Intelligent gas pressure pipeline detection system and method

Abstract

The present application relates to the field of pipeline pressure detection, and specifically to an intelligent gas pressure pipeline detection system and method. It uses pressure sensors deployed at multiple locations of the gas pipeline to monitor and collect gas pressure data in real time, and introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the back end to perform time series analysis and collaborative correlation coding on the gas pressure time series data at these different locations, so as to learn and capture the hidden danger correlation characteristics and patterns of the pressure in the gas pipeline in the time and space dimensions, thereby reflecting the operating status information of the gas pipeline, so as to identify and detect abnormal conditions of the gas pipeline pressure. In this way, intelligent monitoring and abnormal pressure detection of the gas pipeline can be realized, so as to timely discover potential problems in the gas pipeline and take corresponding maintenance measures to ensure the safe operation of the gas pipeline.

Description

Technical Field

[0001] The present application relates to the field of pipeline pressure detection, and specifically to an intelligent gas pressure pipeline detection system and method. Background Art

[0002] Gas pipelines are systems of pipes used to transport fuel gases such as natural gas or liquefied petroleum gas. These pipes are usually laid underground or along the walls of buildings to transport gas from where it is produced to where it is consumed, such as homes, factories or commercial buildings.

[0003] Due to long-term use, gas pipelines may have some abnormal conditions, such as pipeline leakage and pipeline damage. If these abnormalities are not discovered and handled in time, they may lead to accidents and cause serious harm to personnel and the environment. Therefore, it is very important to effectively detect gas pipelines to ensure the safe operation of the gas pipeline system.

[0004] However, traditional gas pipeline inspection methods usually rely on manual inspections, which are not only inefficient, but also easily affected by human factors, resulting in negligence or misjudgment. In addition, manual inspections are difficult to achieve high-frequency and comprehensive gas pipeline inspections, resulting in some potential gas pipeline problems not being discovered and handled in a timely manner, leading to further damage to the gas pipeline and safety hazards. In addition, traditional gas pipeline inspection methods can usually only monitor the status of gas pipelines in local areas, and cannot achieve comprehensive monitoring of the entire pipeline system.

[0005] Therefore, an optimized gas pipeline detection system is desired. Summary of the invention

[0006] This application is made in consideration of the above problems. One purpose of this application is to provide an intelligent gas pressure pipeline detection system and method.

[0007] The embodiment of the present application provides an intelligent gas pressure pipeline detection system, which includes: a gas pressure data acquisition module, which is used to obtain a plurality of gas pressure time series collected by pressure sensors deployed at a plurality of positions of a gas pipeline; a gas pressure time series encoding module, which is used to perform time series encoding based on gas pressure on each gas pressure time series of the plurality of gas pressure time series to obtain a plurality of gas pressure time series associated feature vectors; a gas pressure time series feature context semantic enhancement module, which is used to input the plurality of gas pressure time series associated feature vectors into a context semantic enhancement module based on significance-globality to obtain a plurality of enhanced gas pressure time series associated feature vectors; a gas pressure time series feature node information transmission fusion module, which is used to input the plurality of enhanced gas pressure time series associated feature vectors into a node information transmission aggregation network of fusion node confidence to perform node information transmission aggregation based on position node confidence to obtain a gas pipeline pressure position transmission time series aggregation representation vector; a pipeline pressure anomaly detection module, which is used to input the gas pipeline pressure position transmission time series aggregation representation vector into a gas pressure pipeline detection module based on a classifier to obtain a detection result, and the detection result is used to indicate whether there is an abnormality in the pipeline pressure.

[0008] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, the gas pressure time series encoding module is used to: respectively pass each gas pressure time series of the multiple gas pressure time series through a gas pressure sequence encoder based on a Bi-LSTM model to obtain the multiple gas pressure time series associated feature vectors.

[0009] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, the gas pressure time series feature context semantic enhancement module includes: a prominent representation unit, which is used to respectively calculate the mean and maximum value of each gas pressure time series associated feature vector in the multiple gas pressure time series associated feature vectors to obtain a gas pressure time series associated semantic global representation vector composed of multiple vector means and a gas pressure time series associated semantic prominent representation vector composed of multiple vector maximum values; a first feature capture expression unit, which is used to perform feature capture and expression on the gas pressure time series associated semantic global representation vector to obtain a gas pressure time series global semantic activation representation feature vector; a second feature capture expression unit, which is used to perform feature capture on the gas pressure time series associated semantic prominent representation vector The invention relates to a method for obtaining and expressing a gas pressure time series salient semantic activation representation feature vector; a nonlinear activation unit, which is used to fuse the gas pressure time series global semantic activation representation feature vector and the gas pressure time series salient semantic activation representation feature vector to obtain a gas pressure time series global-salient semantic association activation feature vector, and then perform nonlinear activation processing on the gas pressure time series global-salient semantic association activation feature vector to obtain a gas pressure time series global-salient semantic association weight feature vector; and an enhancement unit, which is used to use the gas pressure time series global-salient semantic association weight feature vector as a weight, perform positional point multiplication on the multiple gas pressure time series association feature vectors, and add the multiple gas pressure time series association feature vectors to obtain the multiple enhanced gas pressure time series association feature vectors.

[0010] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, the first feature capture expression unit is used to: perform one-dimensional convolution encoding on the gas pressure temporal association semantic global representation vector and then generate a new feature capture expression unit. The activation function is processed to obtain a global representation vector of the activation expression of the gas pressure timing-related semantics; the activation expression of the gas pressure timing-related semantics global representation vector is again processed through one-dimensional convolution encoding and then matrix multiplied with the first trainable linear transformation weight matrix to obtain the gas pressure timing global semantic activation representation feature vector.

[0011] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, the second feature capture expression unit is used to: perform one-dimensional convolution encoding on the gas pressure time series associated semantic salient representation vector and then generate a new feature capture expression vector. The activation function is used to process to obtain an activation expression gas pressure time series associated semantic salient representation vector; the activation expression gas pressure time series associated semantic salient representation vector is again subjected to one-dimensional convolution encoding and then matrix multiplied with a second trainable linear transformation weight matrix to obtain the gas pressure time series salient semantic activation representation feature vector.

[0012] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, the nonlinear activation unit is used to: fuse the gas pressure time series global semantic activation representation feature vector and the gas pressure time series prominent semantic activation representation feature vector to obtain the gas pressure time series global-prominent semantic association activation feature vector; respectively use the hyperbolic tangent function and The function activates the gas pressure time series global-salient semantic association activation feature vector to obtain the gas pressure time series global-salient feature hyperbolic tangent activation semantic representation vector and the gas pressure time series global-salient feature S-type activation semantic representation vector; the gas pressure time series global-salient feature hyperbolic tangent activation semantic representation vector and the gas pressure time series global-salient feature S-type activation semantic representation vector are multiplied by position point to obtain the gas pressure time series global-salient semantic association weight feature vector.

