Gas Engineering Construction Safety Management System and Method
By deploying pressure sensors at key nodes of gas pipelines and using artificial intelligence and deep learning analysis technology, combining vibration waveform semantics to identify pipeline abnormalities, the problem that traditional systems cannot effectively detect safety hazards is solved, and high-precision safety management and rapid response.
Patent Information
- Application Number
- CN202411133849.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-08-19
AI Technical Summary
The traditional gas engineering construction safety management system relies on manual inspection and single data monitoring, and cannot effectively detect pipeline safety hazards, and the response speed is slow, increasing the risk of accidents.
Pressure sensors deployed at key nodes of the gas pipeline monitor pressure data in real time, and use artificial intelligence and deep learning algorithms to perform timing feature extraction and collaborative correlation analysis, and interact and fusion with vibration waveform semantics to identify pipeline anomalies.
It has achieved comprehensive monitoring and safety management of gas engineering construction pipelines, improved the accuracy of abnormal identification and detection, timely discovered safety hazards and made maintenance and decision-making adjustments, and improved construction safety.
Smart Images

Figure CN118656763B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of intelligent management, and more specifically, to a gas engineering construction safety management system and method. Background Art
[0002] Gas engineering is an important part of urban infrastructure, providing clean and efficient energy for residents and commercial users. The key to gas engineering construction safety management is to take measures to ensure the safety of personnel, equipment and environment during the design, construction and operation of gas pipeline systems.
[0003] The gas pipeline system is an important energy transmission channel in cities and industrial areas. Safety management is crucial to preventing accidents and ensuring public safety. With the acceleration of urbanization, the scale of the gas pipeline network continues to expand, and the laying and maintenance of pipelines have become more complicated. The safety management of gas pipelines is directly related to public safety and environmental safety. Any leakage or failure may cause serious safety accidents, resulting in casualties and property losses. Therefore, it is crucial to detect safety hazards in gas pipelines during construction, so as to detect abnormalities in time and make corresponding construction strategy adjustments.
[0004] However, the traditional gas engineering construction pipeline safety management system mainly relies on manual inspections, which not only has problems such as high human resource costs and untimely monitoring, but also cannot effectively detect the safety hazards of gas pipelines, and the response speed to emergencies is slow. In addition, there are some existing methods that rely on data monitoring to achieve automated pipeline safety monitoring and management, but the current data monitoring methods usually rely on the monitoring of a single pipeline parameter and threshold comparison to determine whether there is an abnormality. It lacks multi-dimensional data fusion and comprehensive intelligent analysis, and it is difficult to comprehensively evaluate the pipeline status. In addition, this method has a slow speed of identifying and responding to pipeline anomalies, which can easily lead to timely detection of problems after an accident occurs, increasing the risk of accidents.
[0005] Therefore, an optimized gas engineering construction safety management system is desired. Summary of the invention
[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a gas engineering construction safety management system and method, which monitors and collects pressure data at different key nodes in real time through pressure sensors deployed at multiple key nodes of the gas engineering pipeline, and introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the back end to extract time series features and conduct collaborative correlation analysis on the pressure data at different key nodes, and at the same time, the pressure time series global information aggregation features of these different pipeline key nodes are interactively integrated with the vibration waveform semantics of the pipeline to learn the dynamic interaction between the two, which helps to discover hidden pipeline anomalies and safety hazards, thereby further improving the accuracy of pipeline abnormality identification and detection. In this way, it is possible to achieve all-round monitoring and safety management of gas engineering construction pipelines, so as to timely discover abnormal conditions and safety hazards of construction pipelines, so as to carry out corresponding maintenance and gas engineering construction decision adjustments, and achieve more intelligent gas engineering construction safety management.
[0007] According to one aspect of the present application, a gas engineering construction safety management system is provided, which includes: a pipeline key node pressure data acquisition module, which is used to obtain the time series of pressure values collected by pressure sensors deployed at multiple key nodes of the gas engineering pipeline to obtain the time series of multiple key node pipeline pressure values; a pipeline vibration signal acquisition module, which is used to obtain the vibration signal of the gas engineering pipeline; a pipeline pressure time series feature extraction module, which is used to extract the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values based on pipeline pressure to obtain a sequence of pipeline pressure time series feature vectors; a pipeline pressure global information aggregation module, which is used to aggregate the pipeline pressure time series feature vectors; The sequence is passed through a node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector; a pipeline vibration waveform semantic feature extraction module is used to extract semantic features of the vibration signal of the gas engineering pipeline based on the vibration signal waveform to obtain a gas engineering pipeline vibration waveform semantic feature vector; a pipeline state multimodal semantic representation module is used to combine the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector through a pipeline multimodal semantic dynamic interaction fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector; a gas engineering pipeline anomaly detection module is used to determine whether there is an anomaly in the gas engineering pipeline based on the pipeline state semantic multimodal representation vector.
