A method for detecting gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank

Abstract

A method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank includes the following steps: separating a response signal into transmission, elliptical reflection, and echo signals; obtaining a preliminary gas-liquid distribution reconstruction image; obtaining a gas-liquid distribution reconstruction image; dividing the gas-liquid distribution reconstruction image into sets of different bubble regions; obtaining a two-dimensional gas-liquid distribution image; and obtaining a three-dimensional gas-liquid distribution image. In the context of microgravity and a large tank volume, the present invention combines backprojection, elliptical reflection, and local curvature reconstruction methods to weightedly reconstruct the position of the gas-liquid boundary and the approximate distribution of the gas phase region. Blind areas within the large tank measurement area are reconstructed based on the rule that local curvature does not change abruptly. Based on the prior characteristic that the air pillow region is configured as an elliptical body of revolution, morphological processing and elliptical fitting are used to process the gas-liquid distribution image. Using simple measurement equipment, the gas-liquid distribution state within the tank is measured, which is of great significance for spacecraft propellant management, attitude control, and propellant remaining measurement.

Description

Technical Field

[0001] The present invention belongs to the technical field of on-orbit propellant container detection for spacecraft, and in particular relates to a method for detecting the gas-liquid distribution state of a cryogenic liquid fuel tank for a spacecraft. Background Art

[0002] Cryogenic liquid propellant tanks experience complex phenomena such as evaporation phase transitions, pressure fluctuations, heat transfer, and pool boiling. These phenomena generate bubbles dispersed within the liquid phase, which fuse to form a cushion region. Under microgravity, the cushion region ultimately assumes an elliptical shape of rotation. Factors such as cushion region rupture, the addition of inert gas, unstable pressure or temperature distributions, and tank oscillations can lead to complex variations in the gas-liquid distribution within the tank. The uncertainty and difficulty in accurately detecting the gas-liquid state present numerous challenges for spacecraft in-orbit operations. First, direct tank depressurization without an known gas distribution can lead to the accidental discharge of liquid propellant, resulting in serious consequences. Second, changes in the gas-liquid distribution can directly affect the position of the spacecraft's center of mass, thereby impacting attitude control. Third, locating the gas-liquid interface facilitates measurement of remaining propellant, enabling accurate prediction of the end of a satellite's lifespan.

[0003] For gas-liquid distribution measurement, existing technologies include optical methods, electrical tomography methods, and acoustic methods. Among them, optical methods and electrical tomography methods are limited by the non-transparent walls of the tank and the conductivity of the fluid, respectively, and are not suitable for gas-liquid two-phase measurement scenarios in the tank. The ultrasonic tomography method, an acoustic method, uses cyclic excitation and reception of signals through an ultrasonic transducer array. During the excitation and reception process of the array, the attenuation signal and multi-path reflection signal can be collected simultaneously without the use of other experimental devices. Unlike other process tomography methods, the sound wave will be attenuated to a large extent when it is transmitted from the tank wall to the liquid propellant. Part of the energy will be transmitted in the wall in the form of guided waves, and the other part of the acoustic signal will be transmitted, scattered attenuated, and reflected when propagating in the two-phase medium. The fusion of transmission and reflection methods can realize the reconstruction of the gas-liquid interface morphology and position of the two-phase flow. However, no effective method has been found so far. Summary of the Invention

[0004] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide a method for detecting the gas-liquid distribution status of a spacecraft cryogenic liquid fuel tank, which can monitor the three-dimensional gas-liquid distribution inside the tank and is suitable for tank measurement scenarios with large measurement areas, non-circulation and microgravity.

[0005] To achieve the above-mentioned object, the present invention provides a method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank, comprising the following steps performed in sequence:

[0006] 1) A multi-layer ultrasonic transducer array is built on the tank, and ultrasonic transducers in different channels of the ultrasonic transducer array are excited in turn, and the response signals are divided into three groups: transmission, elliptical reflection, and echo signals;

[0007] 2) The transmission, elliptical reflection, and echo signals are inverted using back-projection, elliptical reflection, and local curvature reconstruction methods, respectively. The sensitivity matrices obtained by the three methods are then weighted superimposed to obtain a preliminary reconstructed image of the gas-liquid distribution.

[0008] 3) Based on the morphological processing method, the protrusions and noise points in the preliminary gas-liquid distribution reconstruction image are removed and the missing pixels are supplemented. Then, the measurement blind area reconstruction method is used to reconstruct the measurement blind area to obtain the gas-liquid distribution reconstructed image;

[0009] 4) Dividing the gas-liquid distribution reconstructed image into different bubble region sets based on whether the gas-liquid interface is connected, and then counting and locating the bubbles in each bubble region set;

[0010] 5) After different sets of bubble regions are distinguished, an ideal two-dimensional elliptical model is used to fit different bubble cross sections in the gas-liquid distribution reconstructed image to obtain a two-dimensional gas-liquid distribution image;

[0011] 6) Based on the bubble location and number in the two-dimensional image, multiple frames of the above two-dimensional gas-liquid distribution image are superimposed on the z-axis to obtain a three-dimensional gas-liquid distribution image, thereby fitting the three-dimensional morphology of the gas-liquid interface.