[0013] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, the gas pressure time series feature node information transmission and fusion module includes: a labeling unit, which is used to label the first k-1 enhanced gas pressure time series associated feature vectors among the multiple enhanced gas pressure time series associated feature vectors with serial numbers to obtain a sequence of gas pressure time series associated semantic labeling serial numbers; a mean and variance calculation unit, which is used to calculate the energy function values ​​of the first k-1 enhanced gas pressure time series associated feature vectors to obtain k-1 energy function values, wherein the energy function value is related to the mean and variance of each enhanced gas pressure time series associated feature vector and the labeling serial number of each enhanced gas pressure time series associated feature vector in the sequence; a normalization unit, which is used to calculate the energy function values ​​of the first k-1 enhanced gas pressure time series associated feature vectors to obtain k-1 energy function values. energy function values ​​are normalized to obtain k-1 node confidences; an aggregation unit is used to calculate the position-weighted sum of the first k-1 enhanced gas pressure time series association feature vectors among the multiple enhanced gas pressure time series association feature vectors with the k-1 node confidences as weights to aggregate the first k-1 enhanced gas pressure time series association feature vectors among the multiple enhanced gas pressure time series association feature vectors to obtain a gas pressure time series association semantic rear node set aggregation representation vector; a weighted sum unit is used to calculate the weighted sum between the gas pressure time series association semantic rear node set aggregation representation vector and the kth enhanced gas pressure time series association feature vector among the multiple enhanced gas pressure time series association feature vectors to obtain the gas pipeline pressure position transmission time series aggregation representation vector.

[0014] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, the mean and variance calculation unit is used to: calculate the mean and variance of the predetermined enhanced gas pressure time series association feature vectors among the first k-1 enhanced gas pressure time series association feature vectors to obtain the gas pressure time series association semantic mean and the gas pressure time series association semantic variance; determine the gas pressure time series association semantic annotation serial number of the predetermined enhanced gas pressure time series association feature vector to obtain the predetermined gas pressure time series association semantic annotation serial number; calculate the sum of the gas pressure time series association semantic mean and the gas pressure time series association semantic variance and then multiply it by a constant 4 to obtain a first gas pressure time series association semantic modulation factor; calculate the gas pressure time series association semantic mean and the gas pressure time series association semantic variance respectively. The product of the gas pressure time series association semantic mean and the gas pressure time series association semantic variance and the constant 2 is used to obtain the double gas pressure time series association semantic variance and the double gas pressure time series association semantic mean; after calculating the square of the difference between the gas pressure time series association semantic annotation serial number and the gas pressure time series association semantic mean, it is added with the double gas pressure time series association semantic variance and the double gas pressure time series association semantic mean to obtain the second gas pressure time series association semantic modulation factor; the division between the first gas pressure time series association semantic modulation factor and the second gas pressure time series association semantic modulation factor is calculated to obtain the energy function value corresponding to the predetermined enhanced gas pressure time series association feature vector.

[0015] For example, according to the intelligent gas pressure pipeline detection system of the embodiment of the present application, it also includes a training module for training the gas pressure sequence encoder based on the Bi-LSTM model, the contextual semantic enhancement module based on significance-globality, the node information transmission aggregation network fused with node confidence, and the gas pressure pipeline detection module based on the classifier; wherein the training module includes: a training data acquisition unit, used to obtain training data, the training data includes multiple training gas pressure time series and real detection results collected by pressure sensors deployed at multiple positions of the gas pipeline; a training gas pressure time series encoding unit, used to pass each training gas pressure time series of the multiple training gas pressure time series through the gas pressure sequence encoder based on the Bi-LSTM model to obtain multiple training gas pressure time series associated feature vectors; a training gas pressure time series feature contextual semantic enhancement unit, used to input the multiple training gas pressure time series associated feature vectors into the contextual semantic enhancement module based on significance-globality; a training gas pressure time series feature node information transmission fusion unit, used to input the multiple training enhanced gas pressure time series associated feature vectors into the node information transmission aggregation network of the fusion node confidence to perform node information transmission aggregation based on the position node confidence to obtain a training gas pipeline pressure position transmission time series aggregation representation vector; a training pipeline pressure anomaly detection unit, used to input the training gas pipeline pressure position transmission time series aggregation representation vector into the classifier-based gas pressure pipeline detection module to obtain a training detection result; a classification loss function value calculation unit, used to calculate the cross entropy loss function value between the training detection result and the real detection result to obtain a classification loss function value; a loss training unit, used to train the gas pressure sequence encoder based on the Bi-LSTM model, the context semantic reinforcement module based on significance-globality, the node information transmission aggregation network of the fusion node confidence and the classifier-based gas pressure pipeline detection module based on the classification loss function value.

[0016] An embodiment of the present application also provides an intelligent gas pressure pipeline detection method, which includes: obtaining multiple gas pressure time series collected by pressure sensors deployed at multiple positions of the gas pipeline; performing gas pressure-based time series encoding on each gas pressure time series in the multiple gas pressure time series to obtain multiple gas pressure time series associated feature vectors; inputting the multiple gas pressure time series associated feature vectors into a contextual semantic enhancement module based on significance-globality to obtain multiple enhanced gas pressure time series associated feature vectors; performing node information transfer aggregation based on position node confidence on the multiple enhanced gas pressure time series associated feature vectors to obtain a gas pipeline pressure position transfer time series aggregation representation vector; inputting the gas pipeline pressure position transfer time series aggregation representation vector into a classifier-based gas pressure pipeline detection module to obtain a detection result, and the detection result is used to indicate whether there is an abnormality in the pipeline pressure.

[0017] According to the intelligent gas pressure pipeline detection system and method of the embodiments of the present application, it is possible to realize intelligent monitoring and abnormal pressure detection of the gas pipeline, so as to timely discover potential problems of the gas pipeline and take corresponding maintenance measures to ensure the safe operation of the gas pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0019] Figure 1 A schematic diagram of the application architecture of an intelligent gas pressure pipeline detection system in an embodiment of the present application is shown.

[0020] Figure 2 A structural schematic diagram of an intelligent gas pressure pipeline detection system in an embodiment of the present application is shown.

[0021] Figure 3 A structural schematic diagram of a gas pressure time series feature context semantic enhancement module of an intelligent gas pressure pipeline detection system in an embodiment of the present application is shown.

[0022] Figure 4 A flow chart of an intelligent gas pressure pipeline detection method in an embodiment of the present application is shown.

[0023] Figure 5 The application scenario diagram of the intelligent gas pressure pipeline detection system in the embodiment of the present application is shown.

[0024] Figure 6 A schematic diagram of a storage medium according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0026] The terms used in this specification are those common terms currently widely used in the art in consideration of the functions of the present application, but these terms may vary according to the intention of a person of ordinary skill in the art, precedents, or new technologies in the art. In addition, specific terms may be selected by the applicant, and in this case, their detailed meanings will be described in the detailed description of the present application. Therefore, the terms used in the specification should not be understood as simple names, but rather as a general description based on the meaning of the terms and the present application.

[0027] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0028] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0029] Figure 1 A schematic diagram of the application architecture of the intelligent gas pressure pipeline detection system in an embodiment of the present application is shown, including a server 100 and a terminal device 200.

