Pipeline natural gas leakage source parameter inversion method and system

Abstract

The invention relates to the technical field of natural gas pipelines, and provides a pipeline natural gas leakage source parameter inversion method and system. The method comprises the following steps: firstly, acquiring natural gas concentration data at different monitoring sites at a moment t when the pipeline natural gas leaks, and determining a natural gas predicted concentration data set C at a moment t + delta t when the pipeline natural gas leaks based on a leakage diffusion agent model of the pipeline natural gas constructed by intrinsic orthogonal decomposition and an artificial neural network; and then, performing parameter estimation on a leakage source of the pipeline natural gas based on a leakage source inversion model of the pipeline natural gas constructed by a Bayesian network and an improved particle filtering algorithm. Therefore, specific complex concentration distribution characteristics of a natural gas leakage scene are extracted based on the leakage diffusion model, then leakage source parameter information is calculated through continuous iteration of the leakage source inversion model, the problems of pipeline natural gas leakage source positioning, parameters and the like are effectively solved, and technical support is provided for guaranteeing pipeline safety.

Description

Technical Field

[0001] The present application relates to the technical field of natural gas pipelines, and in particular to a method and system for inverting parameters of pipeline natural gas leakage sources. Background Art

[0002] As an indispensable part of modern energy supply, natural gas is widely used in industry, power generation and residents' lives as a low-carbon fuel. As pipeline materials deteriorate over time due to environmental factors, cracks or holes may form, causing gas leakage. Especially in narrow and confined spaces such as pipelines, the diffusion of natural gas is hindered. Once a leakage accident occurs, the consequences will be more serious.

[0003] Initially, the discovery of natural gas leaks mainly relied on manual inspections and visual observations. Pipelines and the surrounding environment were regularly inspected to see if there were any signs of odor or withered vegetation. With the advancement of technology, methods such as gas detectors and acoustic wave detection were gradually introduced. Gas detectors can monitor changes in gas concentration around pipelines in real time and detect leaks in a timely manner. Acoustic wave detection technology identifies the location of leaks by analyzing the propagation of sound waves in pipelines. Although these methods have significantly improved the accuracy and efficiency of detection, there are still problems such as missed detection and inaccurate positioning. Summary of the invention

[0004] The purpose of the present application is to provide a pipeline natural gas leakage source parameter inversion method and system to solve or alleviate the problems existing in the above-mentioned prior art.

[0005] In order to achieve the above objectives, this application provides the following technical solutions:

[0006] The present application provides a pipeline natural gas leakage source parameter inversion method, comprising: step S101, obtaining natural gas concentration data at different monitoring points at time t when the pipeline natural gas leaks;

[0007] Step S102, based on the leakage and diffusion proxy model of the pipeline natural gas constructed by intrinsic orthogonal decomposition and artificial neural network, obtain the natural gas concentration prediction data set C at time t+Δt when the pipeline natural gas leaks according to the natural gas concentration data;

[0008] Step S103, based on the inversion model of the pipeline natural gas leakage source constructed by the Bayesian network and the improved particle filter algorithm, the parameters of the pipeline natural gas leakage source are estimated according to the natural gas predicted concentration dataset C at the time t+Δt when the pipeline natural gas leaks.

[0009] Preferably, in step S102,

[0010] The gas concentration data is subjected to dimensionality reduction processing based on intrinsic orthogonal decomposition, and the dimensionality reduction data set obtained by the dimensionality reduction processing when the pipeline natural gas leaks is input into the artificial neural network to obtain the natural gas concentration prediction data set C at time t+Δt when the pipeline natural gas leaks; wherein,

[0011]

[0012] In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the predicted natural gas concentration at time t+Δt; i = {1,…,m}, j = {1,…,n}, m and n are both positive integers.

[0013] Preferably, in step S103, based on the leakage source inversion model, according to the natural gas prediction concentration data set C at time t+Δt when the pipeline natural gas leaks and the natural gas concentration data at different monitoring points at time t, the leakage coordinates in the leakage field when the pipeline natural gas leaks are cyclically resampled;

[0014] Based on Gaussian perturbation and Markov transfer matrix, the leakage source parameters corresponding to the leakage coordinates after cyclic resampling are calculated. To update; among them, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the predicted natural gas concentration at time t+Δt.

