Rapid and accurate tracing and positioning method and system for point source high-risk gas leakage

Abstract

The invention discloses a point source high-risk gas leakage rapid and accurate traceability positioning method and system, and relates to the technical field of gas leakage source localization, and the method comprises the steps: determining a mesoscale turbulence diffusion model according to on-site geographical environment information and high-risk gas characteristics, the position of the virtual monitoring point is optimized by verifying the validity of the virtual monitoring point in an actual scene; and then batch numerical simulation is carried out to obtain an associated data set, a neural network model is trained according to the associated data set, a leakage source positioning model is obtained, the predicted actual leakage source position is output according to the gas concentration collected by each actual monitoring point, and accurate traceability of the leakage source is realized. According to the scheme, the diffusion phenomenon is accurately simulated through the optimized diffusion model, so that the obtained data is more practical, and the traceability precision is indirectly improved; and based on a machine learning method of an artificial neural network, the correlation between the concentration distribution and the leakage source position is established, and accurate and rapid traceability positioning of the leakage source is realized.

Description

Technical Field

[0001] The present application relates to the technical field of gas leakage source positioning, and in particular to a method and system for rapid and accurate source tracing and positioning of point source high-risk gas leakage. Background Art

[0002] Gas leak source location technology is a technology that uses instruments or other means to search for the location of toxic gas leaks. This technology can avoid large-scale pollution and damage caused by long-term high-risk gas leaks, and reduce economic losses to a minimum. In terms of high-risk gas leak tracing and location, foreign research started early and has advanced technology. For example, ABB's MobileGuard product in the United States has a mobile detection of natural gas concentration + natural gas leak tracing and location time of 5 minutes. There are two main types of foreign countries: algorithm tracing and robot tracing. Domestic tracing has just started in recent years and has little research. Algorithm tracing is mostly based on the concentration and changes of the monitoring point. The concentration of the monitoring point is predicted by a two-dimensional diffusion model, and it is combined with iterative algorithms such as Bayesian reasoning, random Monte Carlo, and Markov chain Monte Carlo sampling to continuously correct the deviation of the predicted concentration until the deviation is within an acceptable range; this method has a relatively short tracing time, is affected by the diffusion model and scene environmental factors, and has low prediction accuracy. Robotic tracing, also known as active tracing, optimizes the robot's search path through optimization methods. At the same time, based on the detection data of the current search point, the diffusion model is used to infer the direction of the next high concentration point and promptly correct the search path (i.e., metaheuristic / artificial olfactory method) until the highest concentration point (i.e., the leak source) is found or the acceptable deviation range is reached; this method has high traceability accuracy, but the traceability time is relatively long, and it is suitable for continuous leak point search. It can be found that these two common traceability methods are inseparable from the diffusion model, and the time and accuracy consumed by the traceability basically depend on the accuracy of the diffusion model.

[0003] In summary, due to the insufficient accuracy of the traditional diffusion model, the existing high-risk gas leakage source positioning technology has the problem of insufficient leakage source positioning accuracy. Summary of the invention

[0004] The purpose of this application is to provide a method and system for quickly and accurately tracing and locating the source of point source high-risk gas leakage, which can accurately and quickly trace the source of the leakage.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In the first aspect, the present application provides a method for quickly and accurately tracing and locating a point source high-risk gas leak, comprising the following steps:

[0007] According to the spatial distribution information and data of the site, the site is modeled in proportion to obtain a real-scene model of the site.

[0008] The mesoscale turbulent diffusion model is determined according to the on-site geographical environment information and the characteristics of high-risk gases; the mesoscale turbulent diffusion model is a mesoscale optimization model based on the traditional turbulent diffusion model.

[0009] Arrange a set of virtual monitoring points in the actual scene, and optimize the location of the virtual monitoring points by verifying their effectiveness.

[0010] Based on the on-site real-scene model and the mesoscale turbulence diffusion model, batch numerical simulations are carried out to obtain the associated data sets of the virtual leakage source location and the gas concentration collected at each virtual monitoring point.

[0011] Based on the associated data set, a neural network model is trained to obtain a leakage source location model.

[0012] The gas concentration collected at each actual monitoring point is collected and input into the leakage source positioning model, and the predicted actual leakage source location is output to achieve accurate tracing of the leakage source.