[0008] According to another aspect of the present application, a gas engineering construction safety management method is provided, which includes: obtaining a time series of pressure values collected by pressure sensors deployed at multiple key nodes of a gas engineering pipeline to obtain a time series of multiple key node pipeline pressure values; obtaining a vibration signal of the gas engineering pipeline; performing pipeline pressure-based time series feature extraction on the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values to obtain a sequence of pipeline pressure time series feature vectors; passing the sequence of pipeline pressure time series feature vectors through a node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector; performing semantic feature extraction based on the vibration signal waveform on the vibration signal of the gas engineering pipeline to obtain a gas engineering pipeline vibration waveform semantic feature vector; passing the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector through a pipeline multimodal semantic dynamic interaction fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector; based on the pipeline state semantic multimodal representation vector, determining whether there is an abnormality in the gas engineering pipeline.
[0009] Compared with the prior art, the present application provides a gas engineering construction safety management system and method, which monitors and collects pressure data at different key nodes in real time through pressure sensors deployed at multiple key nodes of the gas engineering pipeline, and introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the back end to extract time series features and conduct collaborative correlation analysis on the pressure data at different key nodes. At the same time, the global information aggregation features of the pressure time series of these different pipeline key nodes are interactively integrated with the vibration waveform semantics of the pipeline to learn the dynamic interaction between the two, which helps to discover hidden pipeline anomalies and safety hazards, thereby further improving the accuracy of pipeline abnormality identification and detection. In this way, it is possible to achieve all-round monitoring and safety management of gas engineering construction pipelines, so as to timely discover abnormal conditions and safety hazards of construction pipelines, so as to carry out corresponding maintenance and gas engineering construction decision adjustments, and realize more intelligent gas engineering construction safety management. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.
[0011] Figure 1 It is a block diagram of a gas engineering construction safety management system according to an embodiment of the present application.
[0012] Figure 2 This is a system architecture diagram of a gas engineering construction safety management system according to an embodiment of the present application.
[0013] Figure 3 It is a block diagram of the pipeline pressure global information aggregation module in the gas engineering construction safety management system according to an embodiment of the present application.
[0014] Figure 4 It is a flow chart of a gas engineering construction safety management method according to an embodiment of the present application. DETAILED DESCRIPTION
[0015] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.
[0016] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.
[0017] 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.
[0018] 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. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.
[0019] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.
[0020] With the development of technologies such as the Internet of Things, big data, and artificial intelligence, intelligent monitoring has become a new trend in gas engineering safety management. Through intelligent means, the operating status of gas pipelines can be monitored in real time, and potential safety issues can be discovered and dealt with in a timely manner. In particular, sensor technology is the basis for realizing intelligent monitoring. Pressure sensors and vibration sensors can collect pipeline operating data in real time and provide raw data for subsequent data analysis and safety assessment.
[0021] In the technical solution of the present application, a gas engineering construction safety management system is proposed. Figure 1 It is a block diagram of a gas engineering construction safety management system according to an embodiment of the present application. Figure 2 FIG. 1 is a system architecture diagram of a gas engineering construction safety management system according to an embodiment of the present application. Figure 1 and Figure 2 As shown, according to the embodiment of the present application, the gas engineering construction safety management system 300 includes: a pipeline key node pressure data acquisition module 310, which is used to obtain the time series of pressure values collected by pressure sensors deployed at multiple key nodes of the gas engineering pipeline to obtain the time series of multiple key node pipeline pressure values; a pipeline vibration signal acquisition module 320, which is used to obtain the vibration signal of the gas engineering pipeline; a pipeline pressure time series feature extraction module 330, which is used to extract the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values based on pipeline pressure to obtain a sequence of pipeline pressure time series feature vectors; a pipeline pressure global information aggregation module 340, which is used to aggregate the pipeline pressure time series feature vectors The sequence is passed through a node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector; a pipeline vibration waveform semantic feature extraction module 350 is used to extract semantic features of the vibration signal of the gas engineering pipeline based on the vibration signal waveform to obtain a gas engineering pipeline vibration waveform semantic feature vector; a pipeline state multimodal semantic representation module 360 is used to combine the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector through a pipeline multimodal semantic dynamic interaction fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector; a gas engineering pipeline anomaly detection module 370 is used to determine whether there is an anomaly in the gas engineering pipeline based on the pipeline state semantic multimodal representation vector.
[0022] In particular, the pipeline key node pressure data acquisition module 310 and the pipeline vibration signal acquisition module 320 are used to obtain the time series of pressure values collected by the pressure sensors deployed at multiple key nodes of the gas engineering pipeline to obtain the time series of multiple key node pipeline pressure values; and obtain the vibration signal of the gas engineering pipeline. It should be understood that by deploying pressure sensors at the key nodes of the pipeline, the pressure level of the pipeline can be monitored. Specifically, by analyzing the time series characteristics of the pressure data at different key nodes and the synergy between the nodes, abnormal pressure conditions can be identified, which may indicate pipeline damage or blockage, or leakage problems. The vibration sensor can detect tiny vibrations of the pipeline, which may be caused by pressure fluctuations inside the pipeline or external factors. Analyzing the vibration data to identify abnormal vibration patterns that may cause pipeline damage helps to determine whether there are leakage problems or other safety hazards in the pipeline, so as to identify and respond to potential safety risks in a timely manner.