[0030] The terminal device 200 and the server 100 can be connected via the Internet to achieve mutual communication. Optionally, the above-mentioned Internet uses standard communication technology and / or protocol. The Internet is usually the Internet, but it can also be any network, including but not limited to any combination of a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), a mobile, wired or wireless network, a private network or a virtual private network. In some embodiments, the data exchanged through the network can be represented by technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc. In addition, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), Internet Protocol Security (IPsec), etc. can be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can also be used to replace or supplement the above-mentioned data communication technologies.

[0031] The server 100 can provide various network services for the terminal device 200, wherein the server 100 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Specifically, the server 100 may include a processor 110 (Center Processing Unit, CPU), a memory 120, an input device 130, and an output device 140, etc. The input device 130 may include a keyboard, a mouse, a touch screen, etc., and the output device 140 may include a display device, such as a liquid crystal display (Liquid Crystal Display, LCD), a cathode ray tube (Cathode Ray Tube, CRT), etc.

[0032] The memory 120 may include a read-only memory (ROM) and a random access memory (RAM), and provides the processor 110 with program instructions and data stored in the memory 120. In the embodiment of the present application, the memory 120 may be used to store the corresponding method of the intelligent gas pressure pipeline detection system in the embodiment of the present application.

[0033] The processor 110 calls the program instructions stored in the memory 120, and the processor 110 is used to execute the steps of any one of the intelligent gas pressure pipeline detection methods in the embodiments of the present application according to the obtained program instructions.

[0034] In addition, the application architecture diagram in the embodiment of the present application is intended to more clearly illustrate the technical solution in the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present application is also applicable to similar problems.

[0035] The following is a non-restrictive description of the intelligent gas pressure pipeline detection system provided according to at least one embodiment of the present application through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also fall within the scope of protection of the present application.

[0036] It should be understood that the gas pipeline system is usually a complex network, and different locations of the pipeline have different pressure changes. At the same time, the pressure data at these different locations may be affected by different factors, such as pipeline length, pipeline material, pipeline connection method, etc. Therefore, the time series changes and collaborative information of the pressure data at different locations of the gas pipeline can well reflect the operating status and potential safety hazards of the gas pipeline, which helps to improve the accuracy of pipeline anomaly detection, so as to timely discover the safety hazards of the gas pipeline, so as to make targeted maintenance decisions, ensure the safe and reliable operation of the gas pipeline, and avoid accidents.

[0037] Based on this, in the technical solution of this application, an intelligent gas pressure pipeline detection system is proposed, which can monitor and collect gas pressure data in real time through pressure sensors deployed at multiple locations of the gas pipeline, and introduce data processing and analysis algorithms based on artificial intelligence and deep learning at the back end to perform time series analysis and collaborative correlation coding on the gas pressure time series data at these different locations, so as to learn and capture the hidden danger correlation characteristics and patterns of the pressure in the gas pipeline in the time and space dimensions, thereby reflecting the operating status information of the gas pipeline, so as to identify and detect abnormal conditions of the gas pipeline pressure. In this way, intelligent monitoring and abnormal pressure detection of the gas pipeline can be realized, so as to timely discover potential problems of the gas pipeline and take corresponding maintenance measures to ensure the safe operation of the gas pipeline.

[0038] Figure 2A structural schematic diagram of an intelligent gas pressure pipeline detection system 800 in an embodiment of the present application is shown. The intelligent gas pressure pipeline detection system 800 includes: a gas pressure data acquisition module 810, which is used to obtain multiple gas pressure time series collected by pressure sensors deployed at multiple positions of the gas pipeline; a gas pressure time series encoding module 820, which is used to perform time series encoding based on gas pressure on each gas pressure time series of the multiple gas pressure time series to obtain multiple gas pressure time series associated feature vectors; a gas pressure time series feature context semantic enhancement module 830, which is used to input the multiple gas pressure time series associated feature vectors into a context semantic enhancement module based on significance-globality to obtain multiple enhanced gas pressure time series associated feature vectors; a gas pressure time series feature node information transmission fusion module 840, which is used to input the multiple enhanced gas pressure time series associated feature vectors into a node information transmission aggregation network of fusion node confidence to perform node information transmission aggregation based on position node confidence to obtain a gas pipeline pressure position transmission time series aggregation representation vector; a pipeline pressure anomaly detection module 850, which is used to input the gas pipeline pressure position transmission time series aggregation representation vector into a gas pressure pipeline detection module based on a classifier to obtain a detection result, and the detection result is used to indicate whether there is an abnormality in the pipeline pressure.

[0039] Specifically, in the technical solution of the present application, first, multiple time series of gas pressures collected by pressure sensors deployed at multiple locations of the gas pipeline are obtained. Next, considering that the gas pressure data at each location of the gas pipeline has a time series change law in the time dimension, in order to be able to capture the gas pressure time series characteristics and change patterns at each gas pipeline location respectively, so as to provide time series characteristics and information for the subsequent pipeline pressure anomaly detection task, in the technical solution of the present application, the time series of each gas pressure in the multiple gas pressure time series is respectively passed through the gas pressure sequence encoder based on the Bi-LSTM model to obtain multiple gas pressure time series associated feature vectors. It should be understood that a significant feature of the Bi-LSTM model is that it can process past and future information at the same time. This means that the model not only considers the impact of historical pipeline pressure values ​​on the current state, but also considers the potential impact of future trends on the current state. Therefore, by applying the gas pressure sequence encoder based on the Bi-LSTM model, the forward and backward time series features and long-distance dependencies of the gas pressure at multiple locations of the gas pipeline can be extracted respectively to reflect the gas pressure change pattern at different gas pipeline locations, providing support for subsequent analysis and decision-making.

[0040] Correspondingly, the gas pressure time series encoding module 820 is used to: respectively pass each gas pressure time series of the multiple gas pressure time series through a gas pressure sequence encoder based on a Bi-LSTM model to obtain the multiple gas pressure time series associated feature vectors.

[0041] Then, considering that the pressure time series change of the gas pipeline is not only affected by local factors, but also by the overall state of the pipeline system, for example, if a certain section of the pipeline has a leakage problem, it will be manifested as the pipeline pressure data in the subsequent local pipeline showing a downward trend, and this downward trend may become larger and larger. Based on this, in order to be able to perform a spatial global analysis and feature extraction on the gas pressure time series features at different locations of these gas pipelines, and to identify the contribution of the pressure time series semantics at each pipeline location to the subsequent pipeline pressure anomaly detection task, in the technical solution of the present application, the multiple gas pressure time series associated feature vectors are input into the contextual semantic enhancement module based on significance-globality to obtain multiple enhanced gas pressure time series associated feature vectors. Through the processing of the saliency-global contextual semantic enhancement module, the contextual correlation features and association relationships between the gas pressure time series features at each gas pipeline location based on spatial globality can be automatically learned and captured, and which features contribute more to the subsequent pipeline pressure anomaly detection task can be identified, thereby performing significant enhancement expression on these pipeline gas pressure time series features with greater contributions, so that the model can be better generalized to different environments and conditions, thereby improving the applicability of the detection system under different situations.