[0013] Optionally, in the mesoscale turbulent diffusion model, the method of the energy minimum multi-scale principle is used to structurally quantify the diffusion field of each sub-phase of the flow field, and correct the turbulent diffusion coefficient of the mesoscale turbulent diffusion model; for the laminar sub-phase, the structural quantification indicators of the laminar sub-phase include: the laminar sub-phase flow velocity U lam , fluid density ρ lam , the macroscopic volume fraction 1-u of the laminar subphase and the laminar diffusion coefficient D lam ; For the vortex subphase, the structural quantitative indicators of the vortex subphase include: vortex subphase flow velocity U eddy , fluid density ρ eddy , the macroscopic volume fraction u of the vortex phase and the eddy diffusion coefficient D eddy ; The laminar subphase and the vortex subphase interact at the mesoscale level. The structural quantitative indicators at the mesoscale level include: vortex diameter d eddy and the fluid-eddy interface tension σ.

[0014] Optionally, the turbulent diffusion coefficient of the mesoscale turbulent diffusion model is corrected according to the following formula:

[0015] D turb =f(U lam ,ρ lam ,1-u,D lam ; U eddy ,ρ eddy ,u,D eddy ;σ,d eddy ).

[0016] Among them, D turb represents the turbulent diffusion coefficient of the mesoscale turbulent diffusion model.

[0017] Optionally, a group of virtual monitoring points are arranged in the actual scene, and the positions of the virtual monitoring points are optimized by verifying the validity of the virtual monitoring points, which specifically includes the following steps:

[0018] The on-site monitoring area is divided into multiple grids, and a set of virtual monitoring points are evenly arranged within the possible leakage range.

[0019] The geometric center of a group of virtual monitoring points is used as the coordinate origin to locate the virtual leakage source; by rotating the position of each virtual monitoring point, the influence of natural wind direction on leakage diffusion is simulated.

[0020] According to the location of the virtual leakage source, the effectiveness of the virtual monitoring point layout is verified and the location of the virtual monitoring point is optimized.

[0021] Optionally, each virtual monitoring point adopts a fixed rotation monitoring method or a moving point-line monitoring method; the fixed rotation monitoring method is to install the sensor to a fixed point, and the sensor rotates around the fixed point as the axis to achieve different orientations; the moving point-line monitoring method is to install the sensor to a movable horizontal bar, the positions of both ends of the horizontal bar can be moved as needed, and the sensor can move on the horizontal bar to achieve monitoring of multiple positions.

[0022] Optionally, when carrying out scaled real-life modeling of the site, laser scanning and photogrammetry techniques should be used to scan and record the spatial distribution information and data of the site, and the terrain undulations, building distribution and vegetation coverage of the site should be taken into account.

[0023] Optionally, based on the associated data set, a neural network model is trained to obtain a leakage source location model, which specifically includes the following steps:

[0024] The gas concentration collected at each virtual monitoring point is used as input, and the predicted location of the virtual leakage source is obtained through the activation function and output layer output of the neural network model.

[0025] The loss value of the neural network model is determined according to the loss of the virtual leakage source prediction position and the virtual leakage source position.

[0026] The gradient of the neural network model loss value relative to each parameter in the neural network model is calculated through the back-propagation algorithm, and the weights and biases of the neural network model are updated through the optimizer of the neural network model.

[0027] Jump to the step "Use the gas concentration collected at each virtual monitoring point as input, and obtain the predicted location of the virtual leakage source through the output of the neural network model" until the neural network model reaches the predetermined performance standard or the number of training times to obtain the leakage source positioning model.

[0028] Optionally, the neural network model loss value is determined according to the following formula:

[0029]

[0030] in, is the loss value of the neural network model, is the expected output of the neural network model, y is the output of the neural network model, x is the input of the neural network model, and x i is the feature of the i-th sample, y i is the target value of the i-th sample.

[0031] Optionally, the activation function of the neural network model includes but is not limited to Sigmoid function, Relu function, LeakyRelu function and Tanh function; the activation function of the neural network model includes but is not limited to SGD, Adam and RMSprop; the learning rate is required to be set no higher than 0.01; the number of nodes is set no less than 400.

[0032] In the second aspect, the present application provides a point source high-risk gas leakage rapid and accurate tracing and positioning system, including the following modules:

[0033] The on-site real-scene modeling module is used to perform proportional real-scene modeling of the site based on the on-site spatial distribution information and data to obtain a on-site real-scene model.