[0015] Preferably, the method of performing cyclic resampling of the leakage coordinates in the leakage field when the pipeline natural gas leaks based on the leakage source inversion model and the natural gas predicted concentration dataset C at time t+Δt when the pipeline natural gas leaks and the natural gas concentration data at different monitoring points at time t comprises: based on the Bayesian network in the leakage source inversion model, according to the leakage coordinates (s i ,x j ) at the time t+Δt, the predicted concentration of natural gas and the natural gas concentration data at time t, and calculate the leakage coordinates (s i ,x j )’s position weight Wherein, i={1,…,m}, j={1,…,n}, m and n are both positive integers;

[0016] According to the leakage coordinates (s i ,x j )’s position weight, according to the formula:

[0017]

[0018] Calculate the effective number Neff of leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks; where w′ ij is the leakage coordinates (s i ,x j )’s position weight The weight value after normalization;

[0019] In response to the valid number Neff of leakage coordinates in the leakage field at time t+Δt during the natural gas pipeline leakage being less than a preset threshold, the leakage coordinates in the leakage field during the natural gas pipeline leakage are cyclically resampled until the valid number Neff of leakage coordinates in the leakage field during the natural gas pipeline leakage is greater than or equal to the preset threshold.

[0020] Preferably, in response to the valid number Neff of leakage coordinates in the leakage field at time t+Δt in the pipeline natural gas leakage being less than a preset threshold, cyclically resampling the leakage coordinates in the leakage field in the pipeline natural gas leakage until the valid number Neff of leakage coordinates in the leakage field in the pipeline natural gas leakage is greater than or equal to the preset threshold, comprising:

[0021] In response to the effective number Neff of leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks is less than a preset threshold, the leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks are screened, and the leakage coordinates at time t+Δt are retained (s i ,x j ) whose position weight is greater than or equal to the preset weight threshold (s i ,x j ), and retain the leakage coordinates (s i ,x j )Add noise information;

[0022] According to the position weight of the leakage coordinates after the noise information is added at time t+Δt in the leakage field when the pipeline natural gas leaks, the effective number Neff of leakage coordinates is cyclically judged on the leakage coordinates after the noise information is added at time t+Δt in the leakage field when the pipeline natural gas leaks, until the effective number Neff of leakage coordinates in the leakage field when the pipeline natural gas leaks is greater than or equal to a preset threshold.

[0023] Preferably, it also includes: calculating the leakage coordinates (s i+Δi ,x j+Δj)

[0024] The leakage coordinates (s) with added noise information i+Δi ,x j+Δj ) The data is compared with the natural gas concentration data at time t to determine the leakage coordinates (s i+Δi ,x j+Δj )’s position weight.

[0025] Preferably, according to the formula:

[0026]

[0027] When the pipeline natural gas leaks, the leakage coordinates (s i ,x j )’s position weight Normalize it and get the weight value w ij ;

[0028] In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) at time t, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) predicted natural gas concentration at time t+Δt;

[0029] σ is the leakage coordinate in the leakage field when the pipeline natural gas leaks (s i ,x j ) The measurement error of the natural gas concentration 0 at time t;

[0030] i={1,…,m}, j={1,…,n}, m and n are both positive integers.

[0031] Preferably, the leakage source parameters corresponding to the leakage coordinates after cyclic resampling are obtained based on Gaussian perturbation and Markov transfer matrix. Updates include:

[0032] According to the formula:

[0033]

[0034] The leakage source parameters corresponding to the leakage coordinates after cyclic resampling To update; among them,

[0035]

[0036] In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) at t+

[0037] Predicted natural gas concentration at time Δt;

[0038] is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) at time t, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) The predicted natural gas concentration at time t; The leakage source parameters Leakage source parameters after adding Gaussian perturbation;

[0039] is the leakage coordinates (s) at time t in the leakage field when the pipeline natural gas leaks i ,x j ) to the leakage coordinates at time t+Δt (s i+Δi ,x j+Δj )’s Markov transfer matrix; S is the leakage source parameter The covariance matrix constructed after Cholesky decomposition;

[0040] is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i+Δi ,x j+Δj ) is the likelihood estimate at time t+Δt; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the likelihood estimate at time t;

[0041] μ∈[0,1] is a random function; i={1,…,m}, j={1,…,n}, m and n are both positive integers.

[0042] Preferably, the present application also provides a pipeline natural gas leakage source parameter inversion system, comprising:

[0043] A data acquisition unit configured to obtain natural gas concentration data at different monitoring points at time t when natural gas leaks from a pipeline;

[0044] A simulation prediction unit is configured as a leakage diffusion proxy model of the pipeline natural gas constructed based on intrinsic orthogonal decomposition and artificial neural network, and obtains a natural gas concentration prediction data set C at time t+Δt when the pipeline natural gas leaks according to the natural gas concentration data;

[0045] The inversion estimation unit is configured as an inversion model of the pipeline natural gas leakage source constructed based on a Bayesian network and an improved particle filtering algorithm, and performs parameter estimation on the pipeline natural gas leakage source according to a natural gas predicted concentration data set C at time t+Δt when the pipeline natural gas leaks.

[0046] Beneficial effects:

[0047] The pipeline natural gas leakage source inversion method provided in the embodiment of the present application obtains the natural gas concentration data at different monitoring points at time t when the pipeline natural gas leaks, and determines the natural gas predicted concentration data set C at time t+Δt when the pipeline natural gas leaks based on the leakage diffusion proxy model of the pipeline natural gas constructed by the intrinsic orthogonal decomposition and artificial neural network; then, the leakage source inversion model of the pipeline natural gas constructed based on the Bayesian network and the improved particle filter algorithm is used to estimate the parameters of the leakage source of the pipeline natural gas. Thus, the complex concentration distribution characteristics unique to the natural gas leakage scene are extracted based on the leakage diffusion model, and then the leakage source parameter information is continuously iterated and calculated through the leakage source inversion model, which effectively solves the problems of pipeline natural gas leakage source positioning and parameters, and provides technical support for ensuring pipeline safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The drawings in the specification, which constitute a part of the present application, are used to provide further understanding of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute improper limitations on the present application.

[0049] in:

[0050] Figure 1 A schematic flow chart of a pipeline natural gas leakage source inversion method provided according to some embodiments of the present application;

[0051] Figure 2 A schematic diagram showing a comparison between simulation results of a leakage diffusion proxy model and CFD results in a no-obstacle scenario provided according to some embodiments of the present application;

[0052] Figure 3 A schematic diagram comparing simulation results of a leakage diffusion proxy model and CFD results in an obstacle scenario provided according to some embodiments of the present application;

[0053] Figure 4 A schematic diagram of initial distribution of leakage coordinates of a leakage source inversion model in different scenarios according to some embodiments of the present application;

[0054] Figure 5 A schematic diagram of the convergence of leakage coordinates of a leakage source inversion model in different scenarios according to some embodiments of the present application;

[0055] Figure 6 A posterior probability distribution histogram of the horizontal coordinate of the leakage source in a scene without obstacles provided according to some embodiments of the present application;

[0056] Figure 7 A posterior probability distribution histogram of the horizontal coordinate of the leakage source in a scene with obstacles provided according to some embodiments of the present application;

[0057] Figure 8 A posterior probability distribution histogram of the vertical coordinate of the leakage source in a scene without obstacles provided according to some embodiments of the present application;

[0058] Fig. 9 A posterior probability distribution histogram of the vertical coordinate of the leakage source in a scene with obstacles provided according to some embodiments of the present application;

[0059] Fig.10 A posterior probability distribution histogram of leakage intensity (leaked natural gas concentration) in a no-obstacle scenario provided according to some embodiments of the present application;

[0060] Fig.11 A posterior probability distribution histogram of leakage intensity (leaked natural gas concentration) in an obstacle scenario provided according to some embodiments of the present application;

[0061] Fig.12 A schematic diagram of the structure of a pipeline natural gas leakage source inversion system provided according to some embodiments of the present application. DETAILED DESCRIPTION

[0062] The present application will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments. Each example is provided by way of explanation of the present application and does not limit the present application. In fact, it will be clear to those skilled in the art that modifications and variations may be made in the present application without departing from the scope or spirit of the present application. For example, a feature shown or described as a part of an embodiment may be used in another embodiment to produce yet another embodiment. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the embodiments of the present invention should belong to the scope of protection of the embodiments of the present invention.

