is a bubble image of flowing boiling;

[0113] A grayscale and binarization processing unit, used for converting the bubble image into a grayscale image and performing binarization processing to obtain a binary image containing flow information;

[0114] A feature extraction unit, used for extracting image features by using a wavelet transform method or a grayscale histogram method based on a binary image containing flow information;

[0115] A model building and training unit is used to build a multi-layer deep BP neural network, train the multi-layer deep BP neural network based on the extracted image features, and obtain a gas-liquid two-phase flow pattern recognition model based on the BP neural network;

[0116] A flow pattern recognition unit is used to perform flow pattern recognition on the bubble image to be recognized by using a vapor-liquid two-phase flow pattern recognition model based on a BP neural network to obtain a flow pattern recognition result;

[0117] The vapor content calculation unit is used to perform corresponding vapor content calculations according to different flow pattern identification results to obtain the vapor content value of the two-phase flow.

[0118] Furthermore, the steam content calculation unit includes:

[0119] The first vapor content calculation subunit is used to calculate the vapor content by using a boundary marking method when the flow pattern recognition result is a diffuse bubbly flow, a bubbly flow, a plug flow, and annular flow; the boundary marking method is to mark different areas separated by the phase interface as gas phase or liquid phase, that is, to calculate the vapor content value of the two-phase flow according to the area of ​​different phases;

[0120] The second vapor fraction calculation subunit is used to calculate the vapor fraction using a pixel intensity method when the flow pattern recognition result is a churning flow; the pixel intensity method is to calculate the vapor fraction value of the two-phase flow based on image pixels.

[0121] Furthermore, the execution process of the first steam content fraction calculation subunit is as follows:

[0122] After grayscale processing of the bubble image, the canny operator is used to detect the edge of the grayscale processed image to obtain the phase interface information;

[0123] At the same time, the grayscale processed image is closed, i.e., the operation of dilation followed by erosion, to obtain prominent interface information;

[0124] Based on the prominent interface information, the corresponding image is filled so that the gas phase is marked as white and the vapor content is 1; the liquid phase is marked as black and the vapor content is 0;

[0125] The vapor fraction corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0126] Furthermore, the execution process of the second steam content fraction calculation subunit is as follows:

[0127] After grayscale processing of the image, the pixel intensity of the grayscale processed image is divided into six different levels, each level corresponds to a vapor content value;

[0128] The vapor fraction value corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

[0129] Among them, the execution process of each unit can be executed according to the process steps of a method for obtaining the steam content of a flowing boiling vapor-liquid two-phase flow in Example 1, and will not be repeated in this embodiment.

[0130] At the same time, the present invention also provides a computer-readable storage medium, which stores a computer program, characterized in that when the computer program is executed by a processor, it implements the above-mentioned method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow.

[0131] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0132] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0133] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0134] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0135] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow, characterized in that: The method includes: Acquire a bubble image, wherein the bubble image is a bubble image of flowing boiling; Converting the bubble image into a grayscale image and performing binarization processing to obtain a binary image containing flow information; Based on the binary image containing flow information, wavelet transform or grayscale histogram is used to extract image features; A multi-layer deep BP neural network is constructed and trained based on the extracted image features to obtain a gas-liquid two-phase flow pattern recognition model based on the BP neural network. The flow pattern recognition model of vapor-liquid two-phase flow based on BP neural network is used to recognize the flow pattern of the bubble image to be recognized, and the flow pattern recognition result is obtained; According to different flow pattern identification results, the corresponding vapor fraction calculation is performed to obtain the vapor fraction value of the two-phase flow.

2. The method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to claim 1, characterized in that: The bubble image is obtained from a nuclear reactor flow boiling visualization experiment; In the nuclear reactor flow boiling visualization test, the channel is heated, and the bubbles in the channel are generated by boiling heat transfer, not by injecting air from the channel inlet.

3. The method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to claim 1, characterized in that: The multi-layer deep BP neural network includes an input layer, an intermediate layer and an output layer; The input layer is used to input a feature vector formed by effective feature parameters extracted by wavelet transform, or a feature vector formed by identification parameters extracted by probability distribution of gray levels in gray histogram; The output layer is used to output flow type recognition results.

4. The method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to claim 1, characterized in that: The flow pattern recognition results include diffuse bubbly flow, bubbly flow, plug flow, annular flow and churned flow.

5. The method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to claim 1, characterized in that: According to different flow pattern identification results, the corresponding steam content calculation is performed respectively, including: When the flow pattern identification results are diffuse bubbly flow, bubbly flow, plug flow and annular flow, the boundary marking method is used to calculate the vapor fraction; the boundary marking method is to mark different areas separated by the phase interface as gas phase or liquid phase, that is, to calculate the vapor fraction value of the two-phase flow according to the area of ​​different phases; When the flow pattern identification result is a churning flow, the vapor fraction is calculated using a pixel intensity method; the pixel intensity method is to calculate the vapor fraction value of the two-phase flow based on image pixels.

6. The method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to claim 5, characterized in that: The boundary marking method is used to calculate the steam content, including: After grayscale processing of the bubble image, the canny operator is used to detect the edge of the grayscale processed image to obtain the phase interface information; At the same time, the grayscale processed image is closed, i.e., the operation of dilation followed by erosion, to obtain prominent interface information; Based on the prominent interface information, the corresponding image is filled so that the gas phase is marked as white and the vapor content is 1; the liquid phase is marked as black and the vapor content is 0; The vapor fraction corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

7. The method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to claim 5, characterized in that: The pixel intensity method is used to calculate the vapor content, including: After grayscale processing of the image, the pixel intensity of the grayscale processed image is divided into multiple different levels, each level corresponds to a vapor content value; The vapor fraction value corresponding to each pixel in the image is counted, and then the average is calculated to obtain the vapor fraction value of the two-phase flow in the image.

8. A device for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow, characterized in that: The device includes: An acquisition unit, used for acquiring a bubble image, wherein the bubble image is a bubble image of flowing boiling; A grayscale and binarization processing unit, used for converting the bubble image into a grayscale image and performing binarization processing to obtain a binary image containing flow information; A feature extraction unit, used for extracting image features by using a wavelet transform method or a grayscale histogram method based on a binary image containing flow information; A model building and training unit is used to build a multi-layer deep BP neural network, train the multi-layer deep BP neural network based on the extracted image features, and obtain a gas-liquid two-phase flow pattern recognition model based on the BP neural network; A flow pattern recognition unit is used to perform flow pattern recognition on the bubble image to be recognized by using a vapor-liquid two-phase flow pattern recognition model based on a BP neural network to obtain a flow pattern recognition result; The vapor content calculation unit is used to perform corresponding vapor content calculations according to different flow pattern identification results to obtain the vapor content value of the two-phase flow.

9. The device for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow according to claim 8, characterized in that: The steam content calculation unit comprises: The first vapor content calculation subunit is used to calculate the vapor content by using a boundary marking method when the flow pattern recognition result is a diffuse bubbly flow, a bubbly flow, a plug flow, and annular flow; the boundary marking method is to mark different areas separated by the phase interface as gas phase or liquid phase, that is, to calculate the vapor content value of the two-phase flow according to the area of ​​different phases; The second vapor fraction calculation subunit is used to calculate the vapor fraction using a pixel intensity method when the flow pattern recognition result is a churning flow; the pixel intensity method is to calculate the vapor fraction value of the two-phase flow based on image pixels.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, a method for obtaining the vapor content of a flowing boiling vapor-liquid two-phase flow as described in any one of claims 1 to 7 is implemented.

Images