[0012] In step 1), the specific method of building a multi-layer ultrasonic transducer array on the tank, alternately exciting ultrasonic transducers of different channels in the ultrasonic transducer array, and dividing the response signals into three groups of transmission, elliptical reflection and echo signals is as follows:

[0013] An ultrasonic transducer array is arranged on multiple cross sections of a tank storing propellant therein, and excitation signals are applied to ultrasonic transducers of different channels in the ultrasonic transducer array in turn, and response signals caused by the excitation are collected; for an ultrasonic transducer array with n ultrasonic transducers, the ultrasonic transducer of one channel in the ultrasonic transducer array is excited each time, and after a short delay, response signals of n channels including the ultrasonic transducer itself that is excited are collected. The ultrasonic transducers of n channels are excited in turn within one cycle, and a total of n response signals are collected. 2 Then, the response signals collected within a period are grouped, and the response signals received at relative positions are regarded as transmission signals, the response signals sent and received spontaneously are regarded as echo reflection signals, and the other received signals are regarded as elliptical reflection signals.

[0014] In step 2), the transmission, elliptical reflection and echo signals are inverted respectively by back projection, elliptical reflection method and local curvature reconstruction method, and then the sensitivity matrices obtained by the three methods are weighted superpositioned to obtain a preliminary gas-liquid distribution reconstructed image. The specific method is as follows:

[0015] 2.1 Transmission signal inversion:

[0016] When there is a partial gas phase area blocking the transmission path, the sensitivity matrix P(i, j) of the transmission method is calculated as follows:

[0017]

[0018] Among them, i and j are the channel numbers of the excitation signal, P tr0 is the calibration transmission signal when there is no obstruction, A0 is the diameter of the piezoelectric chip of the ultrasonic transducer, and A tr is the shielding area of ​​the gas phase region. The amplitude of the attenuated transmission signal can be used to infer the size of the shielding area.

[0019] 2.2. Inversion of elliptical reflection signals:

[0020] The sensitivity matrix S obtained by the elliptical reflection method el It mainly depends on the sound wave transmission time and sound pressure attenuation amplitude, and the sensitive matrix S is located at the gas-liquid interface. ij and the divergence angle constraint matrix S δ OK, the formula is:

[0021]

[0022] The gas-liquid interface positioning sensitive matrix S ij The formula is:

[0023]

[0024] Among them, l e is the amplitude parameter of the elliptical reflection, P ij is the amplitude of the elliptical reflection signal, c is the speed of sound in the liquid, t ij is the sound wave transmission time, d ij is the transmission distance of the sound wave.

[0025] The angle between the sound wave reflection path and the straight line where the ultrasonic transducer is located in the elliptical reflection method should be smaller than the divergence angle δ; when a certain channel transmits sound waves and other channels receive sound wave signals, the divergence angle constraint matrix S δ The formula is:

[0026]

[0027] Where j is the number of the receiving channel and d is the transmission distance of the sound wave.

[0028] 2.3. Echo signal inversion:

[0029] The echo signal amplitude P that can be captured by the ultrasonic transducer b The calculation formula is:

[0030]

[0031] θ is the angle of the echo signal relative to the center of curvature at the gas-liquid interface, which can be calculated by the following formula:

[0032]

[0033] Where R is the local curvature radius of the gas-liquid interface, P0 is the reference amplitude of the echo signal, which is measured by arranging a total reflection interface near the ultrasonic transducer; d0 is the diameter of the piezoelectric chip of the ultrasonic transducer. Based on the measurement signal, the sound wave transmission time t can be extracted. ii and the echo signal amplitude P b , thereby solving the local curvature radius of the gas-liquid interface; based on the local curvature radius of the gas-liquid interface, the sensitivity matrix S obtained by the local curvature reconstruction method is obtained rf The ideal fan-shaped reflection interface S needs to be calculated R (d1, R) and the straight line L(θ) forming the reflective sector gas-liquid interface, the formula is:

[0034]

[0035] Where d1 is the shortest distance between the ultrasonic transducer and the reflecting gas-liquid interface;

[0036] After processing the transmission signal, elliptical reflection signal, and echo signal respectively, the sensitivity matrices obtained by the above three methods are weightedly superimposed to obtain a preliminary gas-liquid distribution reconstructed image; in the weight parameter allocation, the weight ratio of the transmission: elliptical reflection: echo inversion method is 1:4:4.

[0037] In step 3), the morphological processing method is used to remove protrusions and noise points in the preliminary gas-liquid distribution reconstructed image and to fill in missing pixels, and then the measurement blind area reconstruction method is used to reconstruct the measurement blind area. The specific method for obtaining the gas-liquid distribution reconstructed image is as follows:

[0038] Before morphological processing, the image must be binarized. Let V th Indicates the threshold for screening the gas-liquid two-phase region, M n×n (i, j) represents the preliminary gas-liquid distribution reconstruction image matrix, and the formula for the binarization process is:

[0039]

[0040] Mean filtering is required before binarization;

[0041] Morphological processing mainly includes dilation and erosion. The dilation operation is a morphological extension operation. By comparing the relationship between pixels, pixels are added around the pixels that meet the conditions, thereby filling the missing pixels in the image. Let P and Q be sets in the two-dimensional integer space Z2. The set Q is morphologically expanded to obtain the set P. This process is defined as:

[0042]

[0043] Similarly, using set Q to erode set P is to concentrate all the pixels contained in set P into set Q, and traverse the pixels in the preliminary gas-liquid distribution reconstructed image for translation; the erosion operation is a shrinking process that removes redundant noise in the morphological image. The definition of set P eroded by set Q is:

[0044]

[0045] After morphological processing, the blind area needs to be reconstructed using the blind area reconstruction method. The curvature radius of the blind area is estimated using the curvature radius measured near the blind area. When the implicit function f(x, y) = 0 is satisfied, the curvature calculation formula of a point on the ellipse boundary is:

[0046]

[0047] The characteristic of the gas-liquid interface approaching the boundary of the ellipsoid indicates that the local curvature of the gas-liquid interface will not change drastically. Let r1 and r2 be the radii of two circular arcs of known curvature, and the curvature calculation formula of the unknown curvature segment between the two is:

[0048]

[0049] Finally, the gas-liquid distribution reconstructed image M(x,y) is obtained.