[0034] The turbulence diffusion model optimization module is used to determine the mesoscale turbulence diffusion model according to the on-site geographical environment information and the characteristics of high-risk gases; the mesoscale turbulence diffusion model is a mesoscale optimization model based on the traditional turbulence diffusion model.

[0035] The monitoring point location optimization module is used to arrange a group of virtual monitoring points in the actual scene on site, and optimize the location of the virtual monitoring points by verifying the validity of the virtual monitoring points.

[0036] The associated data set construction module is used to perform batch numerical simulation based on the on-site real scene model and the mesoscale turbulence diffusion model to obtain an associated data set of the virtual leakage source position and the gas concentration collected at each virtual monitoring point.

[0037] The neural network model training module is used to train a neural network model based on the associated data set to obtain a leakage source positioning model.

[0038] The leakage source precise tracing module is used to collect the gas concentration collected at each actual monitoring point, input it into the leakage source positioning model, output the predicted actual leakage source position, and realize the precise tracing of the leakage source.

[0039] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0040] The present application provides a method and system for rapid and accurate tracing and locating the source of point source high-risk gas leakage. In the method, first, the site is modeled in proportion to the real scene according to the spatial distribution information and data of the site to obtain the real scene model of the site; then, the mesoscale turbulent diffusion model is determined according to the geographical environment information of the site and the characteristics of the high-risk gas; then, a group of virtual monitoring points are arranged in the actual scene of the site, and the positions of the virtual monitoring points are optimized by verifying the effectiveness of the virtual monitoring points, and then, based on the real scene model of the site and the mesoscale turbulent diffusion model, batch numerical simulations are carried out to obtain associated data sets, and a neural network model is trained based on the data to obtain a leakage source positioning model; finally, the gas concentration collected at each actual monitoring point can be input into the leakage source positioning model, and the predicted actual leakage source position is output to achieve accurate tracing of the leakage source. The above-mentioned scheme of the present application optimizes the diffusion model and takes into account the mesoscale structure in turbulence, so that the diffusion model can reflect the eddy motion or scale influence and distinguish between eddy diffusion and molecular diffusion, accurately simulate the diffusion phenomenon, make the obtained data more realistic, and indirectly improve the traceability accuracy; and based on the machine learning method of artificial neural network, establish the association between the concentration distribution and the leakage source location, use the artificial neural network model to take the gas leakage concentration parameter of each monitoring point as the input value, output the precise high-risk gas leakage source location, and realize accurate and rapid traceability of the leakage source. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0042] Figure 1 A flowchart of a method for quickly and accurately tracing and locating a point source high-risk gas leak provided in one embodiment of the present application.

[0043] Figure 2 This is a flowchart of step A3 in a method for quickly and accurately tracing and locating a point source high-risk gas leak provided in one embodiment of the present application.

[0044] Figure 3 A schematic diagram of the layout of virtual monitoring points in a method for rapid and accurate tracing and locating a point source high-risk gas leak provided in one embodiment of the present application.

[0045] Figure 4 This is a flowchart of step A5 in a method for rapid and accurate tracing and locating a point source high-risk gas leak provided in one embodiment of the present application.

[0046] Figure 5A schematic diagram comparing the tracing effects of using CNN and ANN respectively for tracing the source in a method for rapid and accurate tracing and positioning of a point source high-risk gas leak provided in one embodiment of the present application.

[0047] Figure 6 A schematic diagram of the functional modules of a system for rapid and accurate tracing and locating the source of a point source high-risk gas leak provided in one embodiment of the present application.

[0048] Figure 7 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0050] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0051] In an exemplary embodiment, Figure 1 As shown, a method for quickly and accurately tracing and locating a point source high-risk gas leak is provided, comprising the following steps:

[0052] A1. Based on the spatial distribution information and data of the scene, the scene is modeled in proportion to obtain a scene model. Specifically in this embodiment, when the scene is modeled in proportion to the scene, laser scanning and photogrammetry technology should be used to scan and record the spatial distribution information and data of the scene, and factors such as the terrain undulations, building distribution and vegetation coverage of the scene should be taken into account to ensure that the actual scene is restored.