[0023] In particular, the pipeline pressure time series feature extraction module 330 is used to extract the time series features based on pipeline pressure for each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values to obtain a sequence of pipeline pressure time series feature vectors. In a specific example of the present application, the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values is firstly data regularized according to the time dimension to obtain a sequence of pipeline pressure time series input vectors, and then each pipeline pressure time series input vector in the sequence of pipeline pressure time series input vectors is respectively passed through the pipeline pressure time series feature extractor based on the 1D_CNN model to obtain the sequence of pipeline pressure time series feature vectors. Considering that the pipeline pressure values at different key nodes of the gas engineering construction pipeline will change over time in the time dimension, it has the dynamic characteristics and change information of the time series. Therefore, in order to be able to analyze and characterize the pressure time series characteristics and dynamic change information at different key nodes of the pipeline respectively, the pressure interaction and aggregation information between these different pipeline key nodes can be better extracted. Based on this, in the technical solution of the present application, it is necessary to regularize the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values according to the time dimension to obtain a sequence of pipeline pressure time series input vectors. Then, each pipeline pressure time series input vector in the sequence of pipeline pressure time series input vectors is subjected to feature mining in a pipeline pressure time series feature extractor based on the 1D_CNN model to extract the dynamic implicit feature information of the pressure time series at each key node of the gas pipeline, thereby obtaining a sequence of pipeline pressure time series feature vectors. It is worth mentioning that 1D CNN (one-dimensional convolutional neural network) is a deep learning model for processing one-dimensional data sequences, and is commonly used in fields such as time series data, audio processing, and natural language processing.
[0024] In particular, the pipeline pressure global information aggregation module 340 is used to pass the sequence of the pipeline pressure time series feature vectors through the node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector. It should be understood that since each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector contains the pipeline pressure time series characteristics of the gas pipeline at each key node, and the pipeline pressure time series characteristics at these different key nodes have a correlation relationship and mutual influence based on the pipeline global domain. Based on this, in order to be able to analyze the time series changes of pipeline pressure based on the pipeline global domain, so as to better identify pipeline pressure anomalies and leakage conditions and pipeline safety hazards, in the technical solution of the present application, the sequence of the pipeline pressure time series feature vectors is further passed through the node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector. Through the processing of the node energy attention aggregation module based on sequence annotation, it is possible to automatically learn and capture the inter-correlation and synergistic interaction between the pressure time series features at different key nodes of the gas pipeline, identify which key nodes are more relevant to the subsequent pipeline anomaly detection task, and give different importance weights to the pressure time series features at different key nodes. In this way, it is possible to pay attention to the energy surge of different key nodes in the process of aggregating the pressure time series features at different key nodes of these gas pipelines, and apply higher weights to the more critical key nodes related to the identification of pipeline state changes and anomaly detection tasks, so as to pay attention to those nodes with higher energy or importance in the sequence, and provide a basis for the subsequent identification of abnormal conditions and safety hazards in gas engineering pipelines. In addition, the same processing method can also help improve the generalization ability of the model, so that it can adapt to the dynamic changes of pipeline pressure under different conditions and identify gas engineering pipeline anomalies under different conditions. In particular, in a specific example of the present application, such as Figure 3 As shown, the pipeline pressure global information aggregation module 340 includes: a serial number labeling unit 341, which is used to label each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector with a serial number to obtain a sequence of pipeline pressure key node time series semantic labeling serial numbers; an energy function value calculation unit 342, which is used to calculate the energy function value of each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector to obtain a sequence of energy function values, wherein the energy function value is related to the mean and variance of each pipeline pressure time series feature vector and the pipeline pressure key node time series semantic labeling serial number; a weighted enhancement unit 343, which is used to calculate the weighted sum between each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector using the inverse of each energy function value in the sequence of the energy function value as a weight factor to obtain the enhanced pipeline pressure global information aggregation feature vector.
[0025] Specifically, the sequence number labeling unit 341 is used to label each pipeline pressure time series feature vector in the sequence of pipeline pressure time series feature vectors with a sequence number to obtain a sequence of pipeline pressure key node time series semantic labeling sequence numbers. It should be understood that labeling each pipeline pressure time series feature vector in the sequence of pipeline pressure time series feature vectors with a sequence number can help identify and mark key nodes in pipeline pressure data, thereby obtaining a sequence of pipeline pressure key node time series semantic labeling sequence numbers.
