LU508134B1

LU508134B1 - Method and system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with clsvo

Info

Publication number: LU508134B1
Application number: LU508134A
Authority: LU
Inventors: Feng Xie; Dongfeng Li; Yuzhou Zhang; Jianyong Hu; Zhenzhu Meng; Chengju Shan; Dan Liu; Yu Zhou; Jiachun Liu; Hong Liu
Original assignee: Univ Zhejiang Water Resources & Electric Power
Priority date: 2024-07-22
Filing date: 2024-08-29
Publication date: 2025-04-25
Also published as: CN119004167A

Abstract

The present invention discloses a method and system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with coupled level-set and volume-of-fluid (CLSVOF), and relates to the technical field water hammer analysis. The method includes: acquiring random distribution images of bubble swarms in a pipeline, and calculating a random distribution of real-time bubble swarms; constructing a CLSVOF model and calculating information data within each spatial unit under random distribution; constructing an analysis model of the water hammer caused by stagnant gas by the above data; optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient; and performing a real-time prediction and analysis on a phenomenon of the water hammer caused by stagnant gas. The present invention reduces the time complexity by calculating the random distribution of the real-time bubble swarms, considers the correlation on the time series, and improves the stability of the network and the accuracy of the prediction; the characteristic information of the pre-stage images in the continuous images is used for post-stage images to reduce the parameters and computational complexity of the network model; and combined with the time compensation coefficient, the analysis model of the water hammer caused by stagnant gas is constructed and optimized to predict and analyze a phenomenon of the water hammer by stagnant gas in real-time and accurately.

Description

BL-5904

METHOD AND SYSTEM FOR ANALYZING WATER LUS08134

HAMMER CAUSED BY STAGNANT GAS BY COUPLING

REAL GAS CHARACTERISTICS WITH CLSVO

TECHNICAL FIELD

[01] The present invention relates to the technical field of water hammer analysis, and in particular, to a method and system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with coupled level-set and volume-of-fluid (CLSVOF).

BACKGROUND ART

[02] The construction of long-distance pressurized water transmission projects across river basins and regions is an effective method to alleviate the contradiction between supply and demand of social water resources. In order to solve water hammer damage produced in the process of the operation of a long-distance pressurized water transmission pipeline system, the research on theoretical analysis, numerical simulation and physical model test of water hammer caused by stagnant gas, gas-liquid two-phase flow and pipe burst caused by stagnant gas in the long-distance pressurized water transmission pipeline system is reviewed.

[03] According to the basic thought of a CLSVOF method, volume-of-fluid (VOF) and Level Set methods are combined to track the phase interface of incompressible two-phase flow, which overcomes the shortcomings of false velocity potential and difficulty in accurately calculating the normal and curvature of the interface when using the VOF method to solve the momentum transport equation. However, in terms of the water hammer caused by stagnant gas, the research on stagnant water hammer caused by small volume bubbles is not sufficient; and in terms of gas-liquid two-phase flow, there are few studies on the water hammer caused by stagnant gas based on bubble random distribution theory, and a calculation method of stagnant water hammer based on random distribution theory has not been established.

[04] Therefore, how to propose a method and system for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF, and optimize analysis effects of the water hammer by stagnant gas by the random distribution of bubbles is an urgent problem for technicians in this field.

SUMMARY

[05] In view of this, the present invention provides a method and system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF, and optimizes the analysis effects of the water hammer caused by stagnant gas by a random distribution of bubbles. In order to achieve the above objectives, the present invention adopts the following technical solutions.

[06] A method for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF includes: 1

BL-5904

[07] acquiring random distribution images of bubble swarms in a pipeline and LU508134 preprocessing, and calculating a random distribution of real-time bubble swarms based on preprocessed images;

[08] constructing a CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set;

[09] calculating interface locations and volume fraction information of different phases within each spatial unit under random distribution by the CLSVOF model,

[10] constructing an analysis model of the water hammer caused by stagnant gas by the random distribution of bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution;

[11] optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient; and

[12] performing a real-time prediction and analysis on a phenomenon of the water hammer caused by stagnant gas by an optimized analysis model of the water hammer caused by stagnant gas.