[0053] A2. Determine the mesoscale turbulent diffusion model based on the on-site geographic environment information and the characteristics of high-risk gases; the mesoscale turbulent diffusion model is a mesoscale optimization model based on the traditional turbulent diffusion model. Specifically, the diffusion model is selected based on the on-site geographic environment information (temperature, wind speed, wind direction, ground roughness, etc.) and gas characteristics (density, viscosity, molar mass, etc.), considering the influence of mesoscale turbulent eddies on macroscale flow field distribution and microscale molecular diffusion, and proposes corrections to turbulent viscosity and diffusion coefficient based on mesoscale structures to achieve accurate prediction of gas turbulent diffusion processes.

[0054] In this embodiment, in the mesoscale turbulent diffusion model, the method of the energy minimum multi-scale principle is used to structurally quantify the diffusion field of each sub-phase of the flow field, and the turbulent diffusion coefficient of the mesoscale turbulent diffusion model is corrected; for the laminar sub-phase, the structural quantification indicators of the laminar sub-phase include: the laminar sub-phase flow velocity U lam , fluid density ρ lam , the macroscopic volume fraction 1-u of the laminar subphase and the laminar diffusion coefficient D lam ; For the vortex subphase, the structural quantitative indicators of the vortex subphase include: vortex subphase flow velocity U eddy , fluid density ρ eddy , the macroscopic volume fraction u of the vortex phase and the eddy diffusion coefficient D eddy ; The laminar subphase and the vortex subphase interact at the mesoscale level. The structural quantitative indicators at the mesoscale level include: vortex diameter d eddy and the fluid-eddy interface tension σ.

[0055] The turbulent diffusion coefficient of the mesoscale turbulent diffusion model is corrected according to the following formula:

[0056] D turb =f(U lam ,ρ lam ,1-u,D lam ; U eddy ,ρ eddy ,u,D eddy ;σ,d eddy ).

[0057] Among them, D turb The turbulent diffusion coefficient of the mesoscale turbulent diffusion model is expressed. It is required that the modified diffusion model can accurately reveal the microscopic and mesoscale action laws in numerical simulation calculations.

[0058] A3. Arrange a set of virtual monitoring points in the actual scene, and optimize the location of the virtual monitoring points by verifying the effectiveness of the virtual monitoring points. Figure 2 In the flowchart shown, step A33 specifically includes the following steps:

[0059] A31. Divide the on-site monitoring area into multiple grids and evenly arrange a set of virtual monitoring points within the possible leakage range.

[0060] In order to greatly improve the traceability efficiency and accuracy and realize precise traceability positioning, in this embodiment, each virtual monitoring point adopts a fixed rotation monitoring method or a moving point and line monitoring method; the fixed rotation monitoring method is to install the sensor to a fixed point, and the sensor rotates around the fixed point as the axis to achieve different orientations; the moving point and line monitoring method is to install the sensor to a movable horizontal bar, the positions of the two ends of the horizontal bar can be moved as needed, and the sensor can move on the horizontal bar to realize monitoring of multiple positions.

[0061] A32. Use the geometric center of a group of virtual monitoring points as the coordinate origin to locate the virtual leakage source; by rotating the positions of each virtual monitoring point, simulate the influence of natural wind direction on leakage diffusion. The default arrangement of virtual monitoring points is: a rectangle with the center of the area as the coordinate origin, which is applicable to the wind direction of due north by default, such as Figure 3 In the point arrangement form 1 shown in , if the wind direction deviates from the north by a certain number of degrees, then the corresponding rotation of the rectangle by a certain number of degrees in a certain direction is as follows: Figure 3 The second distribution form is shown in .

[0062] A33. According to the location of the virtual leakage source, verify the effectiveness of the virtual monitoring point layout and optimize the location of the virtual monitoring point.

[0063] A4. Based on the on-site real-life model and the mesoscale turbulent diffusion model, batch numerical simulations were performed to obtain the associated data set of the virtual leakage source location and the gas concentration collected at each virtual monitoring point. When performing batch numerical simulations, two methods based on the Gaussian plume model and CFD simulation were also used. Through the diffusion of the leakage source at different locations and different leakage amounts, the concentration of each monitoring point was extracted and placed in the associated data set.

[0064] A5. Based on the associated data set, a neural network model is trained to obtain a leakage source location model. Specifically, the neural network model during training can adopt neural networks such as BP, LSTM and ANN.