[0042] Accordingly, if Figure 3As shown, the gas pressure time series feature context semantic enhancement module 830 includes: a highlight representation unit 831, which is used to respectively calculate the mean and maximum value of each gas pressure time series associated feature vector in the multiple gas pressure time series associated feature vectors to obtain a gas pressure time series associated semantic global representation vector composed of multiple vector means and a gas pressure time series associated semantic highlight representation vector composed of multiple vector maximum values; a first feature capture expression unit 832, which is used to perform feature capture and expression on the gas pressure time series associated semantic global representation vector to obtain a gas pressure time series global semantic activation representation feature vector; a second feature capture expression unit 833, which is used to perform feature capture and expression on the gas pressure time series associated semantic highlight representation vector to obtain a gas pressure A time series salient semantic activation representation feature vector; a nonlinear activation unit 834, which is used to fuse the gas pressure time series global semantic activation representation feature vector and the gas pressure time series salient semantic activation representation feature vector to obtain a gas pressure time series global-salient semantic association activation feature vector, and then perform nonlinear activation processing on the gas pressure time series global-salient semantic association activation feature vector to obtain a gas pressure time series global-salient semantic association weight feature vector; an enhancement unit 835, which is used to use the gas pressure time series global-salient semantic association weight feature vector as a weight, perform positional point multiplication on the multiple gas pressure time series association feature vectors and add the multiple gas pressure time series association feature vectors to obtain the multiple enhanced gas pressure time series association feature vectors.

[0043] The first feature capture and expression unit 832 is used to: perform one-dimensional convolution encoding on the gas pressure temporal association semantic global representation vector and then generate a The activation function is processed to obtain a global representation vector of the activation expression of the gas pressure timing-related semantics; the activation expression of the gas pressure timing-related semantics global representation vector is again processed through one-dimensional convolution encoding and then matrix multiplied with the first trainable linear transformation weight matrix to obtain the gas pressure timing global semantic activation representation feature vector.

[0044] The second feature capture and expression unit 833 is used to: perform one-dimensional convolution encoding on the gas pressure time series associated semantic salient representation vector and then generate a new feature capture and expression unit 833; The activation function is used to process to obtain an activation expression gas pressure time series associated semantic salient representation vector; the activation expression gas pressure time series associated semantic salient representation vector is again subjected to one-dimensional convolution encoding and then matrix multiplied with a second trainable linear transformation weight matrix to obtain the gas pressure time series salient semantic activation representation feature vector.

[0045] The nonlinear activation unit 834 is used to: fuse the gas pressure time series global semantic activation representation feature vector and the gas pressure time series prominent semantic activation representation feature vector to obtain the gas pressure time series global-prominent semantic association activation feature vector; respectively use the hyperbolic tangent function and The function activates the gas pressure time series global-salient semantic association activation feature vector to obtain the gas pressure time series global-salient feature hyperbolic tangent activation semantic representation vector and the gas pressure time series global-salient feature S-type activation semantic representation vector; the gas pressure time series global-salient feature hyperbolic tangent activation semantic representation vector and the gas pressure time series global-salient feature S-type activation semantic representation vector are multiplied by position point to obtain the gas pressure time series global-salient semantic association weight feature vector.

[0046] In a specific example, the gas pressure time series feature context semantic enhancement module 830 is used to: input the multiple gas pressure time series associated feature vectors into the context semantic enhancement module based on significance-globality and process them with the following significance enhancement formula to obtain the multiple enhanced gas pressure time series associated feature vectors; wherein the significance enhancement formula is: ;in, are the multiple gas pressure time series associated feature vectors, represents the global mean operation in the calculation vector, Represents the maximum value operation in the vector. is the gas pressure time series associated semantic mean representation vector, is the maximum value representation vector of gas pressure time series associated semantics, represents one-dimensional convolutional coding, express Activation function, and They represent the trainable weight matrices for linear transformation, and They are the global representation feature vector of gas pressure time series association semantics and the prominent representation feature vector of gas pressure time series association semantics, for function, for function, It means point multiplication by position. It is the plurality of enhanced gas pressure time series associated characteristic vectors.

[0047] It should be understood that in the gas pipeline system, there is a certain correlation and transferability between the pressure data provided by monitoring points at different locations, and they have different qualities and importance. Therefore, in order to more effectively integrate the gas pressure time series features from different gas pipeline monitoring points and ensure that more reliable data and features have a greater weight in analysis and subsequent detection tasks, in the technical solution of the present application, the multiple enhanced gas pressure time series correlation feature vectors are further input into the node information transmission aggregation network of the fused node confidence to obtain the gas pipeline pressure position transmission time series aggregation representation vector. By processing the node information transmission aggregation network of the fused node confidence, the gas pressure time series features at different gas pipeline monitoring nodes after the enhanced expression of significant features can be correlated and globally analyzed, so as to understand and identify the node semantic information with higher confidence in the multiple enhanced gas pressure time series associated feature vectors, and pay higher attention to these node semantic features with higher confidence, and pay less attention to the monitoring node features with lower confidence, so that in the process of fusing the gas pressure time series features between different pipeline space nodes, the key information in different node semantics can be better transmitted and aggregated, which helps to have a more comprehensive understanding of the status of the entire pipeline system and provide a basis for subsequent pipeline pressure abnormality detection.

[0048] Correspondingly, the gas pressure timing feature node information transmission fusion module 840 includes: a labeling unit, which is used to label the first k-1 enhanced gas pressure timing associated feature vectors among the multiple enhanced gas pressure timing associated feature vectors with serial numbers to obtain a sequence of gas pressure timing associated semantic labeling serial numbers; a mean and variance calculation unit, which is used to calculate the energy function values ​​of the first k-1 enhanced gas pressure timing associated feature vectors to obtain k-1 energy function values, wherein the energy function value is related to the mean and variance of each enhanced gas pressure timing associated feature vector and the labeling serial number of each enhanced gas pressure timing associated feature vector in the sequence; a normalization unit, which is used to normalize the k-1 energy function values. Processing to obtain k-1 node confidences; an aggregation unit, used to calculate the position-weighted sum of the first k-1 enhanced gas pressure timing association feature vectors in the multiple enhanced gas pressure timing association feature vectors with the k-1 node confidences as weights to aggregate the first k-1 enhanced gas pressure timing association feature vectors in the multiple enhanced gas pressure timing association feature vectors to obtain the gas pressure timing association semantic rear node set aggregation representation vector; a weighted sum unit, used to calculate the weighted sum between the gas pressure timing association semantic rear node set aggregation representation vector and the kth enhanced gas pressure timing association feature vector in the multiple enhanced gas pressure timing association feature vectors to obtain the gas pipeline pressure position transmission timing aggregation representation vector.

[0049] The mean and variance calculation unit is used to: calculate the mean and variance of the predetermined enhanced gas pressure time series association feature vectors among the first k-1 enhanced gas pressure time series association feature vectors to obtain the gas pressure time series association semantic mean and the gas pressure time series association semantic variance; determine the gas pressure time series association semantic annotation sequence number of the predetermined enhanced gas pressure time series association feature vector to obtain the predetermined gas pressure time series association semantic annotation sequence number; calculate the sum of the gas pressure time series association semantic mean and the gas pressure time series association semantic variance and then multiply it by a constant 4 to obtain a first gas pressure time series association semantic modulation factor; calculate the gas pressure time series association semantic mean and the gas pressure time series association semantic variance respectively. The product of the gas pressure time series association semantic variance and the constant 2 is used to obtain the double gas pressure time series association semantic variance and the double gas pressure time series association semantic mean; after calculating the square of the difference between the gas pressure time series association semantic annotation serial number and the gas pressure time series association semantic mean, it is added with the double gas pressure time series association semantic variance and the double gas pressure time series association semantic mean to obtain the second gas pressure time series association semantic modulation factor; the division between the first gas pressure time series association semantic modulation factor and the second gas pressure time series association semantic modulation factor is calculated to obtain the energy function value corresponding to the predetermined enhanced gas pressure time series association feature vector.