[0050] In step 4), the gas-liquid distribution reconstructed image is divided into different bubble region sets based on whether the gas-liquid interface is connected, and then the number of bubbles in each bubble region set is counted and located in the following specific method:

[0051] Assuming that the current pixel in the gas-liquid distribution reconstructed image is M(p,q), p is the horizontal coordinate of the pixel, q is the vertical coordinate of the pixel, and the total number of bubbles is k, the process of distinguishing different gas-liquid interfaces is expressed as:

[0052]

[0053] Among them, M(i,j) k A set representing different gas-liquid interface images; the pixel point first identified as the gas-liquid interface is stored in a bubble area set. When traversing to a pixel point of a new gas-liquid interface, the pixel point is called the current pixel point, and the distance between the current pixel point and all the pixels in the bubble area set is calculated. If the distance between the current pixel point and all the pixels in the bubble area set exceeds the threshold, it means that the current pixel point is far away from the gas-liquid interface and belongs to other bubbles. Therefore, a new bubble area set needs to be opened and the current pixel point is stored in it; when there is a pixel point in the existing bubble area set and the distance between the current pixel point and the pixel point is less than the threshold, it means that the two pixels are adjacent on the gas-liquid interface, and the current pixel point is included in the bubble area set; therefore, different bubbles can be distinguished by judging whether the gas-liquid interface is connected, and finally the number of bubbles in each bubble area set is counted and located.

[0054] In step 5), the specific method of fitting different bubble cross sections in the gas-liquid distribution reconstructed image with an ideal two-dimensional elliptical model to obtain a two-dimensional gas-liquid distribution image is as follows:

[0055] Under the condition that the internal gas-liquid distribution configuration approaches an ellipse, the gas-liquid distribution reconstructed image after ellipse fitting is expressed as:

[0056]

[0057] Among them, M i (x, y) is the gas-liquid distribution image reconstruction matrix before ellipse fitting, M(x, y)′ is the gas-liquid distribution image reconstruction matrix after ellipse fitting, ρ(j) is the gas-liquid interface curvature solution model, E el Post-processing operation for ellipse approximation of gas-liquid distribution reconstruction image;

[0058] Let {M u (i, j)} is the set of bubble regions in the u-th bubble region, M1 and M2 are any two boundary points in the discrete matrix two-dimensional image, d(M1, M2) is the distance between the two boundary points, and the maximum distance between any two boundary points is calculated as follows:

[0059]

[0060] Post-processing operation E of the gas-liquid interface boundary ellipse approximation by gas-liquid distribution reconstructed image el After processing, it is expressed as:

[0061]

[0062] The u-th bubble area M u (i,j) Post-processing operation E of the ellipse approximation of the gas-liquid distribution reconstructed image elAfter processing, the range of continuous solutions is obtained. After spatial discretization of the solutions, two pairs of discrete gas-liquid boundary positioning points can be obtained. On this basis, the fitting continuous function that approaches an ellipse in the u-th bubble region can be expressed as:

[0063] d 2 (M a1 ,M a2 )y 2 +d 2 (M b1 ,M b2 )=1

[0064] Among them, M a1 、M a2 and M b1 、M b2 Two sets of intersection points for locating the gas-liquid interface; the two-dimensional gas-liquid distribution image is composed of pixel points and is discrete. The continuous function curve needs to be converted into a discrete pixel point image matrix.

[0065] In step 6), the specific method of superimposing multiple frames of the above-mentioned two-dimensional gas-liquid distribution images on the z-axis to obtain a three-dimensional gas-liquid distribution image and fitting the three-dimensional morphology of the gas-liquid interface is:

[0066] Multiple layers of arrays are arranged on the tank. Each layer of array can reconstruct a frame of two-dimensional gas-liquid distribution image of the gas-liquid interface. For an array with w layers, with the layer w as the z-axis, multiple frames of two-dimensional gas-liquid distribution images are superimposed in three-dimensional space, and the three-dimensional morphological features of the gas-liquid distribution are stacked through interpolation and filtering. The three-dimensional gas-liquid distribution image matrix at this time is expressed as:

[0067]

[0068] Among them, G(z1,z2,z3) is a three-dimensional Gaussian filter matrix.