[0026] Specifically, the energy function value calculation unit 342 is used to calculate the energy function value of each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector to obtain a sequence of energy function values, wherein the energy function value is related to the mean and variance of each pipeline pressure time series feature vector and the pipeline pressure key node time series semantic annotation sequence number. In a specific example, the mean and variance of the predetermined pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector are first calculated to obtain the pipeline pressure key node time series semantic mean and the pipeline pressure key node time series semantic variance; then, the pipeline pressure key node time series semantic annotation sequence number of the predetermined pipeline pressure time series feature vector is determined to obtain the predetermined pipeline pressure key node time series semantic annotation sequence number; the sum of the pipeline pressure key node time series semantic variance and the pipeline pressure key node time series semantic mean is calculated and then multiplied by a constant 4 to obtain a first pipeline pressure key node time series semantic modulation factor; then, the pipeline pressure key node time series semantic variance and the pipeline pressure key node time series semantic mean are calculated respectively. The product of the sequence semantic mean and the constant 2 is used to obtain the double pipeline pressure key node sequence semantic variance and the double pipeline pressure key node sequence semantic mean; after further calculating the square of the difference between the pipeline pressure key node sequence semantic annotation serial number and the pipeline pressure key node sequence semantic mean, it is added with the double pipeline pressure key node sequence semantic variance and the double pipeline pressure key node sequence semantic mean to obtain the second pipeline pressure key node sequence semantic modulation factor; finally, the division between the first pipeline pressure key node sequence semantic modulation factor and the second pipeline pressure key node sequence semantic modulation factor is calculated to obtain the energy function value corresponding to the predetermined pipeline pressure sequence feature vector.
[0027] Specifically, the weighted enhancement unit 343 is used to calculate the weighted sum of each pipeline pressure time series feature vector in the sequence of pipeline pressure time series feature vectors using the reciprocal of each energy function value in the sequence of energy function values as a weight factor to obtain the enhanced pipeline pressure global information aggregation feature vector. In a specific example, the reciprocal of each energy function value in the sequence of energy function values is first calculated to obtain a sequence of energy function reciprocal representation values; then, using The function activates the sequence of the inverse representation values of the energy function to obtain the sequence of the inverse representation values of the activated energy function; then calculates the product between each group of corresponding inverse representation values of the activated energy function and the sequence of the pipeline pressure time series feature vectors, and then adds each enhanced pipeline pressure time series feature vector in the obtained sequence of enhanced pipeline pressure time series feature vectors by position to obtain the enhanced pipeline pressure global information aggregation feature vector.
[0028] In summary, in the above embodiment, the sequence of the pipeline pressure time series feature vectors is passed through the node energy attention aggregation module based on sequence annotation to obtain the enhanced pipeline pressure global information aggregation feature vector, including: the sequence of the pipeline pressure time series feature vectors is processed through the node energy attention aggregation module based on sequence annotation with the following node energy enhancement aggregation formula to obtain the enhanced pipeline pressure global information aggregation feature vector; wherein the node energy enhancement aggregation formula is: ;in, is the first in the sequence of pipeline pressure time series feature vectors pipeline pressure time series feature vector, For each pipeline pressure time series feature vector Position feature values, is the length of each pipeline pressure time series feature vector, is the mean of each pipeline pressure time series feature vector, is the variance of each pipeline pressure time series feature vector, is the regularization hyperparameter, is a labeling serial number for each pipeline pressure time series feature vector in the set, is the energy function value of each pipeline pressure time series characteristic vector, for function, is the Hadamard product, is the number of vectors in the sequence of pipeline pressure time series feature vectors, The global information aggregation feature vector of the enhanced pipeline pressure is obtained.
[0029] It is worth mentioning that in other specific examples of the present application, the sequence of the pipeline pressure time series feature vectors can also be passed through a node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector in other ways, for example: input the sequence of the pipeline pressure time series feature vectors; perform sequence annotation on the sequence of pipeline pressure time series feature vectors, identify and mark key nodes; calculate the energy of each node; based on the energy value of the node, calculate the attention weight of each node to determine the contribution of each node to the global information aggregation; use the node's attention weight to perform weighted aggregation on the feature vector of each node to obtain the enhanced pipeline pressure global information aggregation feature vector.
[0030] In particular, the pipeline vibration waveform semantic feature extraction module 350 is used to extract semantic features of the vibration signal of the gas engineering pipeline based on the vibration signal waveform to obtain a semantic feature vector of the gas engineering pipeline vibration waveform. In a specific example of the present application, the vibration signal of the gas engineering pipeline is passed through a vibration signal waveform semantic feature extractor based on a depthwise separable convolution layer to obtain a semantic feature vector of the gas engineering pipeline vibration waveform. It should be understood that for the vibration signal of the gas engineering pipeline, its expression in the time domain is a waveform graph. Therefore, the vibration signal of the gas engineering pipeline is passed through a vibration signal waveform semantic feature extractor based on a depthwise separable convolution layer to obtain a semantic feature vector of the gas engineering pipeline vibration waveform. Specifically, since the standard convolution layer contains 3x3 convolution kernels, batch normalization and nonlinear activation function (ReLU), in the technical solution of the present application, the depth separable convolution layer contains 3x3 depth convolution kernels, batch normalization, 1x1 point-by-point convolution kernels and nonlinear activation function (ReLU6), where ReLU6 is obtained by limiting the maximum output of the ReLU nonlinear activation function to 6, and its main function is to allow the network to learn sparse features earlier and prevent numerical explosion. It should be understood that compared with the traditional convolution layer, the depth separable convolution layer has better performance in feature extraction, which helps to accurately capture important information in the vibration signal. Specifically, the depth separable convolution layer reduces the number of parameters and the amount of calculation of the model by separating the spatial convolution and channel mixing steps, making the model more lightweight while maintaining performance, which is very important for processing a large amount of vibration signal data and can improve processing efficiency. In addition, the depth separable convolution layer of this structure can capture the semantic features of vibration waveforms at different scales, which helps to identify subtle changes from the vibration signal, which may be early signs of pipeline abnormalities. This enables the deep separable convolutional layer to adapt to different vibration signal characteristics, whether it is high-frequency transient changes or low-frequency continuous trends, and can effectively capture the semantic features of the vibration signal waveform and better understand the information contained in the vibration signal.