[0050] In a specific example, the gas pressure time series feature node information transmission fusion module 840 is used to: input the multiple enhanced gas pressure time series associated feature vectors into the node information transmission aggregation network of the fusion node confidence and process them with the following node information transmission formula to obtain the gas pipeline pressure position transmission time series aggregation representation vector; wherein the node information transmission formula is: ;in, is the first of the plurality of enhanced gas pressure time series associated feature vectors Enhanced gas pressure time series correlation feature vector, For each of the enhanced gas pressure time series associated feature vectors, Position feature values, is the length of each enhanced gas pressure time series associated characteristic vector, is the mean value of each enhanced gas pressure time series associated characteristic vector, is the variance of each enhanced gas pressure time series associated feature vector, is the regularization hyperparameter, A serial number is given to each enhanced gas pressure time series associated characteristic vector in the sequence, is the energy function value of each enhanced gas pressure time series associated characteristic vector, for function, The node confidence of each enhanced gas pressure time series associated feature vector, is the first of the plurality of enhanced gas pressure time series associated feature vectors Enhanced gas pressure time series correlation feature vector, A time series aggregation representation vector for the gas pipeline pressure position transmission.

[0051] Furthermore, the gas pipeline pressure position transmission time series aggregation representation vector is input into the classifier-based gas pressure pipeline detection module to obtain the detection result, and the detection result is used to indicate whether there is an abnormality in the pipeline pressure. In other words, the pressure time series semantic information transmission aggregation characteristics at different positions in the gas pipeline are used for classification processing, so as to understand the operating status of the gas pipeline more comprehensively and accurately, and thus identify and detect abnormal conditions of the gas pipeline pressure. In this way, intelligent monitoring and abnormal pressure detection of the gas pipeline can be realized, so as to timely discover potential problems of the gas pipeline and take corresponding maintenance measures to ensure the safe operation of the gas pipeline.

[0052] Here, for the local time domain short-range-long-range bidirectional time series association features of the gas pressure expressed by each of the multiple gas pressure time series association feature vectors, when performing saliency-global-based contextual semantic reinforcement and node information transmission aggregation, due to the saliency-global contextual reinforcement weight difference and local time domain time series node information confidence transmission difference caused by the distribution difference between local time domains under the global time domain of the time series features, the aggregation key information of the obtained gas pipeline pressure position transmission time series aggregation representation vector will be lost, affecting the expression effect of the gas pipeline pressure position transmission time series aggregation representation vector.

[0053] Based on this, in the preferred example, when the gas pipeline pressure position transfer time series aggregation representation vector passes through the classifier-based gas pressure pipeline detection module to obtain the detection result, the gas pipeline pressure position transfer time series aggregation representation vector is optimized. Specifically, the optimization process includes the following steps: clustering all feature values ​​of the gas pipeline pressure position transfer time series aggregation representation vector based on the distance between feature values, such as the difference absolute value distance, and determining the ratio value of the feature number of cluster features relative to the feature number of the gas pipeline pressure position transfer time series aggregation representation vector; arranging the cluster features into a gas pipeline pressure position transfer time series aggregation cluster vector, and dividing the second norm of the gas pipeline pressure position transfer time series aggregation cluster vector by the second norm of the gas pipeline pressure position transfer time series aggregation representation vector to obtain a first gas pipeline pressure position transfer time series aggregation cluster related weight value; dividing the first power value of the gas pipeline pressure position transfer time series aggregation cluster vector exponentially by the first power value of the gas pipeline pressure position transfer time series aggregation representation vector a second power value of the ratio value as an exponent to obtain a second gas pipeline pressure position transmission timing series aggregation clustering related weight value; for each eigenvalue of the gas pipeline pressure position transmission timing series aggregation clustering vector, multiply it by the reciprocal of the difference between the first gas pipeline pressure position transmission timing series aggregation clustering related weight value and the second gas pipeline pressure position transmission timing series aggregation clustering related weight value to obtain the optimized eigenvalue of the gas pipeline pressure position transmission timing series aggregation clustering vector; for each eigenvalue outside the cluster in the gas pipeline pressure position transmission timing series aggregation representation vector, multiply it by the reciprocal of the sum of the first gas pipeline pressure position transmission timing series aggregation clustering related weight value and the second gas pipeline pressure position transmission timing series aggregation clustering related weight value to obtain the optimized out-of-class eigenvalue of the gas pipeline pressure position transmission timing series aggregation representation vector; the optimized eigenvalue of the gas pipeline pressure position transmission timing series aggregation clustering vector and the optimized out-of-class eigenvalue of the gas pipeline pressure position transmission timing series aggregation representation vector are combined to form an optimized gas pipeline pressure position transmission timing series aggregation representation vector.

[0054] That is, the optimization process is expressed as: ;in, is the gas pipeline pressure position transmission time series aggregation representation vector, is the number of eigenvalues ​​of the gas pipeline pressure position transmission time series aggregation representation vector, is the gas pipeline pressure position transmission time series aggregation clustering vector, is the number of eigenvalues ​​of the gas pipeline pressure position transmission time series aggregation clustering vector, is the ratio value, represents the clustering feature set corresponding to the gas pipeline pressure position transmission time series aggregation clustering vector, and Represents the binary norm and the uninorm of the vector respectively. Power, is the characteristic value of each position in the gas pipeline pressure position transmission time series aggregation representation vector, It is the characteristic value of each position in the optimized gas pipeline pressure position transmission time series aggregation representation vector.

[0055] Here, in order to avoid the loss of key suffix semantic information relative to the original feature set as a whole due to aggregation conflict caused by the aggregation feature based on the gas pipeline pressure position transmission time series aggregation representation vector, the adversarial decision of the absolute representation of the set of the gas pipeline pressure position transmission time series aggregation clustering vector and the one-norm aggregation representation vector of the gas pipeline pressure position transmission time series is performed by taking the clustering ratio of the number of features of the gas pipeline pressure position transmission time series aggregation clustering vector relative to the number of features of the gas pipeline pressure position transmission time series aggregation representation vector as the decision function, and respectively comparing the absolute representation of the set of the gas pipeline pressure position transmission time series aggregation clustering vector and the gas pipeline pressure position transmission time series aggregation representation vector with the gas pipeline pressure position transmission time series aggregation clustering vector. The clustering intrinsic conflict representation of the two norms of the gas pipeline pressure position transmission time series aggregation representation vector is positively and negatively interacted to construct the optimized gas pipeline pressure position transmission time series aggregation representation vector based on the aggregation feature and the original feature set as a whole. The solid alignment guardrail is used to achieve the intentional mitigation of the harmful information loss of the optimized gas pipeline pressure position transmission time series aggregation representation vector based on the aggregation risk transferability, and improve the expression effect of the optimized gas pipeline pressure position transmission time series aggregation representation vector, thereby improving the accuracy of the detection results obtained by the gas pipeline pressure position transmission time series aggregation representation vector through the classifier-based gas pressure pipeline detection module. In this way, more accurate monitoring and abnormal identification of gas pipeline pressure can be achieved, so as to timely discover potential problems of the gas pipeline and take corresponding maintenance measures to ensure the safe operation of the gas pipeline.