[0069] In summary, the beneficial effects of the present invention are:

[0070] In microgravity and large tanks, the position of the gas-liquid boundary and the approximate distribution of the gas phase are reconstructed using a weighted method combining back-projection, elliptical reflection, and local curvature reconstruction. Blind spots within the tank's large measurement area are reconstructed based on the principle that local curvature does not abruptly change. Morphological processing and elliptical fitting are used to process the gas-liquid distribution image, based on the prior characteristic that the air pillow region is an elliptical solid of revolution. Using simple measurement equipment, the gas-liquid distribution state within the tank is measured, which is of great significance for spacecraft propellant management, attitude control, and propellant remaining measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1A flow chart of the method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank provided by the present invention;

[0072] Figure 2 Schematic diagram of the ultrasonic transducer array arrangement and measurement system in the present invention;

[0073] Figure 3 Schematic diagram of gas-liquid interface reconstruction using elliptical reflection signals in an embodiment of the present invention;

[0074] Figure 4 Schematic diagram of gas-liquid interface reconstruction using echo reflection signals in an embodiment of the present invention;

[0075] Figure 5 Schematic diagram of gas-liquid distribution measurement blind area reconstruction in an embodiment of the present invention;

[0076] Figure 6 Schematic diagram of three-dimensional reconstruction of gas-liquid distribution in an embodiment of the present invention;

[0077] Figure 7 Schematic diagram of the three-dimensional reconstruction result of the simulation measurement device in an embodiment of the present invention. DETAILED DESCRIPTION

[0078] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0079] like Figures 1 to 6 As shown, the method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank provided by the present invention comprises the following steps performed in sequence:

[0080] 1) A multi-layer ultrasonic transducer array is built on the tank, and ultrasonic transducers in different channels of the ultrasonic transducer array are excited in turn, and the response signals are divided into three groups: transmission, elliptical reflection, and echo signals;

[0081] The specific method is:

[0082] An ultrasonic transducer array 2 is arranged on multiple cross sections of a tank 1 in which a propellant 4 is stored. Excitation signals are applied to ultrasonic transducers of different channels in the ultrasonic transducer array 2 in turn, and response signals caused by the excitation are collected. The ultrasonic transducer has the functions of transmitting excitation and capturing response signals. For an ultrasonic transducer array 2 with n ultrasonic transducers, the ultrasonic transducer of one channel in the ultrasonic transducer array 2 is excited each time, and after a short delay, response signals of n channels including the ultrasonic sensor itself that is excited are collected. The ultrasonic transducers of n channels are excited in turn within one cycle, and a total of n response signals are collected. 2 Then, the response signals collected within a period are grouped, and the response signals received at relative positions are regarded as transmission signals, the response signals sent and received spontaneously are regarded as echo reflection signals, and the other received signals are regarded as elliptical reflection signals.

[0083] 2) The transmission, elliptical reflection, and echo signals are inverted using back-projection, elliptical reflection, and local curvature reconstruction methods, respectively. The sensitivity matrices obtained by the three methods are then weighted superimposed to obtain a preliminary reconstructed image of the gas-liquid distribution.

[0084] Different types of signals are processed using different inversion methods to reconstruct the gas-liquid distribution image. The key to this process is to obtain the sensitivity matrix of the inversion method. The calculation process is:

[0085] 2.1 Transmission signal inversion:

[0086] The basic principle of the multi-channel transmission inversion method is the back-projection method, which uses the attenuation of the transmission signal to invert the distribution of the gas phase region in the acoustic wave transmission path and the obstruction of the acoustic wave path. The larger the area of ​​gas phase obstruction in the acoustic wave propagation path, the smaller the transmission signal received by the ultrasonic transducer. When there is partial gas phase obstruction in the transmission path, the sensitivity matrix P(i,j) of the transmission method is calculated as follows:

[0087]

[0088] Among them, i and j are the channel numbers of the excitation signal, P tr0 is the calibration transmission signal when there is no obstruction, A0 is the diameter of the piezoelectric chip of the ultrasonic transducer, and A tr is the shielding area of ​​the gas phase region. The amplitude of the attenuated transmission signal can be used to infer the size of the shielding area.

[0089] 2.2. Inversion of elliptical reflection signals:

[0090] like Figure 2 As shown, the position of the self-transmitting ultrasonic transducer and the time of the sound wave can be used to draw the gas-liquid interface that triggers the reflection. The sensitivity matrix S is obtained by the elliptical reflection method. elIt mainly depends on the sound wave transmission time and sound pressure attenuation amplitude, and the sensitive matrix S is located at the gas-liquid interface. ij and the divergence angle constraint matrix S δ OK, the formula is:

[0091]

[0092] The gas-liquid interface positioning sensitive matrix S ij The formula is:

[0093]

[0094] Among them, l e is the amplitude parameter of the elliptical reflection, P ij is the amplitude of the elliptical reflection signal, c is the speed of sound in the liquid, t ij is the sound wave transmission time, d ij is the transmission distance of the sound wave.

[0095] In the actual transmission process of ultrasonic waves, the divergence of the sound beam is limited. The angle between the sound wave reflection path in the elliptical reflection method and the straight line where the ultrasonic transducer is located should be smaller than the divergence angle δ. When a certain channel transmits sound waves and other channels receive sound wave signals, the divergence angle constraint matrix S δ The formula is:

[0096]

[0097] Where j is the number of the receiving channel and d is the transmission distance of the sound wave.