[0031] In particular, the pipeline state multimodal semantic representation module 360 is used to obtain a pipeline state semantic multimodal representation vector by combining the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector through a pipeline multimodal semantic dynamic interactive fusion module based on gated response. Since the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector respectively contain the pipeline pressure global information aggregation feature and the gas engineering pipeline vibration waveform semantic feature after key node feature enhancement, data of different modes may contain different aspects of pipeline state. For example, pressure data may reflect pipeline blockage or leakage problems, while vibration data may reveal physical damage or external impact of the pipeline. In addition, there is a mutual influence and synergistic correlation between the features of the two. Therefore, in order to improve the accuracy of pipeline state recognition and anomaly detection, in the technical solution of the present application, the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector are further combined through a pipeline multimodal semantic dynamic interactive fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector. It should be understood that some abnormal conditions and problems of the gas pipeline may not be obvious in a single data source, but can be effectively identified after combining with other data sources. Through the processing of the pipeline multimodal semantic dynamic interactive fusion module based on gated response, the dynamic interactive relationship and importance semantics between the global information aggregation characteristics of the gas pipeline pressure and its vibration waveform semantic characteristics can be learned and captured to extract the interaction and synergistic correlation information between the global characteristics of the gas pipeline pressure and the vibration waveform semantics, and identify the importance of these two features for the task of identifying gas pipeline abnormal conditions in specific scenarios, so as to give different gas pipeline features different weights in the process of information interactive fusion to more comprehensively characterize the state of the gas pipeline, so as to better utilize the comprehensive semantics and dynamic interactive features of the two to perform subsequent gas engineering pipeline abnormality identification and detection tasks.
[0032] In an embodiment of the present application, the process of obtaining a pipeline state semantic multimodal representation vector by passing the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector through a pipeline multimodal semantic dynamic interactive fusion module based on gated response includes: inputting the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector into a feature joint module for cascade processing to obtain a pipeline pressure global information-vibration waveform semantic timing joint feature vector; calculating the matrix multiplication of the pipeline pressure global information-vibration waveform semantic timing joint feature vector and the parameter matrix, and then adding the obtained feature vector to the bias vector by position to obtain a linear transformation pipeline pressure global information-vibration waveform semantic timing joint feature vector; using The function activates the linear transformation pipeline pressure global information-vibration waveform semantic temporal joint feature vector to obtain the pipeline pressure global information-vibration waveform semantic temporal dynamic information fusion response gating value; calculates the positional product between the enhanced pipeline pressure global information aggregation feature vector and the pipeline pressure global information-vibration waveform semantic temporal dynamic information fusion response gating value to obtain the weight modulated enhanced pipeline pressure global information aggregation feature vector; calculates one minus the pipeline pressure global information-vibration waveform semantic temporal dynamic information fusion response gating value, and then multiplies the obtained weight value with the gas engineering pipeline vibration waveform semantic feature vector by position to obtain the weight modulated gas engineering pipeline vibration waveform semantic feature vector; the weight modulated enhanced pipeline pressure global information aggregation feature vector and the weight modulated gas engineering pipeline vibration waveform semantic feature vector are added by position point to obtain the pipeline state semantic multimodal representation vector.
[0033] In summary, in the above embodiment, the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector are passed through a pipeline multimodal semantic dynamic interactive fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector, including: the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector are passed through the pipeline multimodal semantic dynamic interactive fusion module based on gated response to obtain the pipeline state semantic multimodal representation vector according to the following dynamic interactive fusion formula; wherein the dynamic interactive fusion formula is: :in, and They are respectively the global information aggregation feature vector of the enhanced pipeline pressure and the semantic feature vector of the gas engineering pipeline vibration waveform, represents a vector concatenation operation, is the parameter matrix, is the bias vector, yes function, It is the gating value of the pipeline pressure global information-vibration waveform semantic time series dynamic information fusion response. is the pipeline state semantic multimodal representation vector.
[0034] In particular, the gas engineering pipeline anomaly detection module 370 is used to determine whether there is an anomaly in the gas engineering pipeline based on the pipeline state semantic multimodal representation vector. In a specific example of the present application, the pipeline state semantic multimodal representation vector is passed through a classifier-based pipeline anomaly identifier to obtain a recognition result, and the recognition result is used to indicate whether there is an anomaly in the gas engineering pipeline. That is, the multimodal semantic fusion representation information of the gas pipeline state is used for classification processing, so as to effectively identify and detect the abnormal conditions and safety hazards of the gas engineering pipeline. In this way, it is possible to achieve all-round monitoring and safety management of the gas engineering construction pipeline, so as to timely discover the abnormal conditions and safety hazards of the construction pipeline, so as to carry out corresponding maintenance and gas engineering construction decision adjustments, and realize more intelligent gas engineering construction safety management.
[0035] Here, the enhanced pipeline pressure full time domain information aggregation feature vector expresses the local time domain time series correlation feature of the pipeline pressure value based on the aggregated time series feature representation of the local time domain node energy under the global time domain, and the gas engineering pipeline vibration waveform semantic feature vector expresses the image semantic feature of the vibration signal of the gas engineering pipeline. Therefore, after the enhanced pipeline pressure full time domain information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector pass through the pipeline multimodal semantic dynamic interaction fusion module based on gated response, the obtained pipeline state semantic multimodal representation vector will also lack the class probability interpretability relative to the source time series-image semantic distribution representation due to the semantic dynamic interaction differences caused by different modal features, thereby affecting the accuracy of the classification results. Based on this, the present application considers optimizing the pipeline state semantic multimodal representation vector when passing the pipeline state semantic multimodal representation vector through a classifier-based pipeline anomaly identifier to obtain an identification result.
[0036] That is, in a preferred embodiment, the pipeline state semantic multimodal representation vector is passed through a classifier-based pipeline anomaly identifier to obtain an identification result, including the steps of: calculating the autocorrelation matrix of the pipeline state semantic multimodal representation vector, that is: ;in, represents the pipeline state semantic multimodal representation vector, represents the transpose of a vector, represents the transpose of a vector, Denotes the autocorrelation matrix.
[0037] The pipeline state semantic multimodal representation vector is query-multiplied with the self-association matrix to obtain an intermediate feature vector, namely: ;in, represents the pipeline state semantic multimodal representation vector, represents the autocorrelation matrix, represents matrix multiplication, Denotes the intermediate eigenvector.
[0038] Calculate the first Row vector and Inner product sum of row vectors distance, to obtain the first weight matrix and the second weight matrix , expressed as: ;in, and The autocorrelation matrix is represented by Row vector and row vector, represents the transpose of a vector, represents matrix multiplication, represents the autocorrelation matrix, represents the first weight matrix, represents the first weight matrix The eigenvalues of the positions, represents the second weight matrix, represents the second weight matrix The eigenvalues of the positions, Represents the calculation vector distance.
[0039] The first weight matrix and the second weight matrix After matrix multiplication, further matrix multiplication is performed with the intermediate feature vector to obtain the optimized pipeline state semantic multimodal representation vector, namely: ;in, represents the first weight matrix, represents the second weight matrix, represents the intermediate eigenvector, represents matrix multiplication, Represents the optimized pipeline state semantic multimodal representation vector.
[0040] Finally, the optimized pipeline state semantic multimodal representation vector is passed through a classifier-based pipeline anomaly identifier to obtain a recognition result.
[0041] That is, for the bitwise scattering association response space of the eigenvalues of the pipeline state semantic multimodal representation vector, the deep embedding structure in the high-order association space is obtained through the joint embedding space of the vector inner product embedding space and the vector independent embedding space, and the decomposable association response sub-dimension of the bitwise scattering association response space of the eigenvalues of the pipeline state semantic multimodal representation vector is represented as a set of spatial sub-dimensions to further perform the dependent space mapping of the decomposable association response sub-dimension of the deep embedding structure of the pipeline state semantic multimodal representation vector, so as to improve the interpretability of the solution space based on the association distribution of the class regression solution of the pipeline state semantic multimodal representation vector through the complex dimensional representation of the association response, thereby improving the accuracy of the recognition result obtained by the pipeline state semantic multimodal representation vector input based on the classifier pipeline anomaly identifier. In this way, the accuracy of the recognition and detection of abnormal conditions of the gas pipeline can be further improved, so as to timely discover the abnormal conditions and safety hazards of the construction pipeline, so as to carry out corresponding maintenance and gas engineering construction decision adjustments, and realize more intelligent gas engineering construction safety management.
[0042] As described above, the gas engineering construction safety management system 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a gas engineering construction safety management algorithm. In a possible implementation, the gas engineering construction safety management system 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the gas engineering construction safety management system 300 can be a software module in the operating system of the wireless terminal, or can be an application developed for the wireless terminal; of course, the gas engineering construction safety management system 300 can also be one of the many hardware modules of the wireless terminal.
[0043] Alternatively, in another example, the gas engineering construction safety management system 300 and the wireless terminal may also be separate devices, and the gas engineering construction safety management system 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.
[0044] Furthermore, a gas engineering construction safety management method is also provided.
[0045] Figure 4 FIG. 1 is a flow chart of a gas engineering construction safety management method according to an embodiment of the present application. Figure 4As shown, according to the embodiment of the present application, the gas engineering construction safety management method includes the following steps: S1, obtaining the time series of pressure values collected by pressure sensors deployed at multiple key nodes of the gas engineering pipeline to obtain the time series of multiple key node pipeline pressure values; S2, obtaining the vibration signal of the gas engineering pipeline; S3, respectively performing pipeline pressure-based time series feature extraction on the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values to obtain a sequence of pipeline pressure time series feature vectors; S4, passing the sequence of pipeline pressure time series feature vectors through a node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector; S5, performing semantic feature extraction based on the vibration signal waveform on the vibration signal of the gas engineering pipeline to obtain a gas engineering pipeline vibration waveform semantic feature vector; S6, passing the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector through a pipeline multimodal semantic dynamic interaction fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector; S7, based on the pipeline state semantic multimodal representation vector, determining whether the gas engineering pipeline has an abnormality.
[0046] In summary, the gas engineering construction safety management method according to the embodiment of the present application is explained, which uses pressure sensors deployed at multiple key nodes of the gas engineering pipeline to monitor and collect pressure data at different key nodes in real time, and introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the back end to extract time series features and conduct collaborative correlation analysis on the pressure data at different key nodes. At the same time, the global information aggregation features of the pressure time series of these different pipeline key nodes are interactively integrated with the vibration waveform semantics of the pipeline to learn the dynamic interaction between the two, which helps to discover hidden pipeline anomalies and safety hazards, thereby further improving the accuracy of pipeline abnormality identification and detection. In this way, it is possible to achieve all-round monitoring and safety management of gas engineering construction pipelines, so as to timely discover abnormal conditions and safety hazards of construction pipelines, so as to carry out corresponding maintenance and gas engineering construction decision adjustments, and achieve more intelligent gas engineering construction safety management.
[0047] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.
Claims
1. A gas engineering construction safety management system, characterized in that: include: The pipeline key node pressure data acquisition module is used to obtain the time series of pressure values collected by the pressure sensors deployed at multiple key nodes of the gas engineering pipeline to obtain the time series of multiple key node pipeline pressure values; the pipeline vibration signal acquisition module is used to obtain the vibration signal of the gas engineering pipeline; the pipeline pressure time series feature extraction module is used to extract the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values based on pipeline pressure to obtain a sequence of pipeline pressure time series feature vectors; the pipeline pressure global domain information aggregation module is used to aggregate the sequence of the pipeline pressure time series feature vectors through the node energy attention based on sequence annotation An aggregation module is used to obtain an enhanced pipeline pressure global information aggregation feature vector; a pipeline vibration waveform semantic feature extraction module is used to extract semantic features of the vibration signal of the gas engineering pipeline based on the vibration signal waveform to obtain a gas engineering pipeline vibration waveform semantic feature vector; a pipeline state multimodal semantic representation module is used to combine the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector through a pipeline multimodal semantic dynamic interaction fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector; a gas engineering pipeline anomaly detection module is used to determine whether there is an anomaly in the gas engineering pipeline based on the pipeline state semantic multimodal representation vector; Wherein, the pipeline pressure global information aggregation module includes: a serial number labeling unit, which is used to label each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector with a serial number to obtain a sequence of pipeline pressure key node time series semantic labeling serial numbers; an energy function value calculation unit, which is used to calculate the energy function value of each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector to obtain a sequence of energy function values, wherein the energy function value is related to the mean and variance of each pipeline pressure time series feature vector and the pipeline pressure key node time series semantic labeling serial number; a weighted enhancement unit, which is used to calculate the weighted sum between each pipeline pressure time series feature vector in the sequence of the pipeline pressure time series feature vector using the inverse of each energy function value in the sequence of the energy function value as a weight factor to obtain the enhanced pipeline pressure global information aggregation feature vector; Among them, the pipeline state multimodal semantic representation module includes: a pipeline pressure global information-vibration waveform semantic time series joint unit, which is used to input the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector into the feature joint module for cascade processing to obtain the pipeline pressure global information-vibration waveform semantic time series joint feature vector; a linear transformation unit, which is used to calculate the matrix multiplication of the pipeline pressure global information-vibration waveform semantic time series joint feature vector and the parameter matrix, and then add the obtained feature vector to the bias vector by position to obtain the linear transformation pipeline pressure global information-vibration waveform semantic time series joint feature vector; an activation unit, which is used to use The function activates the linear transformation pipeline pressure global information-vibration waveform semantic time series joint feature vector to obtain the pipeline pressure global information-vibration waveform semantic time series dynamic information fusion response gating value; the enhanced pipeline pressure global information weight modulation unit is used to calculate the positional product between the enhanced pipeline pressure global information aggregation feature vector and the pipeline pressure global information-vibration waveform semantic time series dynamic information fusion response gating value to obtain the weight modulated enhanced pipeline pressure global information aggregation feature vector; the gas engineering pipeline vibration waveform weight modulation unit is used to calculate one minus the pipeline pressure global information-vibration waveform semantic time series dynamic information fusion response gating value to obtain the weight value, and multiply the obtained weight value with the gas engineering pipeline vibration waveform semantic feature vector by position to obtain the weight modulated gas engineering pipeline vibration waveform semantic feature vector; the pipeline state semantic multimodal representation unit is used to perform positional multiplication of the weight modulated enhanced pipeline pressure global information aggregation feature vector and the weight modulated gas engineering pipeline vibration waveform semantic feature vector to obtain the pipeline state semantic multimodal representation vector.
2. The gas engineering construction safety management system according to claim 1 is characterized in that: The pipeline pressure time series feature extraction module is used to: perform data regularization on the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values according to the time dimension to obtain a sequence of pipeline pressure time series input vectors; and pass each pipeline pressure time series input vector in the sequence of pipeline pressure time series input vectors through a pipeline pressure time series feature extractor based on the 1D_CNN model to obtain a sequence of pipeline pressure time series feature vectors.
3. The gas engineering construction safety management system according to claim 2 is characterized in that: The energy function value calculation unit is used to: calculate the mean and variance of the pipeline pressure time series feature vectors predetermined in the sequence of the pipeline pressure time series feature vectors to obtain the pipeline pressure key node time series semantic mean and the pipeline pressure key node time series semantic variance; determine the pipeline pressure key node time series semantic annotation sequence number of the predetermined pipeline pressure time series feature vector to obtain the predetermined pipeline pressure key node time series semantic annotation sequence number; calculate the sum of the pipeline pressure key node time series semantic variance and the pipeline pressure key node time series semantic mean and then multiply it by a constant 4 to obtain the first pipeline pressure key node time series semantic annotation sequence number; Sequence semantic modulation factor; respectively calculate the product of the pipeline pressure key node sequence semantic variance and the pipeline pressure key node sequence semantic mean and the constant 2 to obtain the double pipeline pressure key node sequence semantic variance and the double pipeline pressure key node sequence semantic mean; calculate the square of the difference between the pipeline pressure key node sequence semantic annotation serial number and the pipeline pressure key node sequence semantic mean, and then add it to the double pipeline pressure key node sequence semantic variance and the double pipeline pressure key node sequence semantic mean to obtain the second pipeline pressure key node sequence semantic modulation factor; The division between the first pipeline pressure key node timing semantic modulation factor and the second pipeline pressure key node timing semantic modulation factor is calculated to obtain an energy function value corresponding to the predetermined pipeline pressure timing feature vector.
4. The gas engineering construction safety management system according to claim 3 is characterized in that: The weighted enhancement unit is used to: calculate the reciprocal of each energy function value in the sequence of energy function values to obtain a sequence of energy function reciprocal representation values; The function activates the sequence of the inverse representation values of the energy function to obtain the sequence of the inverse representation values of the activated energy function; the product between each group of corresponding inverse representation values of the activated energy function and the pipeline pressure time series feature vector in the sequence of the inverse representation values of the activated energy function and the sequence of the pipeline pressure time series feature vector is calculated to obtain the sequence of enhanced pipeline pressure time series feature vectors, and each enhanced pipeline pressure time series feature vector in the obtained sequence of enhanced pipeline pressure time series feature vectors is summed by position to obtain the enhanced pipeline pressure global information aggregation feature vector.
5. The gas engineering construction safety management system according to claim 4 is characterized in that: The pipeline vibration waveform semantic feature extraction module is used to: pass the vibration signal of the gas engineering pipeline through a vibration signal waveform semantic feature extractor based on a depth-separable convolutional layer to obtain the gas engineering pipeline vibration waveform semantic feature vector.
6. The gas engineering construction safety management system according to claim 5 is characterized in that: The gas engineering pipeline anomaly detection module is used to: pass the pipeline state semantic multimodal representation vector through a classifier-based pipeline anomaly identifier to obtain a recognition result, and the recognition result is used to indicate whether there is an anomaly in the gas engineering pipeline.
7. A gas engineering construction safety management method according to the gas engineering construction safety management system according to claim 1, characterized in that: include: The time series of pressure values collected by pressure sensors deployed at multiple key nodes of a gas engineering pipeline is obtained to obtain the time series of pipeline pressure values at multiple key nodes; the vibration signal of the gas engineering pipeline is obtained; the time series of each key node pipeline pressure value in the time series of the multiple key node pipeline pressure values is respectively extracted based on pipeline pressure to obtain a sequence of pipeline pressure time series feature vectors; the sequence of pipeline pressure time series feature vectors is passed through a node energy attention aggregation module based on sequence annotation to obtain an enhanced pipeline pressure global information aggregation feature vector; the vibration signal of the gas engineering pipeline is subjected to semantic feature extraction based on the vibration signal waveform to obtain a gas engineering pipeline vibration waveform semantic feature vector; the enhanced pipeline pressure global information aggregation feature vector and the gas engineering pipeline vibration waveform semantic feature vector are passed through a pipeline multimodal semantic dynamic interaction fusion module based on gated response to obtain a pipeline state semantic multimodal representation vector; based on the pipeline state semantic multimodal representation vector, it is determined whether there is an abnormality in the gas engineering pipeline.