[13] Optionally, the acquiring random distribution images of bubble swarms in a pipeline and preprocessing includes: capturing videos on a distribution of irregular real-time bubble swarms in the pipeline using a high-speed camera device by a visual window; and performing real-time image frame extraction and noise reduction processing on captured videos, selecting stretching area in morphology, and obtaining images of real-time bubble swarms.

[14] Optionally, the calculating a random distribution of real-time bubble swarms based on preprocessed images includes:

[15] capturing images on the real-time bubble swarms, to obtain the images of the real-time bubble swarms;

[16] extracting a centroid coordinate of each irregular bubble in the real-time bubble swarms, and performing a normalized processing on coordinates; and

[17] determining a first uniformity of bubble mixing in the real-time bubble swarms by adopting an average distance method based on a spatial distribution of the bubble swarms and normalized centroid coordinates.

[18] Optionally, the method further includes:

[19] collecting a number or density data of bubbles in each area of the real-time bubble swarms;

[20] fitting data into a Poisson distribution model, and calculating a mean occurrence rate of a Poisson distribution;

[21] using a Chi-square test to test a fitted Poisson distribution, to verify whether the data conforms to an assumption of the Poisson distribution;

[22] determining coincidence level between the fitted Poisson distribution and observed data by model comparison, and evaluating a second uniformity and Poissonity of the bubble distribution; and

[23] obtaining the random distribution of the real-time bubble swarms by coupling the first uniformity, the second uniformity with the Poissonity. 2

BL-5904

[24] Optionally, the constructing a CLSVOF model for solving an incompressible LU508134 two-phase flow phase interface by coupling VOF with Level Set includes:

[25] initializing a Level Set function and a VOF function, respectively;

[26] updating a velocity field and a pressure field of functions by solving pressure correction equations and momentum equations after initialization;

[27] solving a characterization function convective transportation governing equation after the update;

[28] reconstructing the phase interface according to solution results, and re-initializing the Level Set function; and

[29] obtaining the CLSVOF model used for solving the incompressible two-phase flow phase interface by cyclic iteration.

[30] Optionally, the reconstructing the phase interface according to solution results includes:

[31] obtaining a discrete VOF function and a continuous Level Set function according to the solution results;

[32] reconstructing the interface by adopting a piecewise linear interface reconstruction method, moving the interface in a normal direction to make an area ratio of unit liquid region match a VOF function value, calculating VOF function flux between two adjacent units, and iteratively solving a vertical distance from a center of a unit to the phase interface by a secant method; and

[33] re-initializing the Level Set function into a distance function by using a reconstructed phase interface.

[34] Optionally, the re-initializing the Level Set function includes: determining function symbols and values of the Level Set function, and controlling a Level Set function value of each computing unit to be a minimum value among a distance from a center of a unit to an edge, a distance from a center of a unit to a vertex of a unit, a distance from a center of a unit to an intersection point of the phase interface and a grid edge, and a distance from a center of a unit to a projection point of the phase interface.

[35] Optionally, the constructing an analysis model of water hammer caused by stagnant gas includes:

[36] preprocessing acquired data in the pipeline to obtain the random distribution of the bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution;

[37] performing a normalized processing on preprocessed data and dividing datasets;

[38] constructing and training an analysis model of the water hammer caused by stagnant gas based on a recurrent neural network according to divided datasets; and

[39] acquiring real-time data in the pipeline as samples to be tested, inputting the samples to be tested into the analysis model of the water hammer caused by stagnant gas, and obtaining predictive values of the interface locations and the volume fraction information of different phases at the next moment in the pipeline.