[0098] 2.3. Echo signal inversion:

[0099] like Figure 3 As shown in Figure 1, when a sound beam with a certain width is transmitted to the gas-liquid interface, the sound wave will be reflected in other directions based on the normal line of the gas-liquid interface. Since the gas-liquid interface has a certain curvature, the echo sound beam will diverge. The self-transmitting and self-receiving ultrasonic transducer can only capture a part of the energy in the initial excitation signal. The echo signal amplitude P that the ultrasonic transducer can capture is b The calculation formula is:

[0100]

[0101] θ is the angle of the echo signal relative to the center of curvature at the gas-liquid interface, which can be calculated by the following formula:

[0102]

[0103] Where R is the local curvature radius of the gas-liquid interface, P0 is the reference amplitude of the echo signal, which is measured by arranging a total reflection interface near the ultrasonic transducer; d0 is the diameter of the piezoelectric chip of the ultrasonic transducer. Based on the measurement signal, the sound wave transmission time t can be extracted. ii and the echo signal amplitude P b , thereby solving the local curvature radius of the gas-liquid interface; based on the local curvature radius of the gas-liquid interface, the sensitivity matrix S obtained by the local curvature reconstruction method is obtained rf The ideal fan-shaped reflection interface S needs to be calculated R (d1, R) and the straight line L(θ) forming the reflective sector gas-liquid interface, the formula is:

[0104]

[0105] Where d1 is the shortest distance between the ultrasonic transducer and the reflecting gas-liquid interface. The local curvature reconstruction method reconstructs the fan-shaped gas-liquid interface that triggers the acoustic wave reflection based on the echo signal amplitude and the acoustic wave transmission time information. The larger the radius of the gas-liquid interface curvature, the stronger the echo signal energy.

[0106] After processing the transmission signal, elliptical reflection signal, and echo signal separately, the sensitivity matrices obtained by the three methods are weighted and superimposed to produce a preliminary reconstructed image of the gas-liquid distribution. The superposition weights of the different reconstruction methods reflect their priority. For example, when the transmission signal identifies a region free of bubbles, the back-projection method takes precedence over the elliptical reflection method and the local curvature reconstruction method. The weight parameter allocation for the transmission, elliptical reflection, and echo inversion methods is 1:4:4.

[0107] 3) Based on the morphological processing method, the protrusions and noise points in the preliminary gas-liquid distribution reconstruction image are removed and the missing pixels are supplemented. Then, the measurement blind area reconstruction method is used to reconstruct the measurement blind area to obtain the gas-liquid distribution reconstructed image;

[0108] Morphological operations determine the image parameters of a pixel based on the relationship between it and its surrounding pixels. This relationship can be compared to the convolution operator used to traverse a frame of an image. Before morphological processing, the image must be binarized. Let V th Indicates the threshold for screening the gas-liquid two-phase region, M n×n (i, j) represents the preliminary gas-liquid distribution reconstruction image matrix, and the formula for the binarization process is:

[0109]

[0110] The value of the binarization threshold is related to the numerical distribution of the parameters of the combined reconstruction probability matrix. Since the parameters of the superimposed local area of ​​the weight matrix may be too large, the area of ​​the binarized area is too small. Therefore, it is necessary to perform mean filtering before binarization.

[0111] Morphological processing mainly includes dilation and erosion, which are used to remove bulges and noise in the preliminary gas-liquid distribution reconstruction image or fill in missing pixels. The dilation operation is performed by morphological extension. By comparing the relationship between pixels, pixels are added around the pixels that meet the conditions, thereby filling in the missing pixels in the image. Let P and Q be sets in the two-dimensional integer space Z2. The set Q is morphologically expanded to obtain the set P. This process is defined as:

[0112]

[0113] Similarly, using set Q to erode set P means concentrating all the pixels contained in set P into set Q and traversing the pixels in the preliminary gas-liquid distribution reconstructed image for translation. The erosion operation is a shrinking process that removes redundant noise in the morphological image. The definition of set P eroded by set Q is:

[0114]

[0115] like Figure 5 As shown in Figure 1, after morphological processing, the blind area needs to be reconstructed using the blind area reconstruction method. The curvature radius measured near the blind area is used to estimate the curvature radius at the blind area. When the implicit function f(x, y) = 0 is satisfied, the curvature calculation formula of a point on the ellipse boundary is:

[0116]

[0117] The characteristic of the gas-liquid interface approaching the boundary of the ellipsoid indicates that the local curvature of the gas-liquid interface will not change drastically. Let r1 and r2 be the radii of two circular arcs of known curvature, and the curvature calculation formula of the unknown curvature segment between the two is:

[0118]

[0119] Finally, the gas-liquid distribution reconstructed image M(x,y) is obtained.

[0120] 4) Dividing the gas-liquid distribution reconstructed image into different bubble region sets based on whether the gas-liquid interface is connected, and then counting and locating the bubbles in each bubble region set;

[0121] Assuming that the current pixel in the gas-liquid distribution reconstructed image is M(p,q), p is the horizontal coordinate of the pixel, q is the vertical coordinate of the pixel, and the total number of bubbles is k, the process of distinguishing different gas-liquid interfaces is expressed as:

[0122]

[0123] Among them, M(i,j) k A set of images representing different gas-liquid interfaces. The first pixel identified as a gas-liquid interface is stored in a bubble region set. When a new gas-liquid interface pixel is traversed, this pixel is called the current pixel. The distance between the current pixel and all pixels in the bubble region set is calculated. If the distance between the current pixel and all pixels in the bubble region set exceeds a threshold, it means that the current pixel is far away from the gas-liquid interface and belongs to another bubble. Therefore, a new bubble region set needs to be opened and the current pixel stored there. If the distance between a pixel in the existing bubble region set and the current pixel is less than a threshold, it means that the two pixels are adjacent on the gas-liquid interface, and the current pixel is included in the bubble region set. Therefore, by determining whether the gas-liquid interface is connected, different bubbles can be distinguished. Finally, the number of bubbles in each bubble region set is counted and located.