[0056] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules, and then classify (or predict) unknown data. Logistic regression, SVM, etc. are often used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-classification, but this is prone to errors and is not efficient. Commonly used multi-classification methods include the Softmax classification function.

[0057] Furthermore, in the technical solution of the present application, the intelligent gas pressure pipeline detection system also includes a training module for training the gas pressure sequence encoder based on the Bi-LSTM model, the contextual semantic enhancement module based on significance-globality, the node information transmission aggregation network fused with node confidence, and the classifier-based gas pressure pipeline detection module.

[0058] The training module includes: a training data acquisition unit for acquiring training data, wherein the training data includes a plurality of training gas pressure time series and real detection results collected by pressure sensors deployed at a plurality of positions of the gas pipeline; a training gas pressure time series encoding unit for respectively passing each training gas pressure time series in the plurality of training gas pressure time series through the gas pressure sequence encoder based on the Bi-LSTM model to obtain a plurality of training gas pressure time series associated feature vectors; a training gas pressure time series feature context semantic enhancement unit for inputting the plurality of training gas pressure time series associated feature vectors into the context semantic enhancement module based on significance-globality to obtain a plurality of training enhanced gas pressure time series associated feature vectors; a training gas pressure time series feature node information transmission fusion unit for transmitting the plurality of training enhanced gas pressure time series associated feature vectors into the context semantic enhancement module based on significance-globality to obtain a plurality of training enhanced gas pressure time series associated feature vectors; The feature vector is input into the node information transmission aggregation network of the fused node confidence to perform node information transmission aggregation based on the position node confidence to obtain the training gas pipeline pressure position transmission time series aggregation representation vector; the training pipeline pressure anomaly detection unit is used to input the training gas pipeline pressure position transmission time series aggregation representation vector into the gas pressure pipeline detection module based on the classifier to obtain the training detection result; the classification loss function value calculation unit is used to calculate the cross entropy loss function value between the training detection result and the real detection result to obtain the classification loss function value; the loss training unit is used to train the gas pressure sequence encoder based on the Bi-LSTM model, the context semantic reinforcement module based on significance-globality, the node information transmission aggregation network of the fused node confidence and the gas pressure pipeline detection module based on the classifier based on the classification loss function value.

[0059] Based on the above embodiments, see Figure 4 FIG. 1 is a flow chart of a smart gas pressure pipeline detection method in an embodiment of the present application. For example, the smart gas pressure pipeline detection method can be executed by a server, which can be Figure 1 The server 100 shown in FIG. Figure 4As shown, according to the embodiment of the present application, the intelligent gas pressure pipeline detection method includes the following steps: S510, obtaining multiple gas pressure time series collected by pressure sensors deployed at multiple positions of the gas pipeline; S520, performing gas pressure-based time series encoding on each gas pressure time series in the multiple gas pressure time series to obtain multiple gas pressure time series associated feature vectors; S530, inputting the multiple gas pressure time series associated feature vectors into a context semantic enhancement module based on significance-globality to obtain multiple enhanced gas pressure time series associated feature vectors; S540, performing node information transfer aggregation based on position node confidence on the multiple enhanced gas pressure time series associated feature vectors to obtain a gas pipeline pressure position transfer time series aggregation representation vector; S550, inputting the gas pipeline pressure position transfer time series aggregation representation vector into a classifier-based gas pressure pipeline detection module to obtain a detection result, and the detection result is used to indicate whether there is an abnormality in the pipeline pressure.

[0060] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned intelligent gas pressure pipeline detection method have been referred to above. Figure 2 to Figure 3 The intelligent gas pressure pipeline detection system 800 is described in detail, and therefore, its repeated description will be omitted.

[0061] Figure 5 FIG. 1 is an application scenario diagram of the intelligent gas pressure pipeline detection system according to an embodiment of the present application. Figure 5 As shown, in this application scenario, first, a plurality of time series of gas pressures collected by pressure sensors deployed at a plurality of positions of the gas pipeline are obtained (for example, Figure 5 D), and then, the time series of the multiple gas pressures are input to a server (for example, Figure 5 In S) shown in , the server can use the intelligent gas pressure pipeline detection algorithm to process the time series of the multiple gas pressures to obtain a detection result indicating whether there is an abnormality in the pipeline pressure.

[0062] Based on the above embodiments, another exemplary embodiment of an electronic device is provided in the embodiments of the present application. In some possible implementations, the electronic device in the embodiments of the present application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement the steps of the intelligent gas pressure pipeline detection method in the above embodiments when executing the program.

[0063] For example, in the case of electronic equipment Figure 1Taking the server 100 in the example as an example, the processor in the electronic device is the processor 110 in the server 100, and the memory in the electronic device is the memory 120 in the server 100.

[0064] An embodiment of the present application also provides a computer-readable storage medium. Figure 6 1 shows a schematic diagram of a computer-readable storage medium 1000 according to an embodiment of the present application. Figure 6 As shown, the computer-readable storage medium 1000 stores computer-executable instructions 1001. When the computer-executable instructions 1001 are executed by the processor, the intelligent gas pressure pipeline detection method according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc.

[0065] The embodiment of the present application also provides a computer program product or a computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer readable storage medium. The processor of the computer device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the computer device executes the intelligent gas pressure pipeline detection method according to the embodiment of the present application.

[0066] Those skilled in the art will appreciate that the contents disclosed in this application may be subject to various variations and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0067] In addition, although the present application makes various references to certain units in the system according to embodiments of the present application, any number of different units can be used and run on the client and / or server. The units are only illustrative, and different aspects of the system and method can use different units.

[0068] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present application is not limited to any particular form of combination of hardware and software.

[0069] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.

[0070] The above is an explanation of the present application and should not be considered as a limitation thereof. Although several exemplary embodiments of the present application have been described, it will be readily understood by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application. Therefore, all of these modifications are intended to be included within the scope of the present application as defined by the claims. It should be understood that the above is an explanation of the present application and should not be considered to be limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present application is defined by the claims and their equivalents.