[40] Optionally, the optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient includes: 3

BL-5904

[41] acquiring time information based on the captured videos according to the time LU508134 series;

[42] acquiring and normalizing the time information, a temperature compensation coefficient and a machine compensation coefficient to obtain an eigenvalue cluster;

[43] calculating by using a pre-trained fuzzy neural network based on the eigenvalue cluster, to obtain a time compensation coefficient; and

[44] optimizing memory units and weights of the analysis model of the water hammer caused by stagnant gas by using the time compensation coefficient.

[45] Optionally, a system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF includes:

[46] an acquisition module, which is configured to acquire random distribution images of bubble swarms in a pipeline;

[47] a preprocessing module, which is configured to preprocess, and calculate a random distribution of real-time bubble swarms based on preprocessed images;

[48] a CLSVOF model construction module, which is configured to construct a

CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set;

[49] a calculation module, which is configured to calculate interface locations and volume fraction information of different phases within each spatial unit under random distribution by the CLSVOF model;

[50] an analysis model construction module, which is configured to construct an analysis model of the water hammer caused by stagnant gas by the random distribution of the bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution;

[51] an optimization module, which is configured to optimize the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient; and

[52] a prediction and analysis module, which is configured to perform a real-time prediction and analysis on a phenomenon of the water hammer caused by stagnant gas by an optimized analysis model of the water hammer caused by stagnant gas.

[53] It can be known from the above technical solutions that, compared with the prior art, the present invention provides a method and system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF, which has the following beneficial effects:

[54] the present invention proposes a method for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF, including: acquiring random distribution images of bubble swarms in a pipeline and preprocessing, and calculating a random distribution of real-time bubble swarms based on preprocessed images; constructing a CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set; calculating interface locations and volume fraction information of different phases within each spatial unit under random distribution by the CLSVOF model; constructing an analysis model of water hammer caused by stagnant gas by the random distribution of the bubble swarms, gas 4

BL-5904 characteristics, and the interface locations and the volume fraction information of LUS08134 different phases within each spatial unit under random distribution; optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient, and performing a real-time prediction and analysis on a water hammer caused by stagnant gas phenomenon by an optimized analysis model of the water hammer caused by stagnant gas. By calculating the random distribution of the real-time bubble swarms and considering the spatial distribution of bubbles, the present invention reduces the complexity of time and effectively reduces the error of local inhomogeneity and overall inhomogeneity; and the present invention considers the correlation on time series and learns the correlation information of time prior, thereby improving the stability of the network and the accuracy of prediction. At the same time, in the bubble swarms detection under the pipeline environment of continuous scenes, characteristic information of pre-stage images in the continuous images is used for post-stage images to reduce parameters and computational complexity of the network model; and combined with the time compensation coefficient, the analysis model of the water hammer caused by stagnant gas is constructed and optimized, and the optimized analysis model of the water hammer caused by stagnant gas is used to predict and analyze the phenomenon of the water hammer by stagnant gas in real-time and accurately.

BRIEF DESCRIPTION OF THE DRAWINGS

[55] To more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the drawings required to be used in the description of the embodiments or the prior art are briefly introduced below. It is obvious that the drawings in the description below are merely embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings according to the drawings provided without creative efforts.

[56] FIG 1 is a flow chart schematic diagram of a method for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF provided by the present invention;

[57] FIG 2 is a structural frame diagram of a system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF provided by the present invention; and

[58] FIG 3 is a schematic diagram of a CLSVOF model provided by the present invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[59] The following clearly and completely describes the technical solutions in the embodiments of the present invention with reference to drawings in the embodiments of the present invention. It is clear that the described embodiments are merely a part rather than all of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

BL-5904

[60] An embodiment of the present discloses a method for analyzing water hammer LU508134 caused by stagnant gas by coupling real gas characteristics with CLSVOF, as shown in

FIG 1, inculding:

[61] acquiring random distribution images of bubble swarms in a pipeline and preprocessing, and calculating a random distribution of real-time bubble swarms based on preprocessed images;

[62] constructing a CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set;

[63] calculating interface locations and volume fraction information of different phases within each spatial unit under random distribution by the CLSVOF model,

[64] constructing an analysis model of the water hammer caused by stagnant gas by the random distribution of bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution;

[65] optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient; and

[66] performing a real-time prediction and analysis on a phenomenon of the water hammer caused by stagnant gas by an optimized analysis model of the water hammer caused by stagnant gas.