[0124] 5) After different sets of bubble regions are distinguished, an ideal two-dimensional elliptical model is used to fit different bubble cross sections in the gas-liquid distribution reconstructed image to obtain a two-dimensional gas-liquid distribution image;

[0125] The primary measurement target inside tank 1 is the air pillow region 3. The final configuration of the air pillow region 3 approximates an elliptical body of revolution, with an elliptical cross-section. Therefore, the gas-liquid interface images of different bubble regions can be considered ideal elliptical models during imaging. Under the condition that the internal gas-liquid distribution configuration approaches an ellipse, the reconstructed gas-liquid distribution image after elliptical fitting is expressed as:

[0126]

[0127] Among them, M i (x, y) is the gas-liquid distribution image reconstruction matrix before ellipse fitting, M(x, y)′ is the gas-liquid distribution image reconstruction matrix after ellipse fitting, ρ(j) is the gas-liquid interface curvature solution model, E el Post-processing operation for ellipse approximation of gas-liquid distribution reconstruction image;

[0128] Let {M u (i, j)} is the set of bubble regions in the u-th bubble region, M1 and M2 refer to any two boundary points in the discrete matrix two-dimensional image, d(M1, M2) is the distance between the two boundary points, and the maximum distance between any two boundary points is calculated as follows:

[0129]

[0130] Post-processing operation E of the gas-liquid interface boundary ellipse approximation by gas-liquid distribution reconstructed image el After processing, it can be expressed as:

[0131]

[0132] The u-th bubble area M u (i,j) Post-processing operation E of the ellipse approximation of the gas-liquid distribution reconstructed image el After processing, the range of continuous solutions is obtained. After spatial discretization of the solutions, two pairs of discrete gas-liquid boundary positioning points can be obtained. On this basis, the fitting continuous function that approaches an ellipse in the u-th bubble region can be expressed as:

[0133] d 2 (M a1 ,M a2 )y 2 +d 2 (M b1 ,M b2 )=1

[0134] Among them, M a1 、M a2 and M b1 、M b2 The two sets of intersection points are used to locate the gas-liquid interface. The two-dimensional gas-liquid distribution image is composed of individual pixels and is discrete. The continuous function curve needs to be converted into a discrete pixel image matrix.

[0135] 6) Superimposing multiple frames of the above two-dimensional gas-liquid distribution images on the z-axis to obtain a three-dimensional gas-liquid distribution image, thereby fitting the three-dimensional morphology of the gas-liquid interface.

[0136] like Figure 6 As shown, multiple layers of arrays are arranged on the tank 1. Each layer of the array can reconstruct a frame of two-dimensional gas-liquid distribution image of the gas-liquid interface. For an array with w layers, the number of layers w is used as the z-axis. Multiple frames of two-dimensional gas-liquid distribution images are superimposed in three-dimensional space, and the three-dimensional morphological features of the gas-liquid distribution are stacked through interpolation and filtering. The three-dimensional gas-liquid distribution image matrix at this time is expressed as:

[0137]

[0138] Among them, G(z1, z2, z3) is a three-dimensional Gaussian filter matrix. The acquisition interval of each frame image is the same. The shorter the interval, the higher the density of the three-dimensional gas-liquid distribution image superposition, and the closer the three-dimensional morphology reconstruction result will be to the real gas-liquid interface.

[0139] like Figure 7As shown in the figure, in the ground experimental environment, the 3D reconstruction effect of the simulated measurement device is:

[0140] A weight is used to fix the airbag as the measurement target, replacing the air bubble to be measured in the tank, and water is used instead of liquid propellant. The transducer array is fixed, and the airbag is moved at a constant speed through the ultrasonic transducer array to obtain reconstructed images of different cross-sections of the airbag. 50 frames of images are superimposed on the z-axis, and the reconstruction results reflect the morphology of the airbag. The 3D reconstruction results, 3D reconstruction side view, and reconstructed experimental airbag are shown as follows: Figure 7 As shown in (a), (b), and (c).

[0141] The above description enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank, characterized in that: The method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank comprises the following steps performed in sequence: 1) A multi-layer ultrasonic transducer array is built on the tank, and ultrasonic transducers in different channels of the ultrasonic transducer array are excited in turn, and the response signals are divided into three groups: transmission, elliptical reflection, and echo signals; 2) The transmission, elliptical reflection, and echo signals are inverted using back-projection, elliptical reflection, and local curvature reconstruction methods, respectively. The sensitivity matrices obtained by the three methods are then weighted superimposed to obtain a preliminary reconstructed image of the gas-liquid distribution. 3) Based on the morphological processing method, the protrusions and noise points in the preliminary gas-liquid distribution reconstruction image are removed and the missing pixels are supplemented. Then, the measurement blind area reconstruction method is used to reconstruct the measurement blind area to obtain the gas-liquid distribution reconstructed image; 4) Dividing the gas-liquid distribution reconstructed image into different bubble region sets based on whether the gas-liquid interface is connected, and then counting and locating the bubbles in each bubble region set; 5) After different sets of bubble regions are distinguished, an ideal two-dimensional elliptical model is used to fit different bubble cross sections in the gas-liquid distribution reconstructed image to obtain a two-dimensional gas-liquid distribution image; 6) Superimposing multiple frames of the above two-dimensional gas-liquid distribution images on the z-axis to obtain a three-dimensional gas-liquid distribution image, thereby fitting the three-dimensional morphology of the gas-liquid interface.