Claims

1. An intelligent gas pressure pipeline detection system, characterized in that: include: A gas pressure data acquisition module, used to acquire a time series of multiple gas pressures acquired by pressure sensors deployed at multiple locations of the gas pipeline; A gas pressure time series encoding module, used for performing time series encoding based on gas pressure on each of the multiple gas pressure time series to obtain multiple gas pressure time series associated feature vectors; A gas pressure time series feature context semantics enhancement module, used for inputting the plurality of gas pressure time series associated feature vectors into a context semantics enhancement module based on saliency-globality to obtain a plurality of enhanced gas pressure time series associated feature vectors; A gas pressure time series feature node information transmission fusion module is used to input the multiple enhanced gas pressure time series associated feature vectors into a node information transmission aggregation network that fuses node confidences to perform node information transmission aggregation based on position node confidences to obtain a gas pipeline pressure position transmission time series aggregation representation vector; A pipeline pressure anomaly detection module, used for inputting the gas pipeline pressure position transmission time series aggregation representation vector into a classifier-based gas pressure pipeline detection module to obtain a detection result, wherein the detection result is used to indicate whether there is an abnormality in the pipeline pressure; Wherein, when the gas pipeline pressure position transmission time series aggregation representation vector is passed through the classifier-based gas pressure pipeline detection module to obtain a detection result, the gas pipeline pressure position transmission time series aggregation representation vector is optimized. Specifically, the optimization process includes the following steps: Clustering all eigenvalues ​​of the gas pipeline pressure position transmission time series aggregate representation vector based on the distance between eigenvalues, and determining the ratio of the number of cluster features to the number of features of the gas pipeline pressure position transmission time series aggregate representation vector; Arranging the cluster features into a gas pipeline pressure position transmission time series aggregation cluster vector, and dividing the second norm of the gas pipeline pressure position transmission time series aggregation cluster vector by the second norm of the gas pipeline pressure position transmission time series aggregation representation vector to obtain a first gas pipeline pressure position transmission time series aggregation cluster related weight value; Divide a first power value of a norm of the gas pipeline pressure position transmission time series aggregation clustering vector with the ratio value as an exponent by a second power value of a norm of the gas pipeline pressure position transmission time series aggregation representation vector with the ratio value as an exponent to obtain a second gas pipeline pressure position transmission time series aggregation clustering related weight value; For each eigenvalue of the gas pipeline pressure position transmission time series aggregation clustering vector, multiply it by the inverse of the difference between the first gas pipeline pressure position transmission time series aggregation clustering related weight value and the second gas pipeline pressure position transmission time series aggregation clustering related weight value to obtain the optimized eigenvalue of the gas pipeline pressure position transmission time series aggregation clustering vector; For each eigenvalue outside the cluster in the gas pipeline pressure position transmission time series aggregation representation vector, multiply it by the inverse of the sum of the weight value related to the first gas pipeline pressure position transmission time series aggregation cluster and the weight value related to the second gas pipeline pressure position transmission time series aggregation cluster to obtain the optimized out-of-class eigenvalue of the gas pipeline pressure position transmission time series aggregation representation vector; The optimized eigenvalues ​​of the gas pipeline pressure position transmission time series aggregation clustering vector and the optimized out-of-class eigenvalues ​​of the gas pipeline pressure position transmission time series aggregation representation vector are combined to form an optimized gas pipeline pressure position transmission time series aggregation representation vector.

2. The intelligent gas pressure pipeline detection system according to claim 1 is characterized in that: The gas pressure time series encoding module is used to: respectively pass each gas pressure time series in the multiple gas pressure time series through a gas pressure sequence encoder based on a Bi-LSTM model to obtain the multiple gas pressure time series associated feature vectors.

3. The intelligent gas pressure pipeline detection system according to claim 2 is characterized in that: The gas pressure time series feature context semantic enhancement module includes: a prominent representation unit, used to respectively calculate the mean and maximum value of each gas pressure time series association feature vector in the multiple gas pressure time series association feature vectors to obtain a gas pressure time series association semantic global representation vector composed of multiple vector mean values ​​and a gas pressure time series association semantic prominent representation vector composed of multiple vector maximum values; A first feature capture and expression unit is used to capture and express the gas pressure time series associated semantic global representation vector to obtain a gas pressure time series global semantic activation representation feature vector; A second feature capture and expression unit is used to capture and express the gas pressure time series associated semantic salient representation vector to obtain a gas pressure time series salient semantic activation representation feature vector; a nonlinear activation unit, configured to fuse the gas pressure time series global semantic activation representation feature vector and the gas pressure time series salient semantic activation representation feature vector to obtain a gas pressure time series global-salient semantic association activation feature vector, and then perform nonlinear activation processing on the gas pressure time series global-salient semantic association activation feature vector to obtain a gas pressure time series global-salient semantic association weight feature vector; The enhancement unit is used to use the gas pressure time series global-prominent semantic association weight feature vector as a weight, perform positional point multiplication on the multiple gas pressure time series association feature vectors, and add the multiple gas pressure time series association feature vectors to obtain the multiple enhanced gas pressure time series association feature vectors.

4. The intelligent gas pressure pipeline detection system according to claim 3 is characterized in that: The first feature capture expression unit is used to: After one-dimensional convolution coding is performed on the gas pressure temporal association semantic global representation vector, The activation function is used to process the global representation vector of the activation expression of the gas pressure temporal association semantics; The activation expression gas pressure time series associated semantic global representation vector is again subjected to one-dimensional convolution encoding and then matrix multiplied with the first trainable linear transformation weight matrix to obtain the gas pressure time series global semantic activation representation feature vector.

5. The intelligent gas pressure pipeline detection system according to claim 4 is characterized in that: The second feature capture expression unit is used to: After one-dimensional convolution encoding is performed on the gas pressure time series associated semantic salient representation vector, The activation function is used to process the gas pressure so as to obtain a semantically prominent representation vector for activation expression of the gas pressure temporal correlation; The activation expression gas pressure time series associated semantic salient representation vector is again subjected to one-dimensional convolution encoding and then matrix multiplied with a second trainable linear transformation weight matrix to obtain the gas pressure time series salient semantic activation representation feature vector.

6. The intelligent gas pressure pipeline detection system according to claim 5 is characterized in that: The nonlinear activation unit is used to: The gas pressure time series global semantic activation representation feature vector and the gas pressure time series prominent semantic activation representation feature vector are merged to obtain the gas pressure time series global-prominent semantic association activation feature vector; Using the hyperbolic tangent function and The function performs activation processing on the gas pressure time series global-salient semantic association activation feature vector to obtain the gas pressure time series global-salient feature hyperbolic tangent activation semantic representation vector and the gas pressure time series global-salient feature S-type activation semantic representation vector; The global-salient feature hyperbolic tangent activation semantic representation vector of the gas pressure time series and the global-salient feature S-type activation semantic representation vector of the gas pressure time series are multiplied point by point to obtain the global-salient semantic association weight feature vector of the gas pressure time series.