[0065] The training process of the above neural network model includes: the first step, the initialization stage, randomly setting the weights and biases of the neural network model. The second step, the input data flows in the network through the forward propagation process, each neuron processes the input data through the activation function, and passes the result to the next layer until the output layer. The third step, in the output layer, the model's prediction results are compared with the true labels to calculate the loss value. The fourth step, the backpropagation algorithm is used to calculate the gradient of the loss function relative to each parameter in the network. This process involves the application of the chain rule and the derivative of the activation function. The fifth step, the optimizer uses these gradients to update the weights and biases of the network to reduce the loss value. The process from the second to the fifth step is repeated until the model reaches the predetermined performance standard or the number of training times.

[0066] In this embodiment, if Figure 4 In the flowchart shown, step A5 specifically includes the following steps:

[0067] A51. Using the gas concentration collected at each virtual monitoring point as input, the predicted location of the virtual leakage source is obtained through the activation function and output layer output of the neural network model. Specifically, the activation function of the neural network model includes but is not limited to Sigmoid function, Relu function, Leaky Relu function and Tanh function.

[0068] A52. Determine the neural network model loss value according to the loss of the virtual leakage source prediction position and the virtual leakage source position. In this embodiment, the neural network model loss value is determined according to the following formula:

[0069]

[0070] in, is the loss value of the neural network model, is the expected output of the neural network model, y is the output of the neural network model, x is the input of the neural network model, and x i is the feature of the i-th sample, y i is the target value of the ith sample. In this embodiment, the activation function is the Relu function, the optimizer is set to Adam, the learning rate is set to 0.0005, the number of neural network layers is 4, and the number of nodes in each layer is set to 400.

[0071] A53. Calculate the gradient of the loss value of the neural network model relative to each parameter in the neural network model through the back propagation algorithm, and update the weights and biases of the neural network model through the optimizer of the neural network model. In this embodiment, the activation function of the neural network model includes but is not limited to SGD, Adam and RMSprop; the learning rate is required to be set not higher than 0.01; the number of nodes is set not less than 400.

[0072] Jump to step A51 until the neural network model reaches a predetermined performance standard or training times to obtain a leakage source positioning model.

[0073] In some embodiments, the associated data set obtained by the leakage diffusion numerical simulation is divided into a training set, a validation set and a test set. The training set and the validation set are used to train and optimize the network structure and determine the model parameters. The test set is used to test the model performance and generalization ability, and finally determine the network model and the association between the two.

[0074] A6. Collect the gas concentration collected at each actual monitoring point, input it into the leakage source location model, output the predicted actual leakage source location, and achieve accurate tracing of the leakage source. Generally speaking, using a trained neural network model to trace the gas leak, input the gas concentration data from 20 leakage events, and you can get the model's prediction results within 5 seconds. On average, it takes 0.25 seconds for each leakage event to get a preliminary predicted location.

[0075] Specifically, in this embodiment, in order to better compare the tracing effects of different neural network models, convolutional neural networks (CNN) and artificial neural networks (ANN) were used to trace the gas leak source. Figure 5 The traceability results comparison diagram shown is used to compare the traceability effects of the two.

[0076] As an improved embodiment, after tracing the source of high-level gas, the real-time collected data can also be input into the optimized diffusion model to update the model parameters. Once the preliminary leakage source location is determined, a dynamic simulation will be performed to deduce the leakage diffusion situation in real time. Based on the simulation results, the potential risks of gas diffusion to the surrounding environment and personnel will be evaluated.

[0077] The method for rapid and accurate tracing and locating the point source high-risk gas leakage provided in the present embodiment first performs proportional real-scene modeling of the site according to the spatial distribution information and data of the site to obtain a real-scene model of the site; then determines the mesoscale turbulent diffusion model according to the geographical environment information of the site and the characteristics of the high-risk gas; then arranges a group of virtual monitoring points in the actual scene of the site, optimizes the position of the virtual monitoring points by verifying the effectiveness of the virtual monitoring points, and then performs batch numerical simulation according to the real-scene model of the site and the mesoscale turbulent diffusion model to obtain an associated data set, and trains a neural network model based on the data to obtain a leakage source locating model; finally, the gas concentration collected at each actual monitoring point can be input into the leakage source locating model, and the predicted actual leakage source position can be output to achieve accurate tracing of the leakage source, which can be used for rapid and accurate tracing and locating of leakage sources, and support online perception of the risk of hazardous medium leakage in chemical parks, hidden danger investigation and accident emergency rescue.