2. The method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank according to claim 1, characterized in that: In step 1), the specific method of building a multi-layer ultrasonic transducer array on the tank, alternately exciting ultrasonic transducers of different channels in the ultrasonic transducer array, and dividing the response signals into three groups of transmission, elliptical reflection and echo signals is as follows: Ultrasonic transducer arrays (2) are arranged on multiple cross sections of a tank (1) in which a propellant (4) is stored, and excitation signals are applied to ultrasonic transducers of different channels in the ultrasonic transducer array (2) in turn, and response signals caused by the excitation are collected; for an ultrasonic transducer array (2) with n ultrasonic transducers, the ultrasonic transducer of one channel in the ultrasonic transducer array (2) is excited each time, and after a short delay, the response signals of n channels including the ultrasonic transducer itself that is excited are collected, and the ultrasonic transducers of n channels are excited in turn in one cycle, and a total of n response signals are collected. 2 Then, the response signals collected within a period are grouped, and the response signals received at relative positions are regarded as transmission signals, the response signals sent and received spontaneously are regarded as echo reflection signals, and the other received signals are regarded as elliptical reflection signals.

3. The method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank according to claim 1, characterized in that: In step 2), the transmission, elliptical reflection and echo signals are inverted respectively by back projection, elliptical reflection method and local curvature reconstruction method, and then the sensitivity matrices obtained by the three methods are weighted superpositioned to obtain a preliminary gas-liquid distribution reconstructed image. The specific method is as follows: 2.1 Transmission signal inversion: When there is a partial gas phase area blocking the transmission path, the sensitivity matrix P(i, j) of the transmission method is calculated as follows: Among them, i and j are the channel numbers of the excitation signal, P tr0 is the calibration transmission signal when there is no obstruction, A0 is the diameter of the piezoelectric chip of the ultrasonic transducer, and A tr is the shielding area of ​​the gas phase region. The amplitude of the attenuated transmission signal can be used to infer the size of the shielding area. 2.

2. Inversion of elliptical reflection signals: The sensitivity matrix S obtained by the elliptical reflection method el It mainly depends on the sound wave transmission time and sound pressure attenuation amplitude, and the sensitive matrix S is located at the gas-liquid interface. ij and the divergence angle constraint matrix S δ OK, the formula is: The gas-liquid interface positioning sensitive matrix S ij The formula is: Among them, l e is the amplitude parameter of the elliptical reflection, P ij is the amplitude of the elliptical reflection signal, c is the speed of sound in the liquid, t ij is the sound wave transmission time, d ij is the transmission distance of the sound wave; The angle between the sound wave reflection path and the straight line where the ultrasonic transducer is located in the elliptical reflection method should be smaller than the divergence angle δ; when a certain channel transmits sound waves and other channels receive sound wave signals, the divergence angle constraint matrix S δ The formula is: Where j is the number of the receiving channel and d is the transmission distance of the sound wave; 2.

3. Echo signal inversion: The echo signal amplitude P that can be captured by the ultrasonic transducer b The calculation formula is: θ is the angle of the echo signal relative to the center of curvature at the gas-liquid interface, which can be calculated by the following formula: Where R is the local curvature radius of the gas-liquid interface, P0 is the reference amplitude of the echo signal, which is measured by arranging a total reflection interface near the ultrasonic transducer; d0 is the diameter of the piezoelectric chip of the ultrasonic transducer. Based on the measurement signal, the sound wave transmission time t can be extracted. ii and the echo signal amplitude P b , thereby solving the local curvature radius of the gas-liquid interface; based on the local curvature radius of the gas-liquid interface, the sensitivity matrix S obtained by the local curvature reconstruction method is obtained rf The ideal fan-shaped reflection interface S needs to be calculated R (d1, R) and the straight line L(θ) forming the reflective sector gas-liquid interface, the formula is: Where d1 is the shortest distance between the ultrasonic transducer and the reflecting gas-liquid interface; After processing the transmission signal, elliptical reflection signal, and echo signal respectively, the sensitivity matrices obtained by the above three methods are weightedly superimposed to obtain a preliminary gas-liquid distribution reconstructed image; in the weight parameter allocation, the weight ratio of the transmission: elliptical reflection: echo inversion method is 1:4:

4.