[67] Further, the acquiring random distribution images of bubble swarms in a pipeline and preprocessing includes: capturing videos on a distribution of irregular real-time bubble swarms in the pipeline using a high-speed camera device by a visual window; and performing real-time image frame extraction and noise reduction processing on captured videos, selecting stretching area in morphology, and obtaining images of real-time bubble swarms.

[68] Further, the calculating a random distribution of real-time bubble swarms based on preprocessed images includes:

[69] capturing images on the real-time bubble swarms, to obtain the images of the real-time bubble swarms;

[70] extracting a centroid coordinate of each irregular bubble in the real-time bubble swarms, and performing a normalized processing on coordinates; and

[71] determining a first uniformity of bubble mixing in the real-time bubble swarms by adopting an average distance method based on a spatial distribution of the bubble swarms and normalized centroid coordinates.

[72] Specifically, the normalized processing includes:

[73] performing top-hat transform and filtered transformation preprocessing on the images of the real-time bubble swarms, and segmenting processed images of the real-time bubble bubble swarms, to obtain a series of gray images;

[74] converting the gray images into black and white images, and using a threshold to perform converting into binary images; and

[75] finding the centroid coordinate of each irregular bubble in the real-time bubble swarms based on the binary images, and performing the normalized processing on coordinates. 6

BL-5904

[76] Specifically, the determining a first uniformity of bubble mixing in the real-time LU508134 bubble swarms using an average distance method includes:

[77] calculating an average distance of random points when two-dimensional images reach a uniform state by using a super-cube line selection method in computational geometry theory, and taking the average distance as a standard reference value; and

[78] obtaining an average distance between random irregular bubbles based on the spatial distribution of the bubble swarms and normalized centroid coordinates, and comparing the average distance with the standard reference value, to determine the first uniformity of bubble mixing in the real-time bubble swarms.

[79] By the average distance method, the mixing uniformity of random irregular bubble swarms is effectively determined; and considering the spatial distribution of the bubbles, the complexity of time is reduced, and the error of local inhomogeneity and overall inhomogeneity is effectively reduced.

[80] Further, the method further includes:

[81] collecting a number or density data of bubbles in each area of the real-time bubble swarms;

[82] fitting data into a Poisson distribution model, and calculating a mean occurrence rate of a Poisson distribution;

[83] using a Chi-square test to test a fitted Poisson distribution, to verify whether the data conforms to an assumption of the Poisson distribution;

[84] determining coincidence level between the fitted Poisson distribution and observed data by model comparison, and evaluating a second uniformity and Poissonity of the bubble distribution; and

[85] obtaining the random distribution of the real-time bubble swarms by coupling the first uniformity, the second uniformity with the Poissonity, and performing weighted summation on the first uniformity, the second uniformity and the Poissonity, to obtain parameter values characterizing the random distribution of real-time bubble swarms.

[86] Further, as shown in FIG 3, the constructing a CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set includes:

[87] initializing a Level Set function and a VOF function, respectively;

[88] updating a velocity field and a pressure field of functions after initialization by solving pressure correction equations and momentum equations; governing equations of gas-liquid two-phase flow mainly include continuity equations and momentum equations, and writing governing equations of single phase and mixed phase of two-phase flow in a unified form;

[89] solving a characterization function convective transportation governing equation after the update;

[90] reconstructing the phase interface according to solution results, and re-initializing the Level Set function; and

[91] obtaining the CLSVOF model used for solving the incompressible two-phase flow phase interface by cyclic iteration.

[92] Further, the reconstructing the phase interface according to solution results 7

BL-5904 includes: LUS08134

[93] obtaining a discrete VOF function and a continuous Level Set function according to the solution results;

[94] reconstructing the interface by adopting a piecewise linear interface reconstruction method, moving the interface in a normal direction to make an area ratio of unit liquid region match a VOF function value, calculating VOF function flux between two adjacent units, and iteratively solving a vertical distance from a center of a unit to the phase interface using a secant method; and

[95] re-initializing the Level Set function into a distance function by using the reconstructed phase interface.

[96] Further, the re-initializing the Level Set function includes: determining function symbols of the Level Set function by a phase marker function, determining Level Set function values by calculating a minimum distance from a center of a grid cell to the interface, and controlling a Level Set function value of each computing unit to be a minimum value among a distance from a center of a unit to an edge, a distance from a center of a unit to a vertex of a unit, a distance from a center of a unit to an intersection point of the phase interface and a grid edge, and a distance from a center of a unit to a projection point of the phase interface.

[97] Further, the constructing an analysis model of the water hammer caused by stagnant gas includes:

[98] preprocessing acquired data in the pipeline to obtain the random distribution of the bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution;

[99] performing a normalized processing on the preprocessed data, dividing datasets, and setting original data;

[100] constructing and training an analysis model of the water hammer caused by stagnant gas based on a recurrent neural network; and

[101] acquiring real-time data in the pipeline as samples to be tested, inputting the samples to be tested into the analysis model of the water hammer caused by stagnant gas, and obtaining predictive values of the interface locations and the volume fraction information of different phases at the next moment in the pipeline.

[102] In a specific embodiment, the method further includes predicting the random distribution of the bubble swarms by a deep neural network, with the following steps:

[103] acquiring images to be identified and N consecutive images before and after the image, and forming a set of continuous images sequence according to a time sequence;

[104] inputting the continuous images sequence into a pre-trained deep neural network model; and the deep neural network model includes an encoding network module, a recurrent convolutional network module and a decoding network module;

[105] performing feature extraction on each image of the continuous images sequence in turn by the encoding network module, to obtain a feature map sequence containing semantic features of bubble swarms;

[106] receiving cyclic input of the feature map sequence from the encoding network 8

BL-5904 module by the recurrent convolutional network module, and by multi-layer circular LU508134 convolution and time series feature fusion, outputting a semantic feature map after feature fusion; and

[107] decoding the semantic feature map by the decoding network module, and outputting predicted values of the random distribution of the bubble swarms.

[108] Further, the constructing an analysis model of the water hammer caused by stagnant gas includes:

[109] preprocessing the acquired data in the pipeline to obtain the random distribution of the bubble swarms, a predicted value of the random distribution of the bubble swarms, gas characteristics, and the interface locations and volume fraction information of different phases in each spatial unit under random distribution, and the gas characteristics include bubble size, gas type, gas solubility and gas compressibility, etc.;

[110] performing a normalized processing on the preprocessed data and dividing datasets;

[111] constructing and training an analysis model of the water hammer caused by stagnant gas based on a recurrent neural network according to divided datasets; and

[112] acquiring real-time data from the pipeline as samples to be tested, inputting the samples to be tested into the analysis model of the water hammer caused by stagnant gas, and obtaining predicted values of the interface locations and volume fraction information of different phases at the next moment in the pipeline.

[113] Specifically, steps for constructing and training an analysis model of the water hammer caused by stagnant gas include:

[114] randomly dividing the preprocessed data into a training set and a testing set, then generating a model using the training set, and then testing an accuracy and error of the model using the testing set, to verify the validity of the model,

[115] constructing and training an initial prediction model, and initializing model parameters;

[116] acquiring preprocessed data within a given window and observing a true value within the given window;

[117] calculating a final predicted value within the given window of heating equipment; and

[118] updating the model parameters based on error between the final predicted value and the true value, to complete the construction and training of the final prediction model.

[119] Further, the updating model parameters includes:

[120] updating parameters of a neural network module, an attention mechanism component and a vector autoregression module adopting gradient descent and back propagation;

[121] using a mean square error MSE as an objective function, and a calculation formula is as follows: 2

[122] MSE=_ FE; (0,-0,) ;

[123] where O, represents a true value of the n-th sample, O, represents a predicted 9

BL-5904 value of the n-th sample, and N represents a number of samples; and if current error is LU508134 less than a set error threshold, it is determined that an iteration reaches a preset maximum number of iterations, i.e., obtaining a well-trained prediction model, otherwise continuing to train the prediction model.

[124] Specifically, the gas size includes:

[125] drawing a bubble size frequency density distribution diagram based on the images of the real-time bubble swarms;

[126] constructing a Gauss distribution of a Gauss model, and the Gauss distribution is:

Um

[127] d=(zz/2)"e 7

[128] Where dis a frequency density, which characterizes probability of bubble size;

[129] / is a bubble size, and j is a bubble size expectation, which characterizes an average bubble size; and z is a standard deviation of 2 times; and

[130] fitting the bubble size density distribution by using the Gaussian distribution of the Gauss model according to the obained bubble size frequency density distribution diagram, obtaining a maximum bubble size, a bubble size expectation jr and a standard deviation of 2 times z; and explaining types of bubble sizes and the probability of different bubble sizes appearing according to fitting results of the maximum bubble size, the bubble size expectation ji: and the standard deviation of 2 times z.

[131] The present embodiment fully considers the continuity of the scene, that is, the correlation on the time series, and learns the correlation information of the time prior, so as to improve the stability of the network and the accuracy of the prediction. At the same time, in the bubble swarms detection under the pipeline environment of the continuous scene, the characteristic information of the pre-stage images in the continuous images is used for post-stage images, which reduces parameters and computational complexity of the network model.

[132] Further, the optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient includes:

[133] acquiring time information based on the captured videos according to the time series;

[134] acquiring and normalizing the time information, a temperature compensation coefficient and a machine compensation coefficient to obtain an eigenvalue cluster;

[135] calculating by using a pre-trained fuzzy neural network based on the eigenvalue cluster, to obtain a time compensation coefficient; and

[136] optimizing memory units and weights of the analysis model of the water hammer caused by stagnant gas by using the time compensation coefficient.

[137] Specifically, the fuzzy neural network includes an input layer, a fuzzification layer, a fuzzy inference layer, a defuzzification layer and an output layer, by setting weights and nodes of the fuzzification layer and fuzzy inference layer, a structure of the fuzzy neural network may be changed; the input layer is filled with a normalized eigenvalue for three parameters: the time information, the temperature compensation coefficient, and the machine compensation coefficient, by the fuzzification layer, the

BL-5904 fuzzy inference layer and the defuzzification layer, a time compensation coefficient 1s LU508134 finally output, to complete the time compensation of analysis process for the water hammer caused by stagnation gas.

[138] In a specific embodiment, a training process of the fuzzy neural network includes:

[139] initializing acquired three eigenvalues: temperature compensation coefficient, time information, and machine compensation coefficient, the temperature compensation coefficient is determined according to historical temperature data, the time information is determined according to an acquisition time series, and the machine compensation coefficient is determined according to a running state of an acquisition device;

[140] calculating a corresponding output value and deviation between the output value and an expected output value when these three eigenvalues are used as inputs by utilizing a first set of eigenvalue collected by initialization;

[141] calculating an actual weight value and a result value according to the output value and the deviation by adopting a fuzzy neural network algorithm;

[142] comparing the result value of the time compensation coefficient with an expected value, if the result value is less than the expected value, it shows that the function variables and the result value calculated by the algorithm have reached an expected goal, if not, the actual weight value and function variables that do not meet the performance are calculated as required conditions for the next calculation to perform a re-iteration. until meeting a threshold range.

[143] Specifically, multiplying the time compensation coefficient by the memory units of the analysis model of the water hammer caused by stagnant gas, by adjusting a weight of the memory units, the model pays more attention to the memory of time steps, so as to improve the performance of the model in dealing with time dependence and time offset problems; or multiplying the time compensation coefficient by a connection weight or an output weight in the recurrent neural network, adjusting the importance of the model to the input of different time steps, and making the model more flexible to adapt to the input data of different time steps.

[144] In a specific embodiment, a method and system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF, as shown in

FIG 2, including:

[145] an acquisition module, which is configured to acquire random distribution images of bubble swarms in a pipeline;

[146] a preprocessing module, which is configured to preprocess, and calculate a random distribution of real-time bubble swarms based on preprocessed images;

[147] a CLSVOF model construction module, which is configured to construct a

[148] a calculation module, which is configured to calculate interface locations and volume fraction information of different phases within each spatial unit under random distribution by the CLSVOF model;

[149] an analysis model construction module, which is configured to construct an 11

BL-5904 analysis model of the water hammer caused by stagnant gas by the random distribution LU508134 of the bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution;

[150] an optimization module, which is configured to optimize the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient; and

[151] a prediction and analysis module, which is configured to perform a real-time prediction and analysis on a phenomenon of the water hammer caused by stagnant gas by an optimized analysis model of the water hammer caused by stagnant gas.

[152] The various embodiments in the present specification are all described in a progressive manner. Each of the embodiments focuses on the parts differing from the other embodiments, and the various embodiments may refer to other embodiments for the same or similar parts. A device disclosed by the embodiments is basically similar to a method disclosed by the embodiments, and therefore is described briefly. For related parts, refer to part descriptions in the method.

[153] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the present invention. Thus, the present invention is not intended to be limited to these embodiments shown herein but is to be accorded the broadest scope consistent with the principles and novel features disclosed herein. 12

Claims

BL-5904 WHAT IS CLAIMED IS: LUS08134

1. A method for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with coupled level-set and volume-of-fluid (CLSVOF), comprising: acquiring random distribution images of bubble swarms in a pipeline and preprocessing, and calculating a random distribution of real-time bubble swarms based on preprocessed images; constructing a CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set; calculating interface locations and volume fraction information of different phases within each spatial unit under random distribution by the CLSVOF model, constructing an analysis model of the water hammer caused by stagnant gas by the random distribution of bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution; optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient; and performing a real-time prediction and analysis on a phenomenon of the water hammer caused by stagnant gas by an optimized analysis model of the water hammer caused by stagnant gas.

2. The method for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF according to claim 1, wherein the acquiring random distribution images of bubble swarms in a pipeline and preprocessing comprises: capturing videos on a distribution of irregular real-time bubble swarms in the pipeline using a high-speed camera device by a visual window; and performing real-time image frame extraction and noise reduction processing on captured videos, selecting stretching area in morphology, and obtaining images of real-time bubble swarms.

3. The method for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF according to claim 1, wherein the calculating a random distribution of real-time bubble swarms based on preprocessed images comprises: capturing images on the real-time bubble swarms, to obtain the images of the real-time bubble swarms; extracting a centroid coordinate of each irregular bubble in the real-time bubble swarms, and performing a normalized processing on coordinates; and determining a first uniformity of bubble mixing in the real-time bubble swarms by adopting an average distance method based on a spatial distribution of the bubble swarms and normalized centroid coordinates.

4. The method for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF according to claim 3, wherein the method further comprises: collecting a number or density data of bubbles in each area of the real-time bubble swarms; 13

BL-5904 fitting data into a Poisson distribution model, and calculating a mean occurrence LU508134 rate of a Poisson distribution: using a Chi-square test to test a fitted Poisson distribution, to verify whether the data conforms to an assumption of the Poisson distribution; determining coincidence level between the fitted Poisson distribution and observed data by model comparison, and evaluating a second uniformity and Poissonity of the bubble distribution; and obtaining the random distribution of the real-time bubble swarms by coupling the first uniformity, the second uniformity with the Poissonity.

5. The method for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF according to claim 1, wherein the constructing a CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set comprises: initializing a Level Set function and a VOF function, respectively; updating a velocity field and a pressure field of functions by solving pressure correction equations and momentum equations after initialization; solving a characterization function convective transportation governing equation after the update; reconstructing the phase interface according to solution results, and re-initializing the Level Set function; and obtaining the CLSVOF model used for solving the incompressible two-phase flow phase interface by cyclic iteration.

6. The method for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF according to claim 5, wherein the reconstructing the phase interface according to solution results comprises: obtaining a discrete VOF function and a continuous Level Set function according to the solution results; reconstructing the interface by adopting a piecewise linear interface reconstruction method, moving the interface in a normal direction to make an area ratio of unit liquid region match a VOF function value, calculating VOF function flux between two adjacent units, and iteratively solving a vertical distance from a center of a unit to the phase interface by a secant method; and re-initializing the Level Set function into a distance function by using a reconstructed phase interface.

7. The method for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF according to claim 5, wherein the re-initializing the Level Set function comprises: determining function symbols and values of the Level Set function, and controlling a Level Set function value of each computing unit to be a minimum value among a distance from a center of a unit to an edge, a distance from a center of a unit to a vertex of a unit, a distance from a center of a unit to an intersection point of the phase interface and a grid edge, and a distance from a center of a unit to a projection point of the phase interface.

8. The method for analyzing the water hammer caused by stagnant gas by coupling 14

BL-5904 real gas characteristics with CLSVOF according to claim 1, wherein the constructing an LU508134 analysis model of water hammer caused by stagnant gas comprises: preprocessing acquired data in the pipeline to obtain the random distribution of the bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution; performing a normalized processing on preprocessed data and dividing datasets; constructing and training an analysis model of the water hammer caused by stagnant gas based on a recurrent neural network according to divided datasets; and acquiring real-time data in the pipeline as samples to be tested, inputting the samples to be tested into the analysis model of the water hammer caused by stagnant gas, and obtaining predictive values of the interface locations and the volume fraction information of different phases at the next moment in the pipeline.

9. The method for analyzing the water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF according to claim 1, wherein the optimizing the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient comprises: acquiring time information based on the captured videos according to the time series; acquiring and normalizing the time information, a temperature compensation coefficient and a machine compensation coefficient to obtain an eigenvalue cluster; calculating by using a pre-trained fuzzy neural network based on the eigenvalue cluster, to obtain a time compensation coefficient; and optimizing memory units and weights of the analysis model of the water hammer caused by stagnant gas by using the time compensation coefficient.

10. A system for analyzing water hammer caused by stagnant gas by coupling real gas characteristics with CLSVOF, comprising: an acquisition module, which is configured to acquire random distribution images of bubble swarms in a pipeline; a preprocessing module, which is configured to preprocess, and calculate a random distribution of real-time bubble swarms based on preprocessed images; a CLSVOF model construction module, which is configured to construct a CLSVOF model for solving an incompressible two-phase flow phase interface by coupling VOF with Level Set; a calculation module, which is configured to calculate interface locations and volume fraction information of different phases within each spatial unit under random distribution by the CLSVOF model; an analysis model construction module, which is configured to construct an analysis model of the water hammer caused by stagnant gas by the random distribution of the bubble swarms, gas characteristics, and the interface locations and the volume fraction information of different phases within each spatial unit under random distribution; an optimization module, which is configured to optimize the analysis model of the water hammer caused by stagnant gas by a time compensation coefficient; and

BL-5904 a prediction and analysis module, which is configured to perform a real-time LU508134 prediction and analysis on a phenomenon of the water hammer caused by stagnant gas by an optimized analysis model of the water hammer caused by stagnant gas. 16