7. The intelligent gas pressure pipeline detection system according to claim 6 is characterized in that: The gas pressure time series feature node information transmission and fusion module includes: a labeling unit, configured to label first k-1 enhanced gas pressure time series associated feature vectors among the plurality of enhanced gas pressure time series associated feature vectors with serial numbers to obtain a sequence of gas pressure time series associated semantic labeling serial numbers; a mean and variance calculation unit, used for calculating the energy function values ​​of the first k-1 enhanced gas pressure time series associated characteristic vectors to obtain k-1 energy function values, wherein the energy function value is related to the mean and variance of each enhanced gas pressure time series associated characteristic vector and the marked sequence number of each enhanced gas pressure time series associated characteristic vector in the sequence; A normalization unit, used for normalizing the k-1 energy function values ​​to obtain k-1 node confidences; an aggregation unit, configured to calculate the weighted sum of the first k-1 enhanced gas pressure time series association feature vectors among the multiple enhanced gas pressure time series association feature vectors by taking the k-1 node confidences as weights, so as to aggregate the first k-1 enhanced gas pressure time series association feature vectors among the multiple enhanced gas pressure time series association feature vectors to obtain an aggregated representation vector of the gas pressure time series association semantic rear node set; A weighted sum unit is used to calculate the weighted sum between the gas pressure time series association semantic rear node set aggregation representation vector and the kth enhanced gas pressure time series association feature vector among the multiple enhanced gas pressure time series association feature vectors to obtain the gas pipeline pressure position transmission time series aggregation representation vector.

8. The intelligent gas pressure pipeline detection system according to claim 7 is characterized in that: The mean and variance calculation unit is used to: Calculating the mean and variance of the predetermined enhanced gas pressure time series association feature vectors among the first k-1 enhanced gas pressure time series association feature vectors to obtain the gas pressure time series association semantic mean and the gas pressure time series association semantic variance; Determining the gas pressure time series association semantic annotation sequence number of the predetermined enhanced gas pressure time series association feature vector to obtain a predetermined gas pressure time series association semantic annotation sequence number; Calculate the sum of the gas pressure time series associated semantics mean and the gas pressure time series associated semantics variance, and then multiply the sum by a constant 4 to obtain a first gas pressure time series associated semantics modulation factor; Calculate the product of the gas pressure time series associated semantic mean and the gas pressure time series associated semantic variance and the constant 2 respectively to obtain the double gas pressure time series associated semantic variance and the double gas pressure time series associated semantic mean; After calculating the square of the difference between the gas pressure time series associated semantics annotation serial number and the gas pressure time series associated semantics mean, the square is added to the double gas pressure time series associated semantics variance and the double gas pressure time series associated semantics mean to obtain a second gas pressure time series associated semantics modulation factor; The division between the first gas pressure time series association semantic modulation factor and the second gas pressure time series association semantic modulation factor is calculated to obtain the energy function value corresponding to the predetermined enhanced gas pressure time series association feature vector.

9. The intelligent gas pressure pipeline detection system according to claim 8 is characterized in that: It also includes a training module for training the gas pressure sequence encoder based on the Bi-LSTM model, the contextual semantic reinforcement module based on significance-globality, the node information transmission aggregation network integrating node confidence, and the gas pressure pipeline detection module based on the classifier; Wherein, the training module includes: A training data acquisition unit, used to acquire training data, wherein the training data includes a time series of multiple training gas pressures acquired by pressure sensors deployed at multiple locations of the gas pipeline and actual detection results; A training gas pressure time series encoding unit, used for respectively passing each training gas pressure time series of the multiple training gas pressure time series through the gas pressure sequence encoder based on the Bi-LSTM model to obtain multiple training gas pressure time series associated feature vectors; A training gas pressure time series feature context semantics enhancement unit, used for inputting the plurality of training gas pressure time series associated feature vectors into the saliency-global based context semantics enhancement module to obtain a plurality of training enhanced gas pressure time series associated feature vectors; A training gas pressure time series feature node information transmission fusion unit is used to input the multiple training enhanced gas pressure time series associated feature vectors into the node information transmission aggregation network of the fusion node confidence to perform node information transmission aggregation based on the position node confidence to obtain a training gas pipeline pressure position transmission time series aggregation representation vector; A training pipeline pressure anomaly detection unit, used for inputting the training gas pipeline pressure position transfer time series aggregation representation vector into the classifier-based gas pressure pipeline detection module to obtain a training detection result; A classification loss function value calculation unit, used to calculate the cross entropy loss function value between the training detection result and the real detection result to obtain a classification loss function value; A loss training unit is used to train the gas pressure sequence encoder based on the Bi-LSTM model, the contextual semantic reinforcement module based on significance-globality, the node information transmission aggregation network fused with node confidence, and the gas pressure pipeline detection module based on the classifier based on the classification loss function value.

10. An intelligent gas pressure pipeline detection method, characterized in that: include: Acquire a plurality of time series of gas pressures collected by pressure sensors deployed at a plurality of locations of a gas pipeline; Performing time series coding based on gas pressure on each of the multiple time series of gas pressures to obtain multiple gas pressure time series associated feature vectors; Inputting the plurality of gas pressure time series associated feature vectors into a contextual semantic enhancement module based on saliency-globality to obtain a plurality of enhanced gas pressure time series associated feature vectors; Performing node information transmission aggregation based on position node confidence on the multiple enhanced gas pressure time series associated feature vectors to obtain a gas pipeline pressure position transmission time series aggregation representation vector; Inputting the gas pipeline pressure position transmission time series aggregation representation vector into a classifier-based gas pressure pipeline detection module to obtain a detection result, wherein the detection result is used to indicate whether there is an abnormality in the pipeline pressure; Wherein, when the gas pipeline pressure position transmission time series aggregation representation vector is passed through the classifier-based gas pressure pipeline detection module to obtain a detection result, the gas pipeline pressure position transmission time series aggregation representation vector is optimized. Specifically, the optimization process includes the following steps: Clustering all eigenvalues ​​of the gas pipeline pressure position transmission time series aggregate representation vector based on the distance between eigenvalues, and determining the ratio of the number of cluster features to the number of features of the gas pipeline pressure position transmission time series aggregate representation vector; Arranging the cluster features into a gas pipeline pressure position transmission time series aggregation cluster vector, and dividing the second norm of the gas pipeline pressure position transmission time series aggregation cluster vector by the second norm of the gas pipeline pressure position transmission time series aggregation representation vector to obtain a first gas pipeline pressure position transmission time series aggregation cluster related weight value; Divide a first power value of a norm of the gas pipeline pressure position transmission time series aggregation clustering vector with the ratio value as an exponent by a second power value of a norm of the gas pipeline pressure position transmission time series aggregation representation vector with the ratio value as an exponent to obtain a second gas pipeline pressure position transmission time series aggregation clustering related weight value; For each eigenvalue of the gas pipeline pressure position transmission time series aggregation clustering vector, multiply it by the inverse of the difference between the first gas pipeline pressure position transmission time series aggregation clustering related weight value and the second gas pipeline pressure position transmission time series aggregation clustering related weight value to obtain the optimized eigenvalue of the gas pipeline pressure position transmission time series aggregation clustering vector; For each eigenvalue outside the cluster in the gas pipeline pressure position transmission time series aggregation representation vector, multiply it by the inverse of the sum of the weight value related to the first gas pipeline pressure position transmission time series aggregation cluster and the weight value related to the second gas pipeline pressure position transmission time series aggregation cluster to obtain the optimized out-of-class eigenvalue of the gas pipeline pressure position transmission time series aggregation representation vector; The optimized eigenvalues ​​of the gas pipeline pressure position transmission time series aggregation clustering vector and the optimized out-of-class eigenvalues ​​of the gas pipeline pressure position transmission time series aggregation representation vector are combined to form an optimized gas pipeline pressure position transmission time series aggregation representation vector.

Images