4. The method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank according to claim 1, characterized in that: In step 3), the morphological processing method is used to remove protrusions and noise points in the preliminary gas-liquid distribution reconstructed image and to fill in missing pixels, and then the measurement blind area reconstruction method is used to reconstruct the measurement blind area. The specific method for obtaining the gas-liquid distribution reconstructed image is as follows: Before morphological processing, the image must be binarized. Let V th Indicates the threshold for screening the gas-liquid two-phase region, M n×n (i, j) represents the preliminary gas-liquid distribution reconstruction image matrix, and the formula for the binarization process is: Mean filtering is required before binarization; Morphological processing mainly includes dilation and erosion. The dilation operation is a morphological extension operation. By comparing the relationship between pixels, pixels are added around the pixels that meet the conditions, thereby filling the missing pixels in the image. Let P and Q be sets in the two-dimensional integer space Z2. The set Q is morphologically expanded to obtain the set P. This process is defined as: Similarly, using set Q to erode set P is to concentrate all the pixels contained in set P into set Q, and traverse the pixels in the preliminary gas-liquid distribution reconstructed image for translation; the erosion operation is a shrinking process that removes redundant noise in the morphological image. The definition of set P eroded by set Q is: After morphological processing, the blind area needs to be reconstructed using the blind area reconstruction method. The curvature radius of the blind area is estimated using the curvature radius measured near the blind area. When the implicit function f(x, y) = 0 is satisfied, the curvature calculation formula of a point on the ellipse boundary is: The characteristic of the gas-liquid interface approaching the boundary of the ellipsoid indicates that the local curvature of the gas-liquid interface will not change drastically. Let r1 and r2 be the radii of two circular arcs of known curvature, and the curvature calculation formula of the unknown curvature segment between the two is: Finally, the gas-liquid distribution reconstructed image M(x,y) is obtained.

5. The method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank according to claim 1, characterized in that: In step 4), the gas-liquid distribution reconstructed image is divided into different bubble region sets based on whether the gas-liquid interface is connected, and then the number of bubbles in each bubble region set is counted and located in the following specific method: Assuming that the current pixel in the gas-liquid distribution reconstructed image is M(p,q), p is the horizontal coordinate of the pixel, q is the vertical coordinate of the pixel, and the total number of bubbles is k, the process of distinguishing different gas-liquid interfaces is expressed as: Among them, M(i,j) k A set representing different gas-liquid interface images; the pixel point first identified as the gas-liquid interface is stored in a bubble area set. When traversing to a pixel point of a new gas-liquid interface, the pixel point is called the current pixel point, and the distance between the current pixel point and all the pixels in the bubble area set is calculated. If the distance between the current pixel point and all the pixels in the bubble area set exceeds the threshold, it means that the current pixel point is far away from the gas-liquid interface and belongs to other bubbles. Therefore, a new bubble area set needs to be opened and the current pixel point is stored in it; when there is a pixel point in the existing bubble area set and the distance between the current pixel point and the pixel point is less than the threshold, it means that the two pixels are adjacent on the gas-liquid interface, and the current pixel point is included in the bubble area set; therefore, different bubbles can be distinguished by judging whether the gas-liquid interface is connected, and finally the number of bubbles in each bubble area set is counted and located.

6. The method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank according to claim 1, characterized in that: In step 5), the specific method of fitting different bubble cross sections in the gas-liquid distribution reconstructed image with an ideal two-dimensional elliptical model to obtain a two-dimensional gas-liquid distribution image is as follows: Under the condition that the internal gas-liquid distribution configuration approaches an ellipse, the gas-liquid distribution reconstructed image after ellipse fitting is expressed as: Among them, M i (x, y) is the gas-liquid distribution image reconstruction matrix before ellipse fitting, M(x, y)′ is the gas-liquid distribution image reconstruction matrix after ellipse fitting, ρ(j) is the gas-liquid interface curvature solution model, E el Post-processing operation for ellipse approximation of gas-liquid distribution reconstruction image; Let {M u (i, j)} is the set of bubble regions in the u-th bubble region, M1 and M2 refer to any two boundary points in the discrete matrix two-dimensional image, d(M1, M2) is the distance between the two boundary points, and the maximum distance between any two boundary points is calculated as follows: Post-processing operation E of the gas-liquid interface boundary ellipse approximation by gas-liquid distribution reconstructed image el After processing, it is expressed as: The u-th bubble area M u (i,j) Post-processing operation E of the ellipse approximation of the gas-liquid distribution reconstructed image el After processing, the range of continuous solutions is obtained. After spatial discretization of the solutions, two pairs of discrete gas-liquid boundary positioning points can be obtained. On this basis, the fitting continuous function that approaches an ellipse in the u-th bubble region can be expressed as: d 2 (M a1 ,M a2 )y 2 +d 2 (M b1 ,M b2 )=1 Among them, M a1 、M a2 and M b1 、M b2 Two sets of intersection points for locating the gas-liquid interface; the two-dimensional gas-liquid distribution image is composed of pixel points and is discrete. The continuous function curve needs to be converted into a discrete pixel point image matrix.

7. The method for detecting the gas-liquid distribution state of a spacecraft cryogenic liquid fuel tank according to claim 1, characterized in that: In step 6), the specific method of superimposing multiple frames of the above-mentioned two-dimensional gas-liquid distribution images on the z-axis to obtain a three-dimensional gas-liquid distribution image and fitting the three-dimensional morphology of the gas-liquid interface is: A multi-layer array is arranged on the tank (1), and each layer of the array can reconstruct a frame of a two-dimensional gas-liquid distribution image of the gas-liquid interface. For an array with w layers, the layer w is used as the z-axis, and multiple frames of two-dimensional gas-liquid distribution images are superimposed in a three-dimensional space, and the three-dimensional morphological features of the gas-liquid distribution are stacked through interpolation and filtering. The three-dimensional gas-liquid distribution image matrix at this time is expressed as: Among them, G(z1,z2,z3) is a three-dimensional Gaussian filter matrix.

Images