[0063] In the leakage detection of natural gas pipelines, existing gas leakage detection methods mostly use Bayesian methods or swarm intelligence algorithms. The leakage detection using the Bayesian method requires a large number of sampling samples, a huge amount of calculation, and is difficult to adapt to complex situations; the leakage detection using the swarm intelligence algorithm cannot take into account model and instrument errors, and the positioning accuracy is affected by the search strategy.

[0064] In order to promote the automation and intelligence of pipeline leakage identification and overcome the shortcomings of existing technologies based on traditional algorithms, this application proposes a pipeline natural gas leakage source parameter inversion method. Through the fast and forward diffusion simulation of the proxy model, the high-precision results of CFD calculation are approximately obtained to meet the needs of emergency disposal and provide support for promoting the automation and intelligence of emergency handling of leakage accidents.

[0065] like Figures 1 to 11 As shown, the pipeline natural gas leakage source parameter inversion method includes:

[0066] Step S101, obtaining natural gas concentration data at different monitoring points at time t when pipeline natural gas leaks.

[0067] In this application, a grid layout method is used to arrange sensors in the area of ​​pipeline natural gas. Then, when a natural gas leak occurs in the pipeline, sensors at different monitoring points can be used to realize real-time and uniform monitoring of the natural gas concentration in the pipeline natural gas leakage field, and obtain the natural gas concentration data when the pipeline natural gas leaks. Alternatively, a mid-infrared sensor is mounted on a drone to scan the leakage area during a natural gas leak, and obtain the natural gas concentration data of the leakage area. Then, the obtained natural gas concentration data of the leakage area is preprocessed, including checking and processing missing values, normalizing the acquired data, and other operations.

[0068] Step S102: Based on the pipeline natural gas leakage and diffusion proxy model constructed by intrinsic orthogonal decomposition and artificial neural network, a natural gas concentration prediction dataset C at time t+Δt when the pipeline natural gas leaks is obtained according to the natural gas concentration data.

[0069] By preprocessing the natural gas concentration data in the leakage area obtained by the sensor, a natural gas concentration dataset C in the leakage field when the pipeline natural gas leaks is generated. 0 Then, the natural gas concentration data set C is generated based on the intrinsic orthogonal decomposition of the gas concentration data 0 Dimensionality reduction processing is performed to obtain the dimension reduction data set C′ of pipeline natural gas leakage. Specifically, the natural gas concentration data set C is transformed into 0 Transformed into a left singular matrix, a diagonal matrix, and a right singular matrix; the diagonal matrix is ​​composed of the natural gas concentration data set C 0 Then, the natural gas concentration dataset C 0The singular values ​​in the decomposed diagonal matrix are sorted in descending order, and the first r singular values ​​greater than or equal to the preset threshold are extracted to form a new diagonal matrix, and the elements corresponding to the first r singular values ​​in the left singular matrix and the right singular matrix are extracted to form a new left singular matrix and a new right singular matrix; finally, the new left singular matrix, the new diagonal matrix, and the new right singular matrix are combined to obtain the reduced dimension data set C′. Thus, by analyzing the natural gas concentration data set C in the leakage field when the pipeline natural gas leaks, 0 Perform singular value decomposition to transform the complex high-dimensional natural gas concentration dataset C 0 Dimensionality reduction decomposition is used to simultaneously extract the spatial distribution of the concentration field, gas diffusion path and shape, and other characteristics during pipeline natural gas leakage, as well as the relationship between the natural gas leakage intensity and the concentration field and the influence of the leakage source location on the concentration distribution.

[0070] In this application, low-dimensional features are extracted from complex high-dimensional data by singular value decomposition, and then the data after dimensionality reduction by singular value decomposition is trained with an artificial neural network, and the trained artificial neural network is used for prediction. Here, the artificial neural network includes three mutually learning parts: input layer, hidden layer and output layer, to predict the leakage and diffusion of pipeline natural gas. First, the source term conditions (low-dimensional features extracted from complex high-dimensional data) are input into the hidden layer, and then, in the hidden layer, the nonlinear features are converted into linear features by a nonlinear hyperbolic tangent function; finally, the output features of the hidden layer are linearly summed to obtain the natural gas concentration prediction result (i.e., the predicted natural gas concentration).

[0071] That is, the natural gas concentration dataset C 0 The reduced dimension data set obtained by dimensionality reduction decomposition is input into the artificial neural network, the artificial neural network is trained, and the trained artificial neural network is used for prediction to obtain the natural gas prediction concentration data set C at time t+Δt when the pipeline natural gas leaks. Among them,

[0072]

[0073] In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the predicted natural gas concentration at time t+Δt; i = {1,…,m}, j = {1,…,n}, m and n are both positive integers.

[0074] Therefore, the set of natural gas concentration data is reduced in dimension by intrinsic orthogonal decomposition to obtain a modal energy intensity matrix (reduced dimension data set), which retains the main characteristics of the natural gas concentration field and has a strong mapping relationship with the source term (leakage source) of natural gas leakage. The modal energy intensity matrix (reduced dimension data set) is used for artificial neural network training, and the natural gas concentration distribution is restored using the orthogonal basis. In this process, the intrinsic orthogonal decomposition reduces the output features of the artificial neural network from 1861 to 77, greatly reducing the output dimension of the neural network, effectively reducing the model size and computational consumption, and improving computational efficiency.

[0075] Step S103: Based on the Bayesian network and the improved particle filter algorithm, the leakage source inversion model of the pipeline natural gas is constructed, and according to the natural gas predicted concentration dataset C at the time t+Δt when the pipeline natural gas leaks, the leakage source of the pipeline natural gas is estimated.

[0076] In this application, based on the leakage source inversion model, according to the natural gas prediction concentration data set C at time t+Δt when the pipeline natural gas leaks and the natural gas concentration data at different monitoring points at time t, the leakage coordinates in the leakage field when the pipeline natural gas leaks are cyclically resampled. In this process, first, based on the Bayesian network in the leakage source inversion model, according to the leakage coordinates (s i ,x j ) at the time t+Δt, the predicted concentration of natural gas The natural gas concentration data at time t is used to calculate the leakage coordinates (s i ,x j )’s position weight

[0077] Specifically, according to the formula:

[0078]

[0079] The leakage coordinates (s i ,x j )’s position weight Normalize and get the weight value w′ ij In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) at time t, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j) is the predicted concentration of natural gas at time t+Δt; σ is the leakage coordinate in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the measurement error of the natural gas concentration at time t; i = {1,…,m}, j = {1,…,n}, m and n are both positive integers.

[0080] That is to say, the prior information of the natural gas leakage source is converted into a posteriori information by combining the natural gas concentration data obtained by the sensor through the Bayesian network; that is, according to the formula:

[0081]

[0082] Determine the likelihood probability P(N|M) of the natural gas concentration collected by the sensor when the pipeline natural gas leaks; where M is the leakage source parameter (including location coordinates and leakage amount) when the pipeline natural gas leaks, and N is the natural gas concentration data at the corresponding location coordinates collected when the pipeline natural gas leaks.

[0083] Then, according to the leakage coordinates (s i ,x j ) to determine the effective leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks. That is, according to the formula:

[0084]

[0085] Calculate the effective number Neff of leakage coordinates at the time t+Δt in the leakage field when the pipeline natural gas leaks. Therefore, by defining the number of effective leakage coordinates at the time t+Δt in the leakage field when the pipeline natural gas leaks, it is judged whether particle degradation occurs in the particle filtering process. When the effective number of leakage coordinates at the time t+Δt in the leakage field when the pipeline natural gas leaks is less than the preset threshold, it will lead to a decrease in the diversity of the predicted leakage coordinates, thereby reducing the accuracy of the leakage source prediction. In this regard, when the effective number Neff of leakage coordinates at the time t+Δt in the leakage field when the pipeline natural gas leaks is less than the preset threshold, the leakage coordinates in the leakage field when the pipeline natural gas leaks are cyclically resampled until the effective number Neff of leakage coordinates in the leakage field when the pipeline natural gas leaks is greater than or equal to the preset threshold.

[0086] Specifically, when the effective number Neff of leakage coordinates at time t+Δt in the leakage field during the natural gas pipeline leakage is less than a preset threshold, the leakage coordinates at time t+Δt in the leakage field during the natural gas pipeline leakage are screened, and the leakage coordinates at time t+Δt (s i ,x j ) whose position weight is greater than or equal to the preset weight threshold (s i ,x j), and the leakage coordinates (s i ,x j )Add noise information.

[0087] Based on the leakage diffusion proxy model, the leakage coordinates (s i+Δi ,x j+Δj ) Among them, Δi and Δj are the leakage coordinates after adding noise information (s i ,x j ) coordinate offset. Then, the leakage coordinate (s i+Δi ,x j+Δj ) Compare with the natural gas concentration data at time t to determine the leakage coordinates (s i+Δi ,x j+Δj )’s position weight.

[0088] After adding noise information at time t+Δt, the leakage coordinates (s i+Δi ,x j+Δj ), calculate the number of valid leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks after adding the noise information, and determine whether the valid number Neff of leakage coordinates after adding the noise information is greater than or equal to the preset threshold. If the valid number Neff of leakage coordinates after adding the noise information is still less than the preset threshold, continue to add the noise information to the valid leakage coordinates for resampling and determination of the number of valid coordinates until the valid number Neff of leakage coordinates is greater than or equal to the preset threshold.

[0089] In this application, multiple particles are generated by an improved particle filter algorithm, and weights are assigned. The output of the leakage diffusion proxy model is combined to update the particle state and evaluate the height, and finally the particle with the highest weight is selected as the final leakage source estimated position and concentration. After resampling, the particle weight will be repeatedly copied many times, resulting in the concentration of particles in a few areas. In order to effectively avoid particle depletion in the process of determining the location and concentration of the leakage source, a random disturbance is added to each resampled particle (sampled from the Gaussian kernel and added to the corresponding leakage coordinates) to diffuse the particle to the nearby space. Thus, a random disturbance sample is provided for each particle through the Gaussian kernel, which effectively avoids the concentration of particles in a few areas, offsets the particle position, and redistributes it in space, enhancing the exploration ability of particles.

[0090] Specifically, after determining that the effective number Neff of leakage coordinates is greater than or equal to a preset threshold, for particles corresponding to the effective leakage coordinates, based on Gaussian perturbation and Markov transfer matrix, the leakage source parameters corresponding to the leakage coordinates after cyclic resampling are calculated. The update is performed to achieve parameter estimation of the leakage source of pipeline natural gas.

[0091] Specifically, according to the formula:

[0092]

[0093] The leakage source parameters corresponding to the leakage coordinates after cyclic resampling To update.

[0094]

[0095] In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) predicted natural gas concentration at time t+Δt; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) at time t, Leak source parameter Leakage source parameters after adding Gaussian perturbation;

[0096] is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) The predicted natural gas concentration at time t; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i+Δi ,x j+Δj ) predicted natural gas concentration at time t+Δt;

[0097] is the leakage coordinates at time t in the leakage field when the pipeline natural gas leaks (s i ,x j ) to the leakage coordinates at time t+Δt (s i+Δi ,x j+Δj )’s Markov transfer matrix; S is the leakage source parameter The covariance matrix constructed after Cholesky decomposition;

[0098] is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i+Δi ,x j+Δj ) is the likelihood estimate at time t+Δt; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the likelihood estimate at time t.

[0099] Here, is the updated leakage source parameter Leak source parameters The offset distance between them, i.e. the leakage source parameter Leakage coordinates and leakage source parameters The distance between the leakage coordinates. And according to the formula:

[0100]

[0101] Determine the covariance matrix S. Where A is the intermediate variable, h is the scaling factor during Cholesky decomposition, blkdiag is the block diagonal matrix constructor, u is the number of leakage source parameters, mn represents the total number of leakage coordinates, ∑pos is the leakage coordinate (s i ,x j ), is the leakage coordinate (s i ,x j ) is the state covariance matrix of the predicted natural gas concentration at , is the weighted average coordinate position of mn leakage coordinates, It is the weighted average concentration of the predicted natural gas concentration at mn leakage coordinates.

[0102] Therefore, the natural gas predicted concentration data set is obtained through the leakage diffusion proxy model. Compared with the classic Gaussian plume model, there is no need to consider the influence of obstacles in the diffusion area, and the calculation speed is fast, the feature extraction is accurate, and the calculation efficiency is high. The source term inversion is performed through the leakage source inversion model to effectively solve the problems of pipeline natural gas leakage source positioning and parameters, and provide technical support for ensuring pipeline safety. It has fast processing speed, simple structure, and small error, which can effectively enhance the physical significance of pipeline leakage identification and overcome the problems of long calculation time and low accuracy of traditional models.

[0103] like Fig.12 As shown, the embodiment of the present application also provides a pipeline natural gas leakage source parameter inversion system, including:

[0104] The data acquisition unit 1201 is configured to obtain natural gas concentration data at different monitoring points at time t when the pipeline natural gas leaks;

[0105] The simulation prediction unit 1202 is configured as a pipeline natural gas leakage diffusion proxy model based on intrinsic orthogonal decomposition and artificial neural network, and obtains a natural gas concentration prediction data set C at time t+Δt when the pipeline natural gas leaks according to the natural gas concentration data;

[0106] The inversion estimation unit 1203 is configured as a pipeline natural gas leakage source inversion model constructed based on a Bayesian network and an improved particle filter algorithm, and performs parameter estimation on the pipeline natural gas leakage source according to a natural gas predicted concentration dataset C at time t+Δt when the pipeline natural gas leaks.

[0107] The pipeline natural gas leakage source parameter inversion system provided in the present application can realize the steps and processes of the pipeline natural gas leakage source parameter inversion method of any of the above embodiments and achieve the same technical effects, which will not be repeated here.

[0108] In the description of the present invention, it is to be understood that the terms “center”, “longitudinal”, “lateral”, “length”, “width”, “thickness”, “up”, “down”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside”, “outside”, “clockwise”, “counterclockwise”, “axial”, “radial”, “circumferential”, etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0109] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0110] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral one; it can be a mechanical connection, an electrical connection, or communication with each other; it can be a direct connection, or an indirect connection through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements, unless otherwise clearly defined. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0111] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

[0112] In the present invention, the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" mean that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0113] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A pipeline natural gas leakage source parameter inversion method, characterized in that: include: Step S101, obtaining natural gas concentration data at different monitoring points at time t when pipeline natural gas leaks; Step S102, based on the leakage and diffusion proxy model of the pipeline natural gas constructed by intrinsic orthogonal decomposition and artificial neural network, obtain the natural gas concentration prediction data set C at time t+Δt when the pipeline natural gas leaks according to the natural gas concentration data; Step S103, based on the inversion model of the pipeline natural gas leakage source constructed by the Bayesian network and the improved particle filter algorithm, the parameters of the pipeline natural gas leakage source are estimated according to the natural gas predicted concentration dataset C at the time t+Δt when the pipeline natural gas leaks.

2. The pipeline natural gas leakage source parameter inversion method according to claim 1 is characterized in that: In step S102, The gas concentration data is subjected to dimensionality reduction processing based on intrinsic orthogonal decomposition, and the dimensionality reduction data set obtained by the dimensionality reduction processing when the pipeline natural gas leaks is input into the artificial neural network to obtain the natural gas concentration prediction data set C at time t+Δt when the pipeline natural gas leaks; wherein, In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the predicted natural gas concentration at time t+Δt; i = {1,…,m}, j = {1,…,n}, m and n are both positive integers.

3. The pipeline natural gas leakage source parameter inversion method according to claim 2 is characterized in that: In step S103, Based on the leakage source inversion model, according to the natural gas prediction concentration data set C at time t+Δt when the pipeline natural gas leaks and the natural gas concentration data at different monitoring points at time t, the leakage coordinates in the leakage field when the pipeline natural gas leaks are cyclically resampled; Based on Gaussian perturbation and Markov transfer matrix, the leakage source parameters corresponding to the leakage coordinates after cyclic resampling are calculated. To update; among them, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the predicted natural gas concentration at time t+Δt.

4. The pipeline natural gas leakage source parameter inversion method according to claim 3 is characterized in that: The method based on the leakage source inversion model, according to the natural gas prediction concentration data set C at time t+Δt when the pipeline natural gas leaks and the natural gas concentration data at different monitoring points at time t, cyclically resampling the leakage coordinates in the leakage field when the pipeline natural gas leaks, includes: Based on the Bayesian network in the leakage source inversion model, according to the leakage coordinates (s i ,x j ) at the time t+Δt, the predicted concentration of natural gas and the natural gas concentration data at time t, and calculate the leakage coordinates (s i ,x j )’s position weight Wherein, i={1,…,m}, j={1,…,n}, m and n are both positive integers; According to the leakage coordinates (s i ,x j )’s position weight, according to the formula: Calculate the effective number Neff of leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks; where w′ ij is the leakage coordinates (s i ,x j )’s position weight The weight value after normalization; In response to the valid number Neff of leakage coordinates in the leakage field at time t+Δt during the natural gas pipeline leakage being less than a preset threshold, the leakage coordinates in the leakage field during the natural gas pipeline leakage are cyclically resampled until the valid number Neff of leakage coordinates in the leakage field during the natural gas pipeline leakage is greater than or equal to the preset threshold.

5. The pipeline natural gas leakage source parameter inversion method according to claim 4 is characterized in that: In response to the valid number Neff of leakage coordinates in the leakage field at time t+Δt in the pipeline natural gas leakage being less than a preset threshold, cyclically resampling the leakage coordinates in the leakage field in the pipeline natural gas leakage until the valid number Neff of leakage coordinates in the leakage field in the pipeline natural gas leakage is greater than or equal to the preset threshold, including: In response to the effective number Neff of leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks is less than a preset threshold, the leakage coordinates at time t+Δt in the leakage field when the pipeline natural gas leaks are screened, and the leakage coordinates at time t+Δt are retained (s i ,x j ) whose position weight is greater than or equal to the preset weight threshold (s i ,x j ), and retain the leakage coordinates (s i ,x j )Add noise information; According to the position weight of the leakage coordinates after the noise information is added at time t+Δt in the leakage field when the pipeline natural gas leaks, the effective number Neff of leakage coordinates is cyclically judged on the leakage coordinates after the noise information is added at time t+Δt in the leakage field when the pipeline natural gas leaks, until the effective number Neff of leakage coordinates in the leakage field when the pipeline natural gas leaks is greater than or equal to a preset threshold.

6. The pipeline natural gas leakage source parameter inversion method according to claim 5 is characterized in that: Also includes: Based on the leakage diffusion proxy model, the leakage coordinates (s i+Δi ,x j+Δj ) The leakage coordinates (s) with added noise information i+Δi ,x j+Δj ) The data is compared with the natural gas concentration data at time t to determine the leakage coordinates (s i+Δi ,x j+Δj )’s position weight.

7. The pipeline natural gas leakage source parameter inversion method according to claim 4 is characterized in that: According to the formula: When the pipeline natural gas leaks, the leakage coordinates (s i ,x j )’s position weight Normalize and get the weight value In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) at time t, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) predicted natural gas concentration at time t+Δt; σ is the leakage coordinate in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the measurement error of the natural gas concentration at time t; i={1,…,m}, j={1,…,n}, m and n are both positive integers.

8. The pipeline natural gas leakage source parameter inversion method according to claim 3 is characterized in that: The leakage source parameters corresponding to the leakage coordinates after cyclic resampling are calculated based on Gaussian perturbation and Markov transfer matrix. Updates include: According to the formula: The leakage source parameters corresponding to the leakage coordinates after cyclic resampling To update; among them, In the formula, is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) predicted natural gas concentration at time t+Δt; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) at time t, The leakage source parameters Leakage source parameters after adding Gaussian perturbation; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) The predicted natural gas concentration at time t; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i+Δi ,x j+Δj ) predicted natural gas concentration at time t+Δt; is the leakage coordinates (s) at time t in the leakage field when the pipeline natural gas leaks i ,x j ) to the leakage coordinates at time t+Δt (s i+Δi ,x j+Δj )’s Markov transition matrix; S is the leakage source parameter The covariance matrix constructed after Cholesky decomposition; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i+Δi ,x j+Δj ) is the likelihood estimate at time t+Δt; is the leakage coordinates in the leakage field when the pipeline natural gas leaks (s i ,x j ) is the likelihood estimate at time t; μ∈[0,1], is a random function; i={1,…,m}, j={1,…,n}, m and n are both positive integers.

9. A pipeline natural gas leakage source parameter inversion system, characterized in that: include: A data acquisition unit configured to obtain natural gas concentration data at different monitoring points at time t when natural gas leaks from a pipeline; A simulation prediction unit is configured as a leakage diffusion proxy model of the pipeline natural gas constructed based on intrinsic orthogonal decomposition and artificial neural network, and obtains a natural gas concentration prediction data set C at time t+Δt when the pipeline natural gas leaks according to the natural gas concentration data; The inversion estimation unit is configured as an inversion model of the pipeline natural gas leakage source constructed based on a Bayesian network and an improved particle filtering algorithm, and performs parameter estimation on the pipeline natural gas leakage source according to a natural gas predicted concentration data set C at time t+Δt when the pipeline natural gas leaks.

Images