[0078] Based on the same inventive concept, the embodiment of the present application also provides a system for implementing the method for rapid and accurate tracing and locating the point source high-risk gas leakage involved above. The implementation scheme for solving the problem provided by the system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more system embodiments provided below can refer to the limitations of the method for rapid and accurate tracing and locating the point source high-risk gas leakage above, and will not be repeated here.

[0079] In an exemplary embodiment, Figure 6 As shown, a point source high-risk gas leakage rapid and accurate tracing and positioning system is provided, including the following modules:

[0080] The on-site real-scene modeling module is used to perform proportional real-scene modeling of the site based on the on-site spatial distribution information and data to obtain a on-site real-scene model.

[0081] The turbulence diffusion model optimization module is used to determine the mesoscale turbulence diffusion model according to the on-site geographical environment information and the characteristics of high-risk gases; the mesoscale turbulence diffusion model is a mesoscale optimization model based on the traditional turbulence diffusion model.

[0082] The monitoring point location optimization module is used to arrange a group of virtual monitoring points in the actual scene on site, and optimize the location of the virtual monitoring points by verifying the validity of the virtual monitoring points.

[0083] The associated data set construction module is used to perform batch numerical simulation based on the on-site real scene model and the mesoscale turbulence diffusion model to obtain an associated data set of the virtual leakage source position and the gas concentration collected at each virtual monitoring point.

[0084] The neural network model training module is used to train a neural network model based on the associated data set to obtain a leakage source positioning model.

[0085] The leakage source precise tracing module is used to collect the gas concentration collected at each actual monitoring point, input it into the leakage source positioning model, output the predicted actual leakage source position, and realize the precise tracing of the leakage source.

[0086] certainly, Figure 6 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different functions. Figure 6 One or at least two components of the system shown.

[0087] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 7 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for quickly and accurately tracing and locating a point source high-risk gas leak mentioned in the previous embodiment is implemented.

[0088] Those skilled in the art will understand that Figure 7The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0089] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0090] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0091] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0092] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0093] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0094] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0095] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0096] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for rapid and accurate tracing and locating the source of point source high-risk gas leakage, characterized in that: include: According to the spatial distribution information and data of the site, the site is modeled in proportion to obtain a site model; Determine a mesoscale turbulent diffusion model based on on-site geographic environment information and high-risk gas characteristics; the mesoscale turbulent diffusion model is a mesoscale optimization model based on a traditional turbulent diffusion model; Arranging a set of virtual monitoring points in the actual scene, and optimizing the positions of the virtual monitoring points by verifying the validity of the virtual monitoring points; According to the on-site real-scene model and the mesoscale turbulent diffusion model, batch numerical simulation is performed to obtain a correlation data set of the virtual leakage source position and the gas concentration collected at each virtual monitoring point; Based on the associated data set, a neural network model is trained to obtain a leakage source positioning model; The gas concentration collected at each actual monitoring point is collected and input into the leakage source positioning model, and the predicted actual leakage source position is output to achieve accurate tracing of the leakage source.

2. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 1 is characterized in that: In the mesoscale turbulent diffusion model, the method of the energy minimum multi-scale principle is used to structurally quantify the diffusion field of each sub-phase of the flow field, and to correct the turbulent diffusion coefficient of the mesoscale turbulent diffusion model; For the laminar subphase, the structural quantitative indicators of the laminar subphase include: the laminar subphase flow velocity U lam , fluid density ρ lam , the macroscopic volume fraction 1-u of the laminar subphase and the laminar diffusion coefficient D lam ; For the vortex subphase, the structural quantitative indicators of the vortex subphase include: vortex subphase velocity U eddy , fluid density ρ eddy , the macroscopic volume fraction u of the vortex phase and the eddy diffusion coefficient D eddy ; The laminar subphase and the vortex subphase interact at the mesoscale level. The structural quantitative indicators at the mesoscale level include: vortex diameter d eddy and the fluid-eddy interface tension σ.

3. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 2 is characterized in that: The turbulent diffusion coefficient of the mesoscale turbulent diffusion model is corrected according to the following formula: D turb =f(U lam ,ρ lam ,1-u,D lam ;The eddy ,ρ eddy ,u,D eddy ;σ,d eddy ); Among them, D turb represents the turbulent diffusion coefficient of the mesoscale turbulent diffusion model.

4. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 1 is characterized in that: Arrange a set of virtual monitoring points in the actual scene, and optimize the positions of the virtual monitoring points by verifying the effectiveness of the virtual monitoring points, specifically including: Divide the monitoring area on site into multiple grids and evenly arrange a set of virtual monitoring points within the possible leakage range; The geometric center of a group of virtual monitoring points is used as the coordinate origin to locate the virtual leakage source; the influence of natural wind direction on leakage diffusion is simulated by rotating the position of each virtual monitoring point; According to the virtual leakage source position, the effectiveness of the virtual monitoring point arrangement is verified, and the position of the virtual monitoring point is optimized.

5. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 4 is characterized in that: Each virtual monitoring point adopts a fixed rotation monitoring method or a moving point-line monitoring method; the fixed rotation monitoring method is to install the sensor on a fixed point, and the sensor rotates around the fixed point as the axis to achieve different orientations; the moving point-line monitoring method is to install the sensor on a movable cross bar, the positions of both ends of the cross bar can be moved as needed, and the sensor can move on the cross bar to achieve monitoring of multiple positions.

6. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 1 is characterized in that: When conducting scaled real-life modeling of a site, laser scanning and photogrammetry techniques should be used to scan and record the spatial distribution information and data of the site, and the site's terrain undulations, building distribution, and vegetation coverage should be taken into account.

7. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 1 is characterized in that: Based on the associated data set, a neural network model is trained to obtain a leakage source location model, which specifically includes: The gas concentration collected at each virtual monitoring point is used as input, and the predicted location of the virtual leakage source is obtained through the activation function and output layer output of the neural network model; Determine the loss value of the neural network model according to the loss of the predicted position of the virtual leakage source and the position of the virtual leakage source; The gradient of the loss value of the neural network model relative to each parameter in the neural network model is calculated through the back-propagation algorithm, and the weights and biases of the neural network model are updated through the optimizer of the neural network model; Jump to step "Use the gas concentration collected at each virtual monitoring point as input, and obtain the predicted location of the virtual leakage source through the output of the neural network model" until the neural network model reaches the predetermined performance standard or the number of training times, and obtain the leakage source positioning model.

8. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 7 is characterized in that: The loss value of the neural network model is determined according to the following formula: in, is the loss value of the neural network model, is the expected output of the neural network model, y is the output of the neural network model, x is the input of the neural network model, and x i is the feature of the i-th sample, y i is the target value of the i-th sample.

9. The method for rapid and accurate tracing and locating the point source high-risk gas leakage according to claim 7 is characterized in that: The activation functions of the neural network model include but are not limited to Sigmoid function, Relu function, Leaky Relu function and Tanh function; the activation functions of the neural network model include but are not limited to SGD, Adam and RMSprop; The learning rate should be set no higher than 0.01 and the number of nodes should be set no lower than 400.

10. A rapid and accurate tracing and positioning system for point source high-risk gas leakage, characterized in that: include: The on-site real scene modeling module is used to perform proportional real scene modeling on the site according to the on-site spatial distribution information and data to obtain the on-site real scene model; A turbulence diffusion model optimization module is used to determine a mesoscale turbulence diffusion model according to on-site geographic environment information and high-risk gas characteristics; the mesoscale turbulence diffusion model is a mesoscale optimization model based on a traditional turbulence diffusion model; A monitoring point location optimization module is used to arrange a group of virtual monitoring points in the actual scene on site, and optimize the location of the virtual monitoring points by verifying the validity of the virtual monitoring points; A correlation data set construction module is used to perform batch numerical simulation based on the on-site real scene model and the mesoscale turbulence diffusion model to obtain a correlation data set of the virtual leakage source position and the gas concentration collected at each virtual monitoring point; A neural network model training module is used to train a neural network model based on the associated data set to obtain a leakage source positioning model; The leakage source precise tracing module is used to collect the gas concentration collected at each actual monitoring point, input it into the leakage source positioning model, output the predicted actual leakage source position, and realize the precise tracing of the leakage source.

Images