Risk sensing and early warning method and system for operation state of oil and gas pipe network

Abstract

The invention discloses an oil and gas pipe network operation state risk perception early warning method and system, and the method comprises the steps: carrying out the fusion of the multi-dimensional features of the operation state of a pipe network according to the real-time operation data of the oil and gas pipe network, and generating a spatial-temporal feature matrix; inputting the spatial-temporal feature matrix into a deep Gaussian process model, performing probabilistic modeling on abnormal features in the pipe network operation state, and outputting a risk feature vector with confidence; constructing a three-dimensional risk concentration distribution field according to the risk feature vector, and predicting a propagation path and a diffusion trend of the risk concentration distribution field; and for the prediction result, constructing a self-adaptive early warning threshold generation model based on the generative adversarial network, generating a multi-stage early warning threshold changing along with the operation state of the pipe network, and realizing risk perception early warning of the operation state of the oil and gas pipe network according to the comparison result of the real-time risk characteristics and the early warning threshold. According to the embodiment of the invention, dynamic risk assessment and self-adaptive early warning can be realized, and the safety and emergency response capability of the pipe network are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of oil and gas pipeline networks, and in particular to a method and system for perceiving and warning the risk of operating status of an oil and gas pipeline network. Background Art

[0002] As global energy demand continues to grow, oil and gas pipeline networks, as important energy transportation channels, bear huge transportation pressure and flow. Therefore, ensuring the safe and efficient operation of oil and gas pipeline networks has become a major issue that needs to be urgently addressed in the energy industry. Traditional pipeline network monitoring methods mainly rely on regular inspections and manual monitoring, which makes it difficult to achieve real-time perception and accurate early warning of complex operating conditions. This method has many defects such as slow response and high misjudgment rate when responding to sudden dangers, which may lead to significant economic losses and environmental damage. Summary of the invention

[0003] The purpose of the present invention is to provide a method and system for risk perception and early warning of the operating status of an oil and gas pipeline network, so as to address the deficiencies in the prior art, realize dynamic risk assessment and adaptive early warning, and enhance the safety and emergency response capabilities of the pipeline network.

[0004] An embodiment of the present application provides a method for early warning of oil and gas pipeline network operation status risk perception, the method comprising: Based on the real-time operation data of the oil and gas pipeline network, the multi-dimensional characteristics of the pipeline network operation status are integrated to generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; The spatiotemporal feature matrix is ​​input into the deep Gaussian process model to perform probabilistic modeling on the abnormal features in the operation status of the pipeline network and output a risk feature vector with confidence. According to the risk characteristic vector, a three-dimensional risk concentration distribution field is constructed, and the propagation path and diffusion trend of the risk concentration distribution field are predicted based on the graph neural network; According to the prediction results of the risk concentration distribution field, an adaptive warning threshold generation model based on the generative adversarial network is constructed. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level warning threshold that changes with the operating status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the warning threshold, the risk-aware warning of the operating status of the oil and gas pipeline network is realized.

[0005] Optionally, the multi-dimensional features of the pipeline network operation status are integrated according to the real-time operation data of the oil and gas pipeline network to generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information, including: According to the real-time operation data collected by multi-source sensors of the oil and gas pipeline network, timestamp alignment and missing value filling processing are performed, and a multi-source data alignment algorithm based on dynamic time warping is used to eliminate the difference in sensor sampling frequency and obtain a time-synchronized multi-source data sequence, wherein the real-time operation data includes pressure, temperature, flow and vibration data; The multi-source data sequence is input into a spatiotemporal convolutional neural network, wherein a graph convolutional network is used in the spatial dimension to capture the topological structure characteristics of the pipeline network, and a one-dimensional convolutional kernel is used in the temporal dimension to extract the time series characteristics, thereby obtaining a preliminary spatiotemporal feature representation; For the extracted spatiotemporal feature representation, the multi-head self-attention mechanism is used to calculate the correlation weights between different sensor data, and the features are weighted fused according to the weights to highlight the contribution of key sensor data to risk perception, thus obtaining the weighted spatiotemporal feature representation; The weighted spatiotemporal features are input into the sparse autoencoder, and the dimension is reduced through nonlinear mapping to remove redundant information, retain key features, and generate a low-dimensional spatiotemporal feature matrix containing pressure, temperature, flow and vibration information.

[0006] Optionally, the spatiotemporal feature matrix is ​​input into a deep Gaussian process model, the abnormal features in the operation status of the pipeline network are probabilistically modeled, and a risk feature vector with confidence is output, including: Input the spatiotemporal feature matrix into a multi-layer perceptron, and map the original features to a high-dimensional latent space through multi-layer nonlinear transformation to obtain a deep feature representation, wherein each layer uses an activation function with a residual connection to avoid the gradient vanishing problem and enhance the feature expression capability; For the deep feature representation, a probability model based on Gaussian process regression is constructed, and a linear combination of radial basis kernel function and periodic kernel function is used to capture the nonlinear relationship and periodic change law in the operation state of the pipeline network, and obtain a preliminary probabilistic feature representation; According to the probabilistic feature representation, the predicted mean and variance of each feature point are calculated, and the uncertainty of the model is quantified through a variational inference algorithm to obtain a probabilistic feature distribution with a confidence interval; For the probability feature distribution, an anomaly detection algorithm based on Mahalanobis distance is used to calculate the distance between each feature point and the normal operating condition distribution, and the abnormal feature points are screened out according to a preset confidence threshold to generate a risk feature vector with confidence.

[0007] Optionally, constructing a three-dimensional risk concentration distribution field according to the risk feature vector, and predicting the propagation path and diffusion trend of the risk concentration distribution field based on a graph neural network includes: According to the risk characteristic vector, combined with the topological structure information of the oil and gas pipeline network, each risk characteristic point is mapped to the three-dimensional spatial coordinates of the pipeline network, and the Kriging interpolation algorithm is used to perform spatial interpolation on the discrete risk characteristic points to generate a preliminary three-dimensional risk concentration distribution field; Based on the preliminary three-dimensional risk concentration distribution field, combined with the time dimension information, the spatiotemporal Kriging interpolation algorithm is used, combined with the historical risk propagation data, to dynamically correct the risk concentration distribution field and obtain the three-dimensional risk concentration distribution field that changes with time; The three-dimensional risk concentration distribution field that changes with time is input into the graph neural network, and a graph model is constructed according to the topological structure of the pipeline network. The nodes of the graph model represent the key positions of the pipeline network, and the edges of the graph model represent the connection relationship of the pipeline sections. The graph attention mechanism is used to capture the risk propagation dependency between nodes to obtain the initial prediction result of risk propagation. According to the initial prediction results of risk propagation, a sequence prediction model based on time convolutional network is adopted, combined with historical risk propagation path data, to predict the risk propagation path and diffusion trend in future time steps, and generate the propagation prediction results of dynamic risk concentration distribution field.

[0008] Optionally, the prediction results of the risk concentration distribution field are used to construct an adaptive warning threshold generation model based on a generative adversarial network, and a multi-level warning threshold that changes with the operation status of the pipeline network is generated through a dynamic game between a normal operating condition generator and a real-time discriminator. Based on the comparison results of the real-time risk characteristics and the warning threshold, the risk perception warning of the operation status of the oil and gas pipeline network is realized, including: Based on historical normal operation data, a normal operating condition generator based on a generative adversarial network is trained to generate simulated data that conforms to the normal operation characteristics of the pipeline network. The generator uses a variational autoencoder with conditional constraints to ensure the diversity and authenticity of the generated data. Construct a real-time discriminator based on a deep residual network, input the prediction results of the risk concentration distribution field and the normal operating condition data generated by the generator, learn the boundary characteristics of normal and abnormal states through dynamic game, and output the discrimination results and their confidence; According to the discrimination results, an adaptive clustering algorithm is used to divide the discrimination results into multiple levels, and combined with the dynamic changes of the risk concentration distribution field, a multi-level warning threshold that changes with the operation status of the pipeline network is generated. The multi-level warning threshold includes low risk, medium risk and high risk thresholds; The real-time risk characteristics are compared with the multi-level warning thresholds, and a fuzzy logic-based decision model is used to dynamically adjust the warning level according to the confidence level of the risk characteristics and the degree of threshold deviation, thereby realizing risk-aware warning of the operating status of the oil and gas pipeline network.

[0009] Another embodiment of the present application provides an oil and gas pipeline network operation status risk perception and early warning system, the system comprising: The fusion module is used to fuse the multi-dimensional characteristics of the pipeline network operation status according to the real-time operation data of the oil and gas pipeline network, and generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; The modeling module is used to input the spatiotemporal feature matrix into the deep Gaussian process model, perform probabilistic modeling on the abnormal features in the operation status of the pipeline network, and output a risk feature vector with confidence; A construction module is used to construct a three-dimensional risk concentration distribution field according to the risk characteristic vector, and predict the propagation path and diffusion trend of the risk concentration distribution field based on a graph neural network; The early warning module is used to build an adaptive early warning threshold generation model based on the generative adversarial network according to the prediction results of the risk concentration distribution field. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level early warning threshold that changes with the operation status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the early warning threshold, the risk perception early warning of the operation status of the oil and gas pipeline network is realized.

[0010] Yet another embodiment of the present application provides a storage medium, wherein the storage medium stores a computer program, wherein the computer program is configured to execute any of the above methods when running.

[0011] Yet another embodiment of the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute any of the methods described above.

[0012] Compared with the prior art, the present invention provides a method for risk perception and early warning of the operation status of an oil and gas pipeline network. According to the real-time operation data of the oil and gas pipeline network, the multi-dimensional characteristics of the operation status of the pipeline network are integrated to generate a spatiotemporal feature matrix; the spatiotemporal feature matrix is ​​input into a deep Gaussian process model, the abnormal characteristics in the operation status of the pipeline network are probabilistically modeled, and a risk feature vector with confidence is output; according to the risk feature vector, a three-dimensional risk concentration distribution field is constructed to predict the propagation path and diffusion trend of the risk concentration distribution field; according to the prediction results, an adaptive early warning threshold generation model based on a generative adversarial network is constructed to generate multi-level early warning thresholds that change with the operation status of the pipeline network, and according to the comparison results of the real-time risk characteristics and the early warning thresholds, the risk perception and early warning of the operation status of the oil and gas pipeline network are realized, thereby realizing dynamic risk assessment and adaptive early warning, and enhancing the safety and emergency response capabilities of the pipeline network. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 A hardware structure block diagram of a computer terminal for a method for sensing and warning the operating status risk of an oil and gas pipeline network provided by an embodiment of the present invention; Figure 2A flow chart of a method for sensing and warning the operating status risk of an oil and gas pipeline network provided by an embodiment of the present invention; Figure 3 A schematic structural diagram of an oil and gas pipeline network operation status risk perception and early warning system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, but should not be construed as limiting the present invention.

[0015] The embodiment of the present invention firstly provides a method for risk perception and early warning of the operation status of an oil and gas pipeline network. The method can be applied to electronic devices, such as computer terminals, specifically ordinary computers, etc.

[0016] The following describes it in detail by taking running on a computer terminal as an example. Figure 1 The hardware structure block diagram of a computer terminal of an oil and gas pipeline network operation status risk perception and early warning method provided by an embodiment of the present invention. Figure 1 As shown, the computer device includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

[0017] The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any one of the oil and gas pipeline network operation status risk perception and early warning methods.

[0018] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0019] The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any one of the oil and gas pipeline network operation status risk perception and early warning methods.

[0020] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 1 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0021] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0022] See also Figure 2 The embodiment of the present invention provides a method for risk perception and early warning of oil and gas pipeline network operation status, which may include the following steps: S201, based on the real-time operation data of the oil and gas pipeline network, the multi-dimensional features of the pipeline network operation status are integrated to generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; This step aims to generate a spatiotemporal feature matrix by fusing multi-dimensional features of real-time operation data from the oil and gas pipeline network to reflect the comprehensive operation status of the pipeline network. First, the system collects key data related to the operation of the oil and gas pipeline network from multi-source sensors, including pressure, temperature, flow, and vibration. These data usually exist in the form of time series, and the sampling frequency of each sensor may be different, resulting in the problem of time alignment of the data. To solve this problem, the system uses the dynamic time warping (DTW) algorithm to ensure that the data from different sensors are synchronized in the time dimension, thereby obtaining an integrated multi-source data sequence.

[0023] Then, after data alignment, the system integrates these real-time operation data in two dimensions, space and time, to form a space-time feature matrix. In this process, the space-time convolutional neural network (ST-CNN) is used to extract spatial topological information and time series features. Specifically, the spatial dimension uses a graph convolutional network to capture the topological structure characteristics of the pipeline network, and the time dimension uses a one-dimensional convolution kernel to extract time series features. Finally, the fused space-time feature matrix simultaneously contains multiple operating status information of the pipeline network, laying the foundation for subsequent risk identification and prediction.

[0024] By integrating the multi-dimensional features of the real-time operation data of the oil and gas pipeline network, the generated spatiotemporal feature matrix can effectively reflect the overall operation status of the pipeline network. This process not only improves the efficiency of data utilization, but also provides more accurate input data for subsequent risk identification. For example, by comprehensively considering multiple factors such as pressure, temperature, flow and vibration, the system can have a more comprehensive understanding of the health status of the pipeline network, so as to issue early warnings in time when abnormalities occur. The establishment of such comprehensive features helps to improve the operational safety of the oil and gas pipeline network, provide a basis for risk management, and reduce potential safety hazards.

[0025] Specifically, the real-time operation data collected by multi-source sensors of the oil and gas pipeline network can be used to perform timestamp alignment and missing value filling processing, and a multi-source data alignment algorithm based on dynamic time warping can be used to eliminate the difference in sensor sampling frequency and obtain a time-synchronized multi-source data sequence, wherein the real-time operation data includes pressure, temperature, flow and vibration data; In this step, the system first collects real-time operating data from multi-source sensors in the oil and gas pipeline network. The data types cover key parameters such as pressure, temperature, flow, and vibration. However, data from different sensors often have inconsistent timestamps, which makes subsequent analysis and modeling difficult. In order to effectively solve this problem, the system uses the dynamic time warping (DTW) algorithm to align data of different time frequencies to ensure that all sensor data can be analyzed on the same time scale.

[0026] In the specific implementation process, the system will first pre-process the data of each sensor, including removing noise and filling missing values. Missing values ​​can be filled through interpolation methods, such as linear interpolation or spline interpolation, to ensure the continuity and integrity of the data. Then, the time series of each sensor is compared through the DTW algorithm to identify and adjust the shortest time axis for optimal alignment. Finally, the system will generate a time-synchronized multi-source data sequence, providing a standardized data basis for subsequent feature extraction.

[0027] By time-aligning and processing missing values ​​for multi-source sensor data, the system can obtain high-quality data sequences that can be used for analysis. This step not only improves the integrity of the data, but also lays a solid foundation for subsequent feature extraction. By effectively eliminating differences in sensor sampling frequencies, the accuracy of the data used for subsequent analysis can be ensured, thereby improving the risk identification and early warning capabilities of the operating status of the oil and gas pipeline network. This improvement in data quality helps to achieve real-time monitoring and fault prediction of the pipeline network, and enhance the safety of pipeline network operations.

[0028] In this step, the system first collects real-time operating data from various multi-source sensors installed at different locations in the oil and gas pipeline network to monitor key parameters such as pressure, temperature, flow, and vibration. Since the operating frequencies of different sensors may be different, resulting in inconsistent data timestamps, the data needs to be time-aligned. The system will use the dynamic time warping (DTW) algorithm, which is a method that can handle the similarity of time series. It performs time alignment by calculating the optimal matching path between different time series.

[0029] In the specific implementation process, the data collected by each sensor is first preprocessed, including removing noise and filling missing values. The system will use linear interpolation to fill the missing data, that is, fill the missing part by interpolating the existing data points before and after to ensure the continuity of the data. Next, the DTW algorithm is used for time alignment. The system will build a distance matrix to record the distance between all sensor data points. The distance matrix is ​​used to determine the time matching path to eliminate the impact of different sampling frequencies.

[0030] Finally, the aligned multi-source data sequences will be integrated to form a time-synchronized multi-source data sequence. In this process, the system will ensure that the data of all sensors are balanced at the same timestamp, so that subsequent processing and analysis can be performed on a consistent data basis. The completion of this step provides a solid data foundation for subsequent feature extraction and ensures the integrity and availability of the data.

[0031] The multi-source data sequence is input into a spatiotemporal convolutional neural network, wherein a graph convolutional network is used in the spatial dimension to capture the topological structure characteristics of the pipeline network, and a one-dimensional convolutional kernel is used in the temporal dimension to extract the time series characteristics, thereby obtaining a preliminary spatiotemporal feature representation; In this step, the multi-source data sequence after time alignment and missing value processing will be input into the spatiotemporal convolutional neural network (ST-CNN) to extract richer feature information. This network structure is specially designed to process data in both spatial and temporal dimensions at the same time, and is particularly suitable for describing systems with complex topological structures such as pipeline networks. The spatial dimension of the system uses a graph convolutional network (GCN) to capture the topological features of the pipeline network, such as the connection relationship of the pipelines, the relative position of each node and its attributes. In addition, in the time dimension, the system uses a one-dimensional convolution kernel to extract time series features and captures time change information through a sliding window.

[0032] In specific implementation, the system divides the multi-source data sequence into multiple small windows so that it can process in parallel and extract relevant features of each time period. Through GCN, the system can identify the characteristics of each node in the pipeline network (such as pumping stations, valves, etc.) and its connected pipelines to generate a spatial feature matrix. After obtaining the spatial features, the one-dimensional convolution kernel continues to perform convolution operations on the time series to capture the key change trends and periodic signals in the time series. Ultimately, the spatiotemporal convolutional neural network will generate a preliminary spatiotemporal feature representation that fully reflects the operating status of the oil and gas pipeline network.

[0033] Using spatiotemporal convolutional neural networks to process multi-source data can effectively combine the characteristics of space and time to extract deeper feature information. The fusion of spatial and temporal features provides strong data support for subsequent anomaly detection and risk assessment. By identifying the topological structure and dynamic changes of the pipeline network, the system can better understand the operating status of the pipeline network and thus predict potential risks. This comprehensive feature representation greatly improves the accuracy and efficiency of data analysis and lays a solid foundation for the intelligent management of oil and gas pipeline networks.

[0034] In this step, the system inputs the time-aligned and preprocessed multi-source data sequence into the spatiotemporal convolutional neural network (ST-CNN). The network is designed to capture both spatial and temporal features of the data, and is particularly suitable for processing oil and gas pipeline networks with complex topological structures. The system first structures the input data to ensure that each data point can correspond to a node in the pipeline network, facilitating the subsequent extraction of spatial features.

[0035] To extract the spatial features of the pipeline network, the system uses a graph convolutional network (GCN) to process the data. By processing the topological structure information of the pipeline network, such as the connection relationship of the pipelines and the location of the sensors, GCN can effectively identify the relationship between the nodes in the pipeline network and generate a spatial feature representation. At the same time, GCN can use the characteristics of the graph structure to mine the connection patterns and local characteristics that are particularly important for risk perception. The effect of this process is directly related to the accuracy of the subsequent risk assessment, so the focus is on ensuring that the information of each node is fully transmitted and processed.

[0036] The feature extraction of the time dimension is performed using a one-dimensional convolution kernel. The system performs sliding window processing on each time series and adjusts the size of the convolution kernel to capture the change information in different time periods. For example, in order to identify abnormal fluctuations in flow in a specific time period, the system uses a small convolution kernel to extract local features, while for the analysis of longer time trends, a larger convolution kernel is used. Ultimately, through feature extraction in both space and time dimensions, the system can generate a preliminary spatiotemporal feature representation that comprehensively reflects the operating status of the pipeline network.

[0037] For the extracted spatiotemporal feature representation, the multi-head self-attention mechanism is used to calculate the correlation weights between different sensor data, and the features are weighted fused according to the weights to highlight the contribution of key sensor data to risk perception, thus obtaining the weighted spatiotemporal feature representation; In this step, the system applies a multi-head self-attention mechanism to the previously extracted spatiotemporal feature representations to calculate the correlation weights between different sensor data. The self-attention mechanism allows the system to dynamically focus on the most important information in a specific input data, which is particularly important for risk perception in oil and gas pipeline networks. Specifically, the system calculates the correlation between each group of features based on the extracted spatiotemporal feature representations to determine which sensor data has a greater contribution in risk identification.

[0038] The specific implementation process includes that the system first inputs the spatiotemporal feature representation into the multi-head self-attention module, where each head independently calculates the weighted average of the feature vector, reflecting the importance of specific features to the overall risk assessment. In this way, the system can effectively identify key sensor data. For example, under a certain working condition, flow data and pressure changes may have a higher correlation. Ultimately, the weighted feature representation will highlight the contribution of key sensors to risk perception, which will help subsequent anomaly detection and risk assessment.

[0039] This process significantly enhances the system's sensitivity to key features. By focusing on key sensor data that affects risk perception, the system can more accurately reflect the operating status of the pipeline network and effectively improve the accuracy of risk prediction. For example, changes in flow or temperature may be more significant when pressure is abnormal. The use of a self-attention mechanism can ensure that these key data are given proper attention in risk assessment. The final weighted feature representation provides a solid information foundation for subsequent risk feature extraction and prediction, improving the effectiveness of pipeline network risk management.

[0040] In this step, the system processes the spatiotemporal feature representations extracted in the previous step and uses the multi-head self-attention mechanism to calculate the correlation weights between different sensor data. Specifically, the system uses the self-attention mechanism to identify the features that have a greater impact on risk perception under specific conditions, and by comparing the correlations between different sensor data, it determines which data has a more significant impact on the overall risk assessment.

[0041] In the specific implementation, the system first performs a linear transformation on the spatiotemporal feature representation to generate query, key, and value vectors for multiple heads. Each head independently calculates the attention weights between its features, and then applies these weights to the corresponding value vectors to achieve weighted averaging. Through the multi-head mechanism, the system is able to extract information from different subspaces, thereby comprehensively considering the relationship between multiple features and strengthening the influence of important features. For example, when a drastic change in traffic is detected, the system can automatically adjust the weights so that the data from the flow sensor accounts for a larger proportion in subsequent analysis.

[0042] Finally, the system concatenates the outputs of all heads to form a weighted spatiotemporal feature representation. This weighting process ensures that sensor data that is critical to risk perception is highlighted, reduces the interference of redundant information, and provides more accurate input data for subsequent risk assessment. In this way, the system can effectively improve its sensitivity to the operating status of the oil and gas pipeline network and further enhance the effectiveness of the entire risk management process.

[0043] The weighted spatiotemporal features are input into the sparse autoencoder, and the dimension is reduced through nonlinear mapping to remove redundant information, retain key features, and generate a low-dimensional spatiotemporal feature matrix containing pressure, temperature, flow and vibration information.

[0044] In this step, the system inputs the weighted spatiotemporal features into the sparse autoencoder to achieve dimensionality reduction of the features. Sparse autoencoders are an effective nonlinear dimensionality reduction tool that can remove redundant data while retaining key information. In the analysis of the operating status of oil and gas pipeline networks, due to the complexity of multi-dimensional features, directly processing high-dimensional data may lead to problems such as low computational efficiency and overfitting. Therefore, through the nonlinear mapping of the sparse autoencoder, the system can obtain a low-dimensional feature representation, which is conducive to subsequent data analysis and risk assessment.

[0045] During the specific implementation process, the system will first define the network structure of the autoencoder, including the input layer, hidden layer, and output layer. The input layer receives the weighted spatiotemporal features, and after being transformed by the nonlinear activation function (such as ReLU) of the hidden layer, the system will reconstruct the input data in the output layer. By introducing sparsity constraints, the system can force some nodes in the hidden layer to be zero, thereby achieving the effect of feature selection and dimensionality reduction. Ultimately, the sparse autoencoder will generate a low-dimensional spatiotemporal feature matrix containing pressure, temperature, flow, and vibration information, which serves as the basis for subsequent risk feature modeling.

[0046] Through the dimensionality reduction process of the sparse autoencoder, the system can efficiently remove redundant information and retain features that are critical to risk identification. This feature compression significantly improves the efficiency of data processing, reduces computational complexity, and reduces the risk of overfitting. In addition, the low-dimensional feature matrix after redundancy removal makes subsequent model training and prediction more efficient, ensuring that the system still has good responsiveness and prediction accuracy when facing complex oil and gas pipeline network data. This effective feature processing method provides strong technical support for risk perception and management of oil and gas pipeline networks.

[0047] In this process, the system will input the weighted spatiotemporal features into the sparse autoencoder to achieve feature dimensionality reduction and removal of redundant information. Sparse autoencoder is an efficient neural network structure that can map high-dimensional data to low-dimensional space through learning while retaining the most critical information in the data. In the specific implementation process, the system first constructs the network structure of the sparse autoencoder, including the input layer, hidden layer and output layer. The input layer directly receives the weighted spatiotemporal features, while the hidden layer is responsible for the compression and mapping of the features.

[0048] The system uses nonlinear activation functions to perform nonlinear transformations of features in the hidden layer. For example, ReLU or LSTM activation functions can help the model capture complex feature associations. During training, the system optimizes network weights through a back-propagation algorithm, with the goal of making the data reconstructed by the output layer as close as possible to the data in the input layer. To enhance sparsity, the system introduces L1 regularization in training to force some activations of neurons to zero, thereby removing redundant features and retaining only key feature information.

[0049] Finally, after being processed by the sparse autoencoder, the system will generate a low-dimensional spatiotemporal feature matrix that contains key pressure, temperature, flow and vibration information. This transformation provides a more streamlined and efficient data input for subsequent risk modeling and anomaly detection, and is more suitable for large-scale data processing and analysis.

[0050] Exemplary: 1. Multi-source data collection and preprocessing 1.1 Sensor Type Pressure sensor (such as Honeywell ST3000 series, sampling frequency 1Hz); Temperature sensor (such as PT100 thermocouple, sampling frequency 0.5Hz); Flow meter (such as Emerson high-precision turbine flow meter, sampling frequency 2Hz); Vibration sensor (such as ICP accelerometer, sampling frequency 10Hz).

[0051] 1.2 Data alignment and missing value processing Input: pressure sequence P = [p1, p2, ...] (1 Hz), temperature sequence T = [t1, t3, ...] (0.5 Hz).

[0052] Operation: Align the time axis by calculating the optimal path through DTW, for example, interpolating the temperature series to 1 Hz.

[0053] Missing value filling: linear interpolation is used (for example, when t2 is missing, t2=(t1+t3) / 2).

[0054] Output: The synchronized multi-source data matrix, each row represents the timestamp, and each column represents the sensor parameters, as shown in Table 1: Table 1 Timestamp Pressure (MPa) Temperature (°C) Flow rate (m³ / s) Vibration (g) t1 5.2 30.1 10.5 0.01 t2 5.3 30.5 10.7 0.02 2. Spatiotemporal feature extraction 2.1. Spatiotemporal Convolutional Neural Network (ST-CNN) Graph model construction: nodes are pumping stations (such as nodes A and B), valves (such as V1 and V2), and edges represent pipeline connections.

[0055] Adjacency matrix: If node A is connected to V1, then A[0][1]=1.

[0056] Feature aggregation: Aggregate neighboring node features using GCN layers (such as the PyTorch Geometric library).

[0057] Time dimension (1D convolution): The convolution kernel size is 3, and the sliding window extracts traffic time series features (such as local fluctuation trends).

[0058] 2.2 Multi-head self-attention mechanism Input: Spatiotemporal feature matrix (dimension [timesteps × number of nodes × number of features]).

[0059] Operation: Linear transformation generates Q, K, and V vectors (number of heads = 4); calculates attention weights (such as the correlation weight between flow and pressure = 0.7); outputs the feature matrix after weighted fusion.

[0060] 2.3 Sparse Autoencoder (SAE) Structure: input layer (64 dimensions) → hidden layer (32 dimensions, ReLU activation, L1 regularization) → output layer (64 dimensions).

[0061] Output: low-dimensional matrix after dimensionality reduction (such as 32 dimensions).

[0062] S202, inputting the spatiotemporal feature matrix into the deep Gaussian process model, probabilistically modeling the abnormal features in the operation status of the pipeline network, and outputting a risk feature vector with confidence; In this step, the spatiotemporal feature matrix generated by the previous processing is input into the deep Gaussian process model to perform probabilistic modeling of abnormal characteristics of the operating status of the oil and gas pipeline network. The advantage of the deep Gaussian process model is that it can flexibly capture complex nonlinear relationships in the data and handle uncertainty. By performing multi-layer nonlinear transformations on the input spatiotemporal features, the model can map the original features to a high-dimensional latent space to extract deeper feature information and enhance the expressiveness of the features. In this process, each layer of the network uses an activation function with residual connections to avoid the common gradient vanishing problem in traditional deep network training, thereby ensuring effective training and stability of the model.

[0063] After the model training is completed, the system will construct a probability model of Gaussian process regression based on the extracted deep feature representation. This model can effectively capture the nonlinear relationship and periodic change law in the operation status of the pipeline network by combining the radial basis kernel function and the periodic kernel function. This structure enables the model to not only obtain the predicted mean of the feature, but also calculate the corresponding predicted variance, and finally generate a probability feature distribution with a confidence interval for each feature point. This information will help with subsequent risk assessment and decision making.

[0064] By inputting the spatiotemporal feature matrix into the deep Gaussian process model, the system can accurately identify and quantitatively evaluate potential abnormal features in the operation status of the pipeline network. The implementation of this step not only enhances the ability to perceive risks, but also provides high-confidence decision support in dynamic situations. By outputting risk feature vectors with confidence, the system can quickly judge the safety status of the oil and gas pipeline network in actual operation, provide a scientific basis for the formulation of emergency response strategies, thereby effectively reducing the risk of accidents and ensuring the safety and stability of the oil and gas pipeline network.

[0065] Specifically, the spatiotemporal feature matrix can be input into a multi-layer perceptron, and the original features can be mapped to a high-dimensional latent space through multi-layer nonlinear transformation to obtain a deep feature representation, wherein each layer uses an activation function with a residual connection to avoid the gradient vanishing problem and enhance the feature expression capability; In this step, the system inputs the processed spatiotemporal feature matrix into a multilayer perceptron (MLP) to achieve deep learning and abstraction of the original features. The design of the multilayer perceptron includes multiple fully connected layers, and the neurons in each layer are connected to all the neurons in the previous layer. This structure enables the MLP to capture complex data patterns. Through layer-by-layer nonlinear transformations, the system can gradually map the original features into a high-dimensional latent space, thereby extracting deeper and more abstract feature information.

[0066] In order to solve the gradient vanishing problem in traditional deep networks, the system uses an activation function with residual connections in each layer. Residual connections allow the network to skip the direct output of a certain layer and combine the output of the previous layer with the output of the next layer, thereby effectively improving the training performance of the model. This structure can not only accelerate convergence, but also enhance the ability to express features, so that the model can still maintain good learning effects when facing high-dimensional features. After multiple layers of processing, the system will obtain a deep feature representation that contains rich information about the input spatiotemporal features.

[0067] The implementation of this process enables the system to perform efficient feature extraction and representation learning in high-dimensional feature space, greatly improving the model's expressiveness and complex pattern recognition capabilities. The resulting deep feature representation will provide a more comprehensive foundation for subsequent probabilistic modeling, making the identification of risk features more accurate, thereby facilitating safety monitoring and risk management of oil and gas pipeline networks.

[0068] At the initial stage of this process, the system first designs the architecture of a multi-layer perceptron (MLP), which usually consists of an input layer, multiple hidden layers, and an output layer. The input layer is responsible for receiving the processed spatiotemporal feature matrix, which contains important information such as pressure, temperature, flow, and vibration. The hidden layer of each layer uses a fully connected structure to use the output of the current layer as the input of the next layer. For example, assuming that the feature dimension of the input layer is 32, after the first hidden layer, the feature dimension may be expanded to 64. This change is achieved through the learned weight matrix.

[0069] In each hidden layer, the system applies nonlinear activation functions, such as ReLU (Rectified Linear Unit), to increase the nonlinear expression ability of the model. This process also introduces the design of residual connections. Specifically, certain features are directly passed from the previous layer to the subsequent layer without being transformed in the current layer. This method effectively solves the gradient vanishing problem that may occur in deep networks, allowing the network to be trained faster and more efficiently. For example, in this design, if the output of a layer is small, the residual connection can retain the information of the previous layer, preventing the information from disappearing during the propagation process, thereby increasing the learning ability of the network.

[0070] Finally, after multiple layers of nonlinear transformation, the system will obtain deep feature representation. This representation not only extracts important information from the original features, but also captures more complex feature correlations. These deep features will be used for anomaly detection in the subsequent Gaussian process regression model because they can provide more discriminative inputs and improve the overall performance and accuracy of the model.

[0071] For the deep feature representation, a probability model based on Gaussian process regression is constructed, and a linear combination of radial basis kernel function and periodic kernel function is used to capture the nonlinear relationship and periodic change law in the operation state of the pipeline network, and obtain a preliminary probabilistic feature representation; In this step, the system constructs a Gaussian process regression (GPR) model based on the deep feature representation obtained in the previous step. Gaussian process regression is a flexible non-parametric Bayesian modeling method that can effectively handle complex functional relationships and is suitable for state prediction of oil and gas pipeline networks. When constructing the GPR model, the system uses a linear combination of radial basis kernel functions and periodic kernel functions. This combination strategy can simultaneously capture the nonlinear relationship and periodic variation law in the operation status of the pipeline network.

[0072] The radial basis kernel function can well represent the similarity between input features. Its characteristics enable the model to reflect the similarity of data points that are close to each other in the feature space. The periodic kernel function is specifically used to model periodic patterns in data. By introducing periodic components in the Gaussian process, the system can better adapt to the periodic fluctuations of the network operation status. After processing these kernel functions, the system will obtain a preliminary probabilistic feature representation, which contains a rich description of the risk status.

[0073] By constructing a Gaussian process regression model, the system can synthesize the operating status of the oil and gas pipeline network into a probabilistic feature representation. This representation not only reflects the current state of the pipeline network, but also predicts the future state and provides the confidence level of each feature point. This information not only enriches the dimension of risk assessment, but also provides a reliable basis for subsequent anomaly detection, allowing the system to respond in a timely manner in a dynamic environment and reduce the risk of accidents.

[0074] In this step, the system will use deep feature representation to construct a Gaussian process regression (GPR) model. First, the system selects a kernel function, which is a combination of a radial basis kernel function and a periodic kernel function. The radial basis kernel function (RBF) can effectively capture the similarity between feature points, while the periodic kernel function targets the periodic characteristics of the data, which is especially important in an environment with seasonal changes such as oil and gas pipeline networks. For example, flow rates often show different patterns of change in summer and winter, and combining periodic kernel functions can help the model accurately capture these characteristics.

[0075] In specific implementation, the system will first define the prior distribution of the Gaussian process and construct a covariance matrix based on the training data. By linearly combining the radial basis kernel with the periodic kernel, the system can obtain a covariance function that comprehensively considers the effects of distance and periodicity. In this process, the system will optimize the hyperparameters of the kernel function, such as selecting appropriate length scales and periodicity parameters, to ensure that the model performs best in practical applications. After the model training is completed, the system can generate a probability distribution that represents the possible states of the input features under specific conditions.

[0076] After this processing, the initial probabilistic feature representation obtained can reflect the complexity of the pipeline network operation status. For example, the flow characteristics of a certain node may fluctuate abnormally within a specific time window, and the Gaussian process model can identify this abnormal state by measuring its difference from the normal feature distribution. The final output is the mean and variance corresponding to each feature point, forming a set of probabilistic features that can guide subsequent risk assessment.

[0077] According to the probabilistic feature representation, the predicted mean and variance of each feature point are calculated, and the uncertainty of the model is quantified through a variational inference algorithm to obtain a probabilistic feature distribution with a confidence interval; In this process, the system uses the variational inference algorithm to further analyze the probabilistic feature representation obtained by the Gaussian process regression model to calculate the predicted mean and variance of each feature point. Variational inference is an effective inference method that is mainly used to deal with uncertainty in the model. It approximates the true distribution by constructing an approximate distribution. In this step, the system will minimize the difference between the approximate posterior distribution and the true posterior distribution through the optimization algorithm, thereby quantifying the uncertainty and risk of the model.

[0078] Specifically, the system will calculate the predicted mean and predicted variance for each feature point, where the predicted mean represents the best estimate of the feature point state, and the predicted variance indicates the uncertainty of this estimate. This information is obtained by weighted calculation of the observed data. Furthermore, the system will construct a confidence interval based on the obtained variance and generate a probability feature distribution for each feature point, reflecting the uncertainty characteristics under different risk states.

[0079] The implementation of this process enables the system to quantitatively evaluate the operating status of the oil and gas pipeline network and provide uncertainty information for the prediction model. The probability characteristic distribution with confidence intervals not only helps to identify potential risk points in the pipeline network, but also provides a reliable basis for subsequent decision-making. This capability enables oil and gas pipeline network operators to make more robust decisions in complex environments, reduce safety hazards, and ensure the safety of operations.

[0080] In this step, the system will use the probabilistic feature representation obtained from the Gaussian process regression model to calculate the predicted mean and variance of each feature point. In the specific implementation, the system will apply the variational inference algorithm to each feature point to quantify the uncertainty of the model. Variational inference is an approximate inference technique. The system will obtain an approximation of the posterior distribution by setting an appropriate prior distribution and optimizing it using observed data. The system optimizes the parameters of the approximate posterior distribution by maximizing the variational lower bound to make it as close to the true posterior distribution as possible.

[0081] For example, the system can set a Gaussian distribution prior for each feature point, where the mean represents the possible state and the variance reflects the uncertainty of the state. When making inferences, the system independently calculates each feature point, introduces sample points, and updates the mean and variance based on the observed data. Ultimately, the system will generate a probabilistic feature distribution containing a confidence interval, which provides a basis for quantifying the risk level.

[0082] This process improves the model's explanatory power, because the predicted mean with confidence interval not only indicates the possible state of each feature point, but also enables decision makers to understand the credibility of the data. By combining the mean and variance, the system can provide the necessary information for subsequent risk assessment, allowing managers to make more scientific decisions when judging risks.

[0083] For the probability feature distribution, an anomaly detection algorithm based on Mahalanobis distance is used to calculate the distance between each feature point and the normal operating condition distribution, and the abnormal feature points are screened out according to a preset confidence threshold to generate a risk feature vector with confidence.

[0084] In this step, the system will analyze the obtained probability feature distribution through anomaly detection algorithm based on Mahalanobis distance to identify abnormal feature points. Mahalanobis distance is an effective distance measurement method that can take into account the covariance matrix of the data to better reflect the relationship between different features. In this process, the system first needs to establish a distribution model of normal working conditions, which is usually obtained through historical data statistics. This model will be used as a reference standard.

[0085] Next, for each feature point, the system calculates the Mahalanobis distance between it and the normal operating condition distribution to evaluate the abnormality of the current feature point. If the Mahalanobis distance exceeds the preset confidence threshold, the system marks the feature point as abnormal. This process can efficiently screen out potential risk points and integrate these risk features into risk feature vectors with confidence, which is convenient for subsequent processing and decision-making.

[0086] Through the anomaly detection algorithm based on Mahalanobis distance, the system can effectively identify the abnormal state of the oil and gas pipeline network and provide timely feedback on potential risks. The implementation of this step greatly enhances the safety monitoring capability of the oil and gas pipeline network, allowing operators to take countermeasures in the first place and reduce the risk of accidents. The risk feature vector with confidence provides strong data support for subsequent decision-making, making the system more flexible and efficient in the complex oil and gas pipeline network environment.

[0087] In this step, the system will use the Mahalanobis distance-based anomaly detection algorithm to identify potential abnormal feature points based on the previously obtained probability feature distribution. First, the system will establish a data distribution model for normal operating conditions, which is usually achieved by statistically analyzing historical normal operating data. This distribution model will provide a reference standard for subsequent anomaly detection. The system will calculate the mean and covariance of each feature point under normal conditions to determine the clustering area of ​​normal feature points.

[0088] Next, the system calculates the Mahalanobis distance from each feature point to the normal operating condition distribution. Unlike the Euclidean distance, the Mahalanobis distance can take into account the correlation and covariance between features, making the distance metric more appropriate. During the calculation process, the system uses the Mahalanobis distance formula to calculate the Mahalanobis distance. Through such calculations, the system can effectively determine the degree of deviation of each feature point from the normal state.

[0089] Finally, the system will filter out abnormal feature points based on the preset confidence threshold. If the calculated Mahalanobis distance exceeds the threshold, the feature point is considered abnormal. After this process, the risk feature vector with confidence generated by the system will contain the feature points marked as abnormal and their related information. This not only provides data support for subsequent risk management, but also helps the operation team take timely measures to prevent potential safety accidents.

[0090] Exemplary: 1. Multilayer Perceptron (MLP) and Residual Connections Technical features: The spatiotemporal feature matrix is ​​input into a multi-layer perceptron for nonlinear transformation, and an activation function with residual connection is used to avoid gradient disappearance.

[0091] Input data: The dimension of the spatiotemporal feature matrix is ​​32 (for example, it contains features of 8 time steps each for pressure, temperature, flow, and vibration).

[0092] Network structure: input layer 32 dimensions → hidden layer 1 (64 dimensions, ReLU activation, residual connection) → hidden layer 2 (128 dimensions, ReLU activation, residual connection) → output layer (64-dimensional deep features).

[0093] Residual connection design: The output of each layer is H(x)=F(x)+x, where F(x) is the result after transformation of the current layer.

[0094] Weight initialization: He normal distribution (adapted to ReLU).

[0095] Batch Normalization: Batch normalization is added after each layer to accelerate convergence.

[0096] Actual scenario: The pressure sensor data of an oil pipeline fluctuates within a time window. MLP captures long-term dependencies through residual connections to avoid signal attenuation due to increasing number of layers.

[0097] 2. Gaussian Process Regression (GPR) and Kernel Function Combination Technical features: A linear combination of radial basis kernel (RBF) and periodic kernel is used to model nonlinearity and periodicity.

[0098] Kernel function definition: RBF kernel: k_RBF(x,x')=exp(-||x-x'||² / (2l²)), length scale l=0.5; periodic kernel: k_PER(x,x')=exp(-2sin²(π|x-x'| / p) / σ²), period p=24 (daily cycle), variance σ=1; combined kernel: k_combined=0.6*k_RBF+0.4*k_PER.

[0099] Training data: flow data of a natural gas pipeline (including daily peak periodic fluctuations).

[0100] Hyperparameter optimization: Optimize l, p, σ by maximum marginal likelihood estimation.

[0101] Output: The predicted mean flow rate of a node in the next three hours is 50 m³ / s, with a variance of 2.5 and a confidence interval of [45,55].

[0102] 3. Variational Inference Algorithms Quantify Uncertainty Technical features: The mean and variance of feature points are calculated through variational inference to generate confidence intervals.

[0103] Variational distribution setting: approximate posterior q(f)~N(μ,Σ), where μ is the mean vector and Σ is the diagonal covariance matrix.

[0104] Optimization objective: maximize the evidence lower bound (ELBO) using stochastic gradient descent (learning rate 1e-3).

[0105] Example result: The predicted mean of the feature point pressure is 10 MPa, the variance is 0.4, and the corresponding 95% confidence interval is [9.2, 10.8]. The temperature feature deviates from the confidence interval, triggering anomaly detection.

[0106] Application scenario: In liquefied natural gas pipelines, variational inference quantifies the measurement uncertainty of temperature sensors and assists in determining whether it is a normal fluctuation.

[0107] 4. Anomaly Detection Based on Mahalanobis Distance Technical features: Calculate the Mahalanobis distance between feature points and normal distribution to filter out anomalies.

[0108] 4.1. Normal working condition modeling The historical data mean μ=[10MPa,50℃,200m³ / h].

[0109] Covariance matrix Σ (calculated using 1000 sets of normal data): Real-time data point: x=[12MPa,60℃,180m³ / h].

[0110] 4.2. Mahalanobis distance calculation Threshold setting: The confidence level of 95% corresponds to the chi-square distribution threshold of 7.81 (3 degrees of freedom).

[0111] Result: 15.6>7.81, marked as high-risk abnormality.

[0112] S203, constructing a three-dimensional risk concentration distribution field according to the risk characteristic vector, and predicting the propagation path and diffusion trend of the risk concentration distribution field based on a graph neural network; In this step, the system first maps each risk feature point to the three-dimensional spatial coordinates of the pipeline network based on the risk feature vector generated previously and the topological structure information of the oil and gas pipeline network. This process involves mapping the nodes of the pipeline network (such as valves, pumping stations, etc.) to the three-dimensional coordinate axis so as to intuitively represent the risk level of each feature point in space. Subsequently, the Kriging interpolation algorithm is used to perform spatial interpolation on these discrete risk feature points to form a preliminary three-dimensional risk concentration distribution field. This distribution field reflects the risk level and distribution status of the oil and gas pipeline network at different spatial locations. In addition, in order to dynamically capture changes in risks, the system will also combine information from the time dimension and correct the risk concentration through the spatiotemporal Kriging interpolation algorithm to ensure that the generated distribution field can reflect the temporal evolution of the risk.

[0113] By constructing a three-dimensional risk concentration distribution field, the system can intuitively display the risk status of the oil and gas pipeline network, allowing operators to quickly identify potential high-risk areas during monitoring. The real-time updated risk concentration distribution field provides data support for subsequent risk management, allowing decision makers to monitor and respond more accurately when dealing with possible risk events. This process not only improves the accuracy of pollutant diffusion prediction, but also provides a strong basis for dynamically adjusting pipeline network operation strategies.

[0114] Specifically, each risk characteristic point can be mapped to the three-dimensional spatial coordinates of the pipeline network according to the risk characteristic vector and combined with the topological structure information of the oil and gas pipeline network, and the Kriging interpolation algorithm is used to perform spatial interpolation on the discrete risk characteristic points to generate a preliminary three-dimensional risk concentration distribution field; In this process, the system first needs to clarify the composition and topological structure of the oil and gas pipeline network. The topological structure usually includes the connection relationship of the pipeline and the coordinate information of the main nodes. The system will associate each risk feature point with these nodes to ensure that each feature point can find a corresponding position in three-dimensional space. With the help of the Kriging interpolation algorithm, these discrete risk feature points on the pipeline are integrated into a continuous risk concentration distribution field. Kriging interpolation is a statistically based spatial interpolation technology that can use the spatial relationship of known feature points to estimate the value of unknown points. For example, if the pressure of a section of the pipeline is abnormal, Kriging interpolation can use the information of the surrounding normal points to predict the risk concentration of this section of the pipeline, thereby generating a continuous risk distribution field.

[0115] The implementation of this step will make the oil and gas pipeline network spatially organized and enable visual monitoring of risks in different areas. This intuitive visual presentation helps operators respond quickly and promptly detect and deal with potential leaks or other safety hazards. In addition, the generated three-dimensional risk concentration distribution field provides basic data for subsequent risk prediction and management, which will help enhance the safety and stability of the pipeline network.

[0116] In this step, the system first needs to obtain the topological structure information of the oil and gas pipeline network, including the three-dimensional spatial coordinates of key nodes such as pipelines, valves and pump stations. This information usually comes from the design drawings or digital models of the pipeline network. The system associates each risk feature point (for example, abnormal pressure and flow in a certain section of the pipeline) with the corresponding three-dimensional spatial coordinates. In order to ensure that these feature points can be effectively displayed in three-dimensional space, the system will use identifiers to match the feature points with the actual pipeline network locations one by one.

[0117] Next, the system uses the Kriging interpolation algorithm to perform spatial interpolation on these discrete risk feature points. Kriging interpolation is an effective statistical interpolation method that can rely on the distribution of known points to predict the values ​​of unknown points. Specifically, the system will construct a covariance matrix based on the risk values ​​of the feature points and their spatial relationships to characterize the relationships between the feature points. In this way, the system can construct a continuous risk concentration distribution field in the three-dimensional space of the pipeline network. For example, if the pressure value of a certain section of the pipeline is significantly higher than that of other parts, the Kriging interpolation method can use the risk value of that point to infer the risk concentration of the surrounding area, thereby forming a smooth risk concentration map.

[0118] Finally, after Kriging interpolation processing, the system will obtain a preliminary three-dimensional risk concentration distribution field. This distribution field will provide basic data for subsequent dynamic analysis and decision-making, allowing pipeline network operators to intuitively identify high-risk areas and take quick action to ensure the safety and stability of the pipeline network.

[0119] Based on the preliminary three-dimensional risk concentration distribution field, combined with the time dimension information, the spatiotemporal Kriging interpolation algorithm is used, combined with the historical risk propagation data, to dynamically correct the risk concentration distribution field and obtain the three-dimensional risk concentration distribution field that changes with time; In this step, the system will combine historical risk propagation data and introduce the dynamic time dimension into the generation process of the risk concentration distribution field. First, the system will extract the risk propagation trend in the historical data, such as the evolution of risks under specific conditions in the past few months. These data will provide valuable reference for the system. Next, the system will apply the spatiotemporal Kriging interpolation algorithm, which can introduce time factors on the basis of spatial interpolation and dynamically correct the risk concentration distribution field. By analyzing the time series, the system can capture how the risk concentration changes over time. For example, assuming that a section of the pipeline is at risk due to abnormal temperature in a certain period of time in the past, then in future risk assessments, the system will adjust the current risk distribution according to the changing pattern of historical data.

[0120] By introducing the time dimension, the system can generate a risk concentration distribution field with strong timeliness, which ensures that risks that change over time can be captured and updated in a timely manner. This dynamic feature not only improves the accuracy of risk monitoring, but also provides a more reliable data basis for subsequent risk assessment and decision-making. Operators can take proactive measures to reduce potential risks and ensure the safe and stable operation of the oil and gas pipeline network by analyzing the changing trends in the time dimension.

[0121] In this step, the system combines the preliminary three-dimensional risk concentration distribution field with the information of the time dimension for dynamic correction. First, the system extracts historical risk propagation data, including records of risk events in the past few months or years. These data can include changes in risk characteristics of each node, such as leakage events and flow fluctuations in a certain section of pipeline at a specific time point. These historical data will be used to build a time series model to analyze the risk propagation trend in the pipeline network.

[0122] Next, the system will use the space-time Kriging interpolation algorithm to achieve this dynamic correction. The space-time Kriging interpolation algorithm is different from the traditional spatial interpolation method. It takes into account the influence of time and can better reflect the changing process of risk concentration. The system will input historical risk propagation data, establish a time series model, analyze the change pattern of risk over time, and calculate the risk concentration in a certain time period in the future under the current pipeline network status. For example, if historical data shows that a certain section of pipeline is prone to leakage in the rainy season, the system will take this seasonal trend into account and adjust the current risk concentration forecast accordingly.

[0123] Finally, after dynamic correction, the system will obtain a three-dimensional risk concentration distribution field that changes with time. This dynamic risk concentration distribution field can provide more accurate information support for the management and maintenance of oil and gas pipeline networks, allowing operators to take timely preventive measures based on the latest risk status and reduce the probability of accidents.

[0124] The three-dimensional risk concentration distribution field that changes with time is input into the graph neural network, and a graph model is constructed according to the topological structure of the pipeline network. The nodes of the graph model represent the key positions of the pipeline network, and the edges of the graph model represent the connection relationship of the pipeline sections. The graph attention mechanism is used to capture the risk propagation dependency between nodes to obtain the initial prediction result of risk propagation. In this step, the system first inputs the time-varying three-dimensional risk concentration distribution field into the graph neural network (GNN) to establish a graph model that reflects the topological structure of the oil and gas pipeline network. Specifically, the nodes in the graph model represent the key locations of the pipeline network (such as pumping stations, valves, etc.), while the edges represent the connection relationship of the pipe sections. With this representation, the system can better understand the interrelationships and dependencies within the pipeline network. Next, the system will use the graph attention mechanism to capture the risk propagation dependencies between different nodes. The graph attention mechanism can dynamically adjust the influence weight of each node with its neighboring nodes by learning the importance of each node. For example, assuming that the risk level of a node is high, the system will pay more attention to other nodes connected to the node to ensure that important information is not missed in the risk propagation analysis.

[0125] By combining the graph neural network with the topological structure of the pipeline network, the system can mine the mutual influence between each node and its corresponding risk propagation path. This capability enables decision makers to more clearly identify the direction and potential impact of risk propagation, enhancing the scientificity and accuracy of real-time monitoring and decision-making. In addition, the initial prediction results of risk propagation will provide a basis for subsequent risk management measures, allowing the management team to respond in a timely manner and prevent accidents.

[0126] In this step, the system inputs the time-varying three-dimensional risk concentration distribution field into the graph neural network (GNN) to build a graph model. First, the system needs to clarify the topological structure of the oil and gas pipeline network and define the nodes and edges of the graph model. Nodes generally represent key locations in the pipeline network, such as pumping stations, valves, and monitoring points, while edges represent the connection relationship between these nodes. The system will combine the risk feature values ​​with the nodes to build a graph model that represents the risk status.

[0127] Next, the system uses the graph attention mechanism to analyze and capture the risk propagation dependencies between different nodes. The graph attention mechanism is a dynamic weighted model that can assign different importance weights to each node so that more attention can be paid to nodes that have a greater impact on the final result in the risk propagation analysis. For example, if a valve has a high risk value and the valve is connected to multiple pipe sections, the system will give a higher weight to the nodes connected to the valve to ensure that their impact on the risk propagation analysis is fully considered.

[0128] Finally, after being processed by the graph neural network, the system will obtain the initial prediction results of risk propagation. These results can not only reveal the distribution characteristics of the current risks in the pipeline network, but also reflect the possible propagation paths and affected areas, providing a reference for subsequent risk countermeasures.

[0129] According to the initial prediction results of risk propagation, a sequence prediction model based on time convolutional network is adopted, combined with historical risk propagation path data, to predict the risk propagation path and diffusion trend in future time steps, and generate the propagation prediction results of dynamic risk concentration distribution field.

[0130] In this step, the system uses the temporal convolutional network (TCN) to further analyze and predict the preliminary prediction results of risk propagation. The temporal convolutional network is a model that specializes in processing time series data. It captures the time series characteristics in the data through convolution operations. Based on the results of the preliminary prediction, the system will input the historical risk propagation path data, combined with the advantages of the temporal convolutional network, to generate risk propagation predictions for future time steps. In this process, the model will be trained based on historical data to learn the propagation pattern and diffusion trend of risks. For example, by analyzing the changes in risks over a period of time in the past, the system can determine how risks will propagate in the pipeline network under certain conditions.

[0131] The implementation of this process enables the system to predict future risk states, significantly improving managers' ability to respond to emergencies. By generating propagation prediction results of the dynamic risk concentration distribution field, the system can identify high-risk areas in advance, giving decision makers sufficient time to take countermeasures. In addition, such predictions not only reduce the probability of accidents, but also improve the overall safety of the oil and gas pipeline network and ensure the stability of energy supply.

[0132] In this step, the system will use a sequence prediction model based on a temporal convolutional network (TCN), using the previously obtained initial risk propagation prediction results and historical risk propagation path data. First, the system needs to organize the historical risk propagation data to provide rich training information for the TCN model. These historical data may contain risk concentration values ​​at different time points and their corresponding propagation paths, such as the risk distribution of a certain section of pipeline in a certain period of time.

[0133] Next, the system will pass this data into the temporal convolutional network. TCN uses convolutional layers to process time series data, which can effectively capture features in the time dimension. When inputting data, the system will take into account the temporal nature of risk propagation and train through multiple time windows to ensure that the model can learn the potential propagation patterns in historical data. For example, if the system identifies that a risk spreads rapidly under certain conditions, TCN will remember this pattern and apply it when predicting future risks.

[0134] Finally, the system outputs the risk propagation path and diffusion trend of the future time step based on TCN prediction. In this way, the system can not only provide a prediction of the future development of risks, but also form a propagation prediction result of a dynamic risk concentration distribution field. These prediction results will provide the necessary data support for the operation and management of the oil and gas pipeline network, enabling decision makers to proactively respond to possible risks, optimize operation strategies, and ensure the safe and efficient operation of the pipeline network.

[0135] Exemplary: 1. Data collection and generation of spatiotemporal feature matrix 1.1 Multi-source sensor data Pressure sensor: P1 (50MPa), P2 (48MPa), P3 (52MPa); Temperature sensor: T1 (25°C), T2 (28°C); Flow meter: F1 (100m³ / s), F2 (105m³ / s); Vibration sensor: V1 (0.5mm / s²), V2 (0.6mm / s²).

[0136] 1.2 Dynamic Time Warping (DTW) Alignment Assuming that the sampling frequency of the pressure sensor is 1 Hz and that of the flow meter is 0.5 Hz, the DTW algorithm stretches the time axis of the flow data to align it with the time axis of the pressure data to ensure time synchronization.

[0137] 1.3. Spatiotemporal Convolutional Neural Network (ST-CNN) Spatial dimension: Graph convolutional network (GCN) processes pipeline network topology (such as pipeline connection relationship: node A→node B→node C).

[0138] Time dimension: One-dimensional convolution kernel (window size = 5) extracts the periodic fluctuation characteristics of traffic data.

[0139] Output: spatiotemporal feature matrix (size: 10×10×4, including four channels of pressure, temperature, flow, and vibration).

[0140] 2. Deep Gaussian Process Model and Risk Feature Vector 2.1. Multilayer Perceptron (MLP) The input spatiotemporal feature matrix is ​​mapped to a high-dimensional space through a three-layer residual network (dimension of each layer: 256→512→256), and the deep feature representation (dimension: 128) is output.

[0141] 2.2 Gaussian Process Regression (GPR) Kernel function combination: RBF kernel (length scale = 1.0) + periodic kernel (period = 24 hours).

[0142] Prediction results: The mean pressure of a certain pipe section is 51MPa, and the variance is 0.5 (confidence interval: 50.2~51.8MPa).

[0143] 2.3. Mahalanobis distance anomaly detection Calculate the distance D = 2.1 between the current pressure characteristic and the normal operating condition (mean 50 MPa, covariance matrix Σ).

[0144] If the threshold = 2.0, it is judged as abnormal and a risk feature vector [risk type = high pressure, confidence = 90%] is generated.

[0145] 3. Construction of three-dimensional risk concentration distribution field 3.1 Kriging interpolation Discrete risk point coordinates: Node A (1, 2, 0) and Node B (2, 3, 0), with risk values ​​of 0.8 and 0.5 respectively.

[0146] After interpolation, a continuous field is generated, such as the risk concentration = 0.65 at the coordinates (1.5, 2.5, 0).

[0147] 3.2. Space-time Kriging Correction Combined with historical data (such as a 20% increase in risk at point A during last year’s rainy season), the current risk concentration is dynamically adjusted to 0.78.

[0148] 4. Graph Neural Networks and Risk Propagation Prediction 4.1 Graph Model Construction Node: {Pump station 1, valve 2, monitoring point 3} Edge: {(Pumping station 1→Valve 2, length = 5km), (Valve 2→Monitoring point 3, length = 3km)} 4.2 Graph Attention Mechanism (GAT) The risk weight of pump station 1 = 0.7, valve 2 = 0.3, and the predicted risk transmission path is: pump station 1 → valve 2 → monitoring point 3.

[0149] 4.3 Temporal Convolutional Network (TCN) Input historical risk path data (such as the risk diffusion speed in the past 24 hours = 0.5km / h), and predict that the risk will cover monitoring point 3 in the next 6 hours.

[0150] S204, based on the prediction results of the risk concentration distribution field, an adaptive warning threshold generation model based on the generative adversarial network is constructed. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level warning threshold that changes with the operation status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the warning threshold, the risk perception warning of the operation status of the oil and gas pipeline network is realized.

[0151] In this step, the system realizes dynamic monitoring and risk perception of the operating status of the oil and gas pipeline network by constructing an adaptive warning threshold generation model based on the generative adversarial network (GAN). The model mainly consists of two parts: a normal operating condition generator and a real-time discriminator. The task of the generator is to generate simulation data that conforms to the normal state of the pipeline network based on historical normal operating data. These data not only need to reflect the typical characteristics of the pipeline network, but also need to have a certain degree of diversity and authenticity in order to better simulate the real operating environment. The real-time discriminator is responsible for comparing the generated normal operating condition data with the predicted results of the risk concentration distribution field, learning to identify the boundary characteristics of normal and abnormal states, thereby improving the accuracy of the model. This process continuously optimizes the two parts through dynamic game, making the warning threshold more targeted and effective.

[0152] The early warning thresholds generated based on this dynamic game mechanism can timely reflect the real-time operation status of the pipeline network, allowing the risk management team to quickly identify potential risk areas and take corresponding preventive measures. Specifically, as the operating status of the oil and gas pipeline network changes, the system will generate corresponding low-risk, medium-risk and high-risk thresholds. These thresholds will serve as important reference standards for the monitoring system to help decision makers make scientific judgments and decisions at critical moments. In this way, the safety and stability of the oil and gas pipeline network can be further improved, thereby effectively reducing the probability of accidents and ensuring smooth energy transportation.

[0153] Specifically, according to the historical normal operation data, a normal operating condition generator based on a generative adversarial network can be trained to generate simulated data that meets the normal operation characteristics of the pipeline network; the generator uses a variational autoencoder with conditional constraints to ensure the diversity and authenticity of the generated data; In this step, the system will use historical normal operation data to train the generator in order to generate simulated data that conforms to the normal state of the oil and gas pipeline network. The design of the generator is based on the variational autoencoder (VAE) architecture. By introducing conditional constraints, the system can ensure that the generated data is not only diverse, but also truly reflects the operating characteristics of the pipeline network. The key to this method is to encode historical data to obtain the distribution of the latent space to generate new samples. Specifically, the system will extract various features of the samples, including pressure, flow, temperature, and vibration. Through structured data processing methods, the system can form a representative data set for the generator to learn.

[0154] By generating simulated data that matches historical data, the system can monitor the normal status of the oil and gas pipeline network even when real-time data is lacking. This powerful generation capability enables managers to fully consider various changes under normal operation when evaluating the safety of the pipeline network, providing a scientific basis for subsequent risk assessment. In addition, the generated diverse data helps identify potential abnormal behaviors and provides more comprehensive information support for the risk monitoring system.

[0155] In this step, the system first collects historical normal operation data, including information such as pressure, flow, temperature, and vibration. These data come from various monitoring sensors in the oil and gas pipeline network, such as pipeline pressure values ​​recorded by pressure sensors, flow data monitored by flow meters, and real-time data provided by temperature and vibration sensors. During the data collection phase, the system needs to preprocess the raw data, such as removing outliers and filling missing values, to ensure the quality and accuracy of the data. These processed data will become training samples for the generator.

[0156] Next, a normal operating condition generator based on a generative adversarial network (GAN) is built. The generator uses a variational autoencoder (CVAE) architecture with conditional constraints, which ensures that the generated data is both diverse and consistent with the actual operating characteristics. The generator uses the learned features together with the conditional input to generate new samples. For example, if the input condition is "normal operating pressure", the generator will generate diversified pressure values ​​and corresponding flow, temperature and other information under this condition based on historical data. During the model training process, the generator will be continuously optimized to reduce the gap between the generated data and the real data.

[0157] Finally, by comparing with the real data, the system will continuously adjust the parameters of the generator so that it can output simulated data that is more in line with the actual situation. This process will be achieved by using adversarial loss and reconstruction loss to ensure that the generated data is not only statistically similar to the real data, but also physically usable. The trained generator can generate a large amount of simulated data representing normal working conditions, which will provide a reliable basis for risk prediction and early warning strategies.

[0158] Construct a real-time discriminator based on a deep residual network, input the prediction results of the risk concentration distribution field and the normal operating condition data generated by the generator, learn the boundary characteristics of normal and abnormal states through dynamic game, and output the discrimination results and their confidence; In this step, the system builds a real-time discriminator, which uses the architecture of a deep residual network to improve the recognition ability of the model. The main function of the discriminator is to compare the predicted results of the input risk concentration distribution field with the normal operating data generated by the generator, and learn and distinguish what is a normal state and what is an abnormal state. In the dynamic game process, the discriminator will continuously optimize its performance and improve the accuracy of the input data. Specifically, the system will train the discriminator to identify the boundary features between normal data and abnormal risk features by comparing them, ensuring that the model can make effective risk judgments in practical applications.

[0159] The establishment of a real-time discriminator will greatly enhance the risk monitoring capabilities of the oil and gas pipeline network. By accurately identifying normal and abnormal conditions, the system can warn of potential risks in advance and provide key decision-making support for managers. This function not only improves the scientific nature of risk management, but also effectively reduces safety accidents caused by misjudgment, ensuring the smooth operation, safety and stability of the pipeline network.

[0160] In this step, the system needs to build a real-time discriminator based on a deep residual network (ResNet). The main task of the discriminator is to distinguish the normal operating data generated by the generator from the real-time risk concentration distribution field prediction results. First, the system will design the network structure so that it consists of multiple residual blocks to ensure the effective transmission of information in the deep network and avoid the problem of gradient disappearance. The discriminator will input the normal operating samples generated by the generator and the real-time risk data, extract the key features through multi-layer nonlinear transformation, and establish the boundary between normal and abnormal states.

[0161] During the training process, the discriminator will engage in a dynamic game, that is, it will continuously update itself while competing with the generator. If the data generated by the generator is realistic, the discriminator will have difficulty distinguishing it from real data; conversely, if the data generated by the generator is not realistic enough, the discriminator will be able to easily distinguish it. Through this adversarial training, the discriminator can more accurately learn the characteristic differences between normal and abnormal states. For example, when real-time sensor data fluctuates abnormally, the discriminator can promptly identify the state and mark it as abnormal.

[0162] Finally, after multiple rounds of training, the discriminator will output the discrimination results and confidence level of each input data. The higher the confidence level, the more certain the model is that the sample is normal or abnormal. In this way, the system can monitor the operation status of the pipeline network in real time and detect potential risks in a timely manner.

[0163] According to the discrimination results, an adaptive clustering algorithm is used to divide the discrimination results into multiple levels, and combined with the dynamic changes of the risk concentration distribution field, a multi-level warning threshold that changes with the operation status of the pipeline network is generated. The multi-level warning threshold includes low risk, medium risk and high risk thresholds; In this step, the system will rely on the results of the discriminator output and apply the adaptive clustering algorithm to divide the risk status into multiple levels. The advantage of the adaptive clustering algorithm is that it can automatically determine the number and category of clusters according to the distribution of the current data, making the classification results more reasonable and effective. Combined with the dynamic changes of the risk concentration distribution field, the system will generate multi-level warning thresholds, including low risk, medium risk and high risk thresholds. These thresholds not only reflect the risk characteristics of the current pipeline network operation status, but also provide quantitative reference standards for real-time monitoring.

[0164] Through the cluster-based multi-level warning threshold setting, the system can achieve a detailed division of the operating status risks of the oil and gas pipeline network, which not only improves the response speed of the warning system, but also enables managers to arrange monitoring and emergency measures in a targeted manner. The multi-level warning mechanism ensures that decision makers can take the initiative to respond under different risk conditions, thereby improving the safety management level of the oil and gas pipeline network. At the same time, this method reduces the subjectivity of human judgment and provides support for scientific decision-making.

[0165] In this step, the system will apply the adaptive clustering algorithm to divide the discrimination results into multiple levels according to the results output by the discriminator. The key feature of the adaptive clustering algorithm is that it can automatically determine the number and category of clusters based on the distribution of data. The system will first extract abnormal and normal sample data from the discrimination results and sort them according to their confidence scores. Next, the system will extract features from these data, including their temporal features, spatial features, etc., in order to conduct more effective clustering analysis.

[0166] By introducing an adaptive clustering algorithm, the system can classify data into three main categories: low risk, medium risk, and high risk based on the discrimination results. For example, if the confidence score of the discrimination result is above a certain threshold, the sample can be classified as low risk, the one in the middle range is medium risk, and the one below the threshold is high risk. This classification will help managers take corresponding response measures in a timely manner under different conditions, thereby optimizing risk management strategies.

[0167] Finally, the system integrates the generated multi-level warning thresholds into a dynamic update system so that the warning thresholds at all levels can be adjusted in real time when the operating status of the oil and gas pipeline network changes. Combined with the dynamic changes in the risk concentration distribution field, the system can ensure the effectiveness and accuracy of risk monitoring, thereby better responding to potential risks.

[0168] The real-time risk characteristics are compared with the multi-level warning thresholds, and a fuzzy logic-based decision model is used to dynamically adjust the warning level according to the confidence level of the risk characteristics and the degree of threshold deviation, thereby realizing risk-aware warning of the operating status of the oil and gas pipeline network.

[0169] In this step, the system compares the real-time risk characteristics with the generated multi-level warning thresholds, and dynamically adjusts the warning level in combination with the fuzzy logic-based decision model. The fuzzy logic decision model can handle uncertainty and ambiguity. Based on the analysis of the confidence of the risk characteristics and the degree of deviation from the warning threshold, the system will generate a more accurate warning level. For example, if the real-time risk characteristics are significantly higher than the high-risk threshold, the system will issue a high-risk warning; if the risk characteristics are in the low-risk range, the system may remain normal.

[0170] Dynamic adjustment decisions based on fuzzy logic can significantly improve the risk management capabilities of oil and gas pipeline networks. This mechanism not only ensures the efficiency of real-time monitoring, but also helps managers respond in a timely manner in a complex and changing operating environment, reducing the probability of risk accidents. Through accurate risk perception and early warning, the system improves the safe operation level of the pipeline network and creates conditions for ensuring the stability of energy supply.

[0171] In this step, the system compares the risk characteristics collected in real time with the generated multi-level warning thresholds, and uses a fuzzy logic-based decision model to dynamically adjust the warning level. First, the system integrates and compares the real-time risk characteristic data with multiple warning thresholds to assess the risk level of the current pipeline network. For example, when the real-time monitoring shows that the pressure value of a certain section of the pipeline exceeds the medium risk threshold, the system will mark the section of the pipeline as a high-risk area.

[0172] Next, the fuzzy logic model considers the confidence of the risk feature and the degree of deviation from the warning threshold for calculation. By establishing fuzzy rules, such as "if the risk feature is above the medium risk threshold and the confidence is high, the warning level is set to high risk", the system can flexibly respond to different scenarios. This flexibility is a significant advantage of the fuzzy logic model because it can handle uncertainty and ambiguity, making decisions more scientific and reasonable.

[0173] Finally, based on the calculation results, the system will dynamically adjust the warning level and output the corresponding warning information. This information will be presented to the pipeline monitoring personnel through a visual interface to help them make decisions and take measures in a timely manner, such as adding monitoring personnel, reducing pressure, and performing equipment maintenance. This risk perception and warning mechanism will significantly improve the safety of the oil and gas pipeline network, reduce the risk of accidents, and ensure the stable operation of the pipeline network.

[0174] 1. Historical data collection and preprocessing 1.1. Source of sample data Pressure data: Pressure sensor (model PTX5000) of a natural gas pipeline in Xi’an, sampling frequency 1Hz, normal range 4-6MPa.

[0175] Flow data: turbine flowmeter (model TUF-2000), data interval 5 seconds, normal flow range 20-30m³ / s.

[0176] Temperature data: infrared temperature sensor (model IR-TC300), sampling frequency 0.5Hz, normal temperature range 10-50℃.

[0177] Vibration data: accelerometer (model ADXL345), sampling frequency 100Hz, normal vibration amplitude <0.5g.

[0178] 1.2 Preprocessing steps Missing value filling: Cubic spline interpolation is used to fill the missing 5-minute data in the temperature data.

[0179] Outlier elimination: Use the 3σ principle to eliminate abnormal points in the pressure data that exceed ±3 standard deviations (such as a sudden 8MPa peak).

[0180] 2. Normal Condition Generator (CVAE) Training 2.1 Model Architecture Encoder: 3-layer fully connected network (256 nodes in the input layer and 128 nodes in the hidden layer), outputting the mean μ and variance σ of the latent variables.

[0181] Decoder: 3-layer deconvolution network, input condition is "flow rate = 25m³ / s±5%", generates corresponding pressure, temperature and vibration data.

[0182] Loss function: ELBO (evidence lower bound) loss + KL divergence constrained latent space distribution.

[0183] 2.2. Generate data examples Enter the condition "pressure = 5MPa" to generate 10 sets of simulation data: Flow rate: 24.8m³ / s, 25.1m³ / s,...; Temperature: 32℃, 31.5℃,...; Vibration: 0.3g, 0.35g, ...

[0184] 3. Real-time discriminator (ResNet) construction 3.1 Network Structure Residual block: 4 residual modules, each module contains 2 convolutional layers (kernel size 3×3) + skip connection.

[0185] Input: Risk concentration distribution field (64×64 grid) + simulation data output by the generator.

[0186] Output: judgment result (0 / 1) and confidence level (0-1, such as 0.92 means "normal").

[0187] 3.2 Dynamic Game Example The generator generates a set of "pressure = 7MPa (abnormal)" data, and the discriminator judges it as "abnormal" with a confidence level of 0.85.

[0188] After adjusting the parameters, the generator generated the data "pressure = 5.2MPa", and the discriminator confidence dropped to 0.55 (close to misjudgment).

[0189] 4. Adaptive clustering (DBSCAN algorithm) 4.1 Parameter settings Neighborhood radius ε=0.5, minimum number of samples min_samples=10.

[0190] Input: 500 sets of data confidence output by the discriminator (such as [0.1, 0.3, 0.8, ...]).

[0191] 4.2 Clustering Results Low risk: Confidence level > 0.7 (such as 0.8, 0.9), corresponding pressure < 5.5MPa.

[0192] Medium risk: 0.4≤confidence≤0.7 (such as 0.6), corresponding to pressure 5.5-6.5MPa.

[0193] High risk: Confidence level <0.4 (such as 0.2), corresponding to pressure >6.5MPa or vibration >1g.

[0194] 5. Fuzzy logic decision model 5.1 Rule Base Example IF pressure deviation threshold>10%AND confidence>0.8THEN high risk warning.

[0195] IF flow fluctuation is ±5% AND temperature is normal THEN medium risk warning.

[0196] 5.2. Dynamic Adjustment Case Real-time data: Pressure = 6.8MPa (13% above threshold), confidence level 0.75.

[0197] Fuzzy output: Risk value = 0.82 (high risk), triggering a red alarm and activating the pressure relief valve.

[0198] 6. Visualization and warning output 6.1 System interface example Three-dimensional risk field: In the three-dimensional model of the pipeline, high-risk sections are displayed in red (such as the Xi'an section) and low-risk sections are displayed in green.

[0199] 6.2. Warning log [2023-10-01 14:30] Warning ID: #2023-001; Location: Xi'an section KP32+500; Risk level: High risk (abnormal pressure); Recommended Action: Immediately repair valve No. 3.

[0200] It can be seen that according to the real-time operation data of the oil and gas pipeline network, the multi-dimensional characteristics of the pipeline network operation status are fused to generate a spatiotemporal feature matrix; the spatiotemporal feature matrix is ​​input into the deep Gaussian process model, the abnormal characteristics in the pipeline network operation status are probabilistically modeled, and the risk feature vector with confidence is output; according to the risk feature vector, a three-dimensional risk concentration distribution field is constructed, and the propagation path and diffusion trend of the risk concentration distribution field are predicted; according to the prediction results, an adaptive early warning threshold generation model based on the generative adversarial network is constructed to generate multi-level early warning thresholds that change with the operation status of the pipeline network, and according to the comparison results of the real-time risk characteristics and the early warning threshold, the risk perception early warning of the operation status of the oil and gas pipeline network is realized, thereby realizing dynamic risk assessment and adaptive early warning, and enhancing the safety and emergency response capabilities of the pipeline network.

[0201] Another embodiment of the present invention provides an oil and gas pipeline network operation status risk perception and early warning system, see Figure 3 , the system may include: The fusion module 301 is used to fuse the multi-dimensional features of the operation status of the pipeline network according to the real-time operation data of the oil and gas pipeline network, and generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; Modeling module 302, used to input the spatiotemporal feature matrix into the deep Gaussian process model, perform probabilistic modeling on the abnormal features in the operation status of the pipeline network, and output a risk feature vector with confidence; A construction module 303 is used to construct a three-dimensional risk concentration distribution field according to the risk characteristic vector, and predict the propagation path and diffusion trend of the risk concentration distribution field based on a graph neural network; The early warning module 304 is used to construct an adaptive early warning threshold generation model based on a generative adversarial network according to the prediction results of the risk concentration distribution field. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level early warning threshold that changes with the operation status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the early warning threshold, the risk perception early warning of the operation status of the oil and gas pipeline network is realized.

[0202] It can be seen that according to the real-time operation data of the oil and gas pipeline network, the multi-dimensional characteristics of the pipeline network operation status are fused to generate a spatiotemporal feature matrix; the spatiotemporal feature matrix is ​​input into the deep Gaussian process model, the abnormal characteristics in the pipeline network operation status are probabilistically modeled, and the risk feature vector with confidence is output; according to the risk feature vector, a three-dimensional risk concentration distribution field is constructed, and the propagation path and diffusion trend of the risk concentration distribution field are predicted; according to the prediction results, an adaptive early warning threshold generation model based on the generative adversarial network is constructed to generate multi-level early warning thresholds that change with the operation status of the pipeline network, and according to the comparison results of the real-time risk characteristics and the early warning threshold, the risk perception early warning of the operation status of the oil and gas pipeline network is realized, thereby realizing dynamic risk assessment and adaptive early warning, and enhancing the safety and emergency response capabilities of the pipeline network.

[0203] An embodiment of the present invention further provides a storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0204] Specifically, in this embodiment, the above storage medium may be configured to store a computer program for performing the following steps: S201, based on the real-time operation data of the oil and gas pipeline network, the multi-dimensional features of the pipeline network operation status are integrated to generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; S202, inputting the spatiotemporal feature matrix into the deep Gaussian process model, probabilistically modeling the abnormal features in the operation status of the pipeline network, and outputting a risk feature vector with confidence; S203, constructing a three-dimensional risk concentration distribution field according to the risk characteristic vector, and predicting the propagation path and diffusion trend of the risk concentration distribution field based on a graph neural network; S204, based on the prediction results of the risk concentration distribution field, an adaptive warning threshold generation model based on the generative adversarial network is constructed. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level warning threshold that changes with the operation status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the warning threshold, the risk perception warning of the operation status of the oil and gas pipeline network is realized.

[0205] It can be seen that according to the real-time operation data of the oil and gas pipeline network, the multi-dimensional characteristics of the pipeline network operation status are fused to generate a spatiotemporal feature matrix; the spatiotemporal feature matrix is ​​input into the deep Gaussian process model, the abnormal characteristics in the pipeline network operation status are probabilistically modeled, and the risk feature vector with confidence is output; according to the risk feature vector, a three-dimensional risk concentration distribution field is constructed, and the propagation path and diffusion trend of the risk concentration distribution field are predicted; according to the prediction results, an adaptive early warning threshold generation model based on the generative adversarial network is constructed to generate multi-level early warning thresholds that change with the operation status of the pipeline network, and according to the comparison results of the real-time risk characteristics and the early warning threshold, the risk perception early warning of the operation status of the oil and gas pipeline network is realized, thereby realizing dynamic risk assessment and adaptive early warning, and enhancing the safety and emergency response capabilities of the pipeline network.

[0206] An embodiment of the present invention further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0207] Specifically, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0208] Specifically, in this embodiment, the processor may be configured to perform the following steps through a computer program: S201, based on the real-time operation data of the oil and gas pipeline network, the multi-dimensional features of the pipeline network operation status are integrated to generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; S202, inputting the spatiotemporal feature matrix into the deep Gaussian process model, probabilistically modeling the abnormal features in the operation status of the pipeline network, and outputting a risk feature vector with confidence; S203, constructing a three-dimensional risk concentration distribution field according to the risk characteristic vector, and predicting the propagation path and diffusion trend of the risk concentration distribution field based on a graph neural network; S204, based on the prediction results of the risk concentration distribution field, an adaptive warning threshold generation model based on the generative adversarial network is constructed. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level warning threshold that changes with the operation status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the warning threshold, the risk perception warning of the operation status of the oil and gas pipeline network is realized.

[0209] It can be seen that according to the real-time operation data of the oil and gas pipeline network, the multi-dimensional characteristics of the pipeline network operation status are fused to generate a spatiotemporal feature matrix; the spatiotemporal feature matrix is ​​input into the deep Gaussian process model, the abnormal characteristics in the pipeline network operation status are probabilistically modeled, and the risk feature vector with confidence is output; according to the risk feature vector, a three-dimensional risk concentration distribution field is constructed, and the propagation path and diffusion trend of the risk concentration distribution field are predicted; according to the prediction results, an adaptive early warning threshold generation model based on the generative adversarial network is constructed to generate multi-level early warning thresholds that change with the operation status of the pipeline network, and according to the comparison results of the real-time risk characteristics and the early warning threshold, the risk perception early warning of the operation status of the oil and gas pipeline network is realized, thereby realizing dynamic risk assessment and adaptive early warning, and enhancing the safety and emergency response capabilities of the pipeline network.

[0210] The above describes in detail the structure, features and effects of the present invention based on the embodiments shown in the drawings. The above is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the drawings. Any changes made according to the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which still do not exceed the spirit covered by the description and drawings, should be within the protection scope of the present invention.

Claims

1. A method for early warning of oil and gas pipeline network operation status risk perception, characterized in that: include: Based on the real-time operation data of the oil and gas pipeline network, the multi-dimensional characteristics of the pipeline network operation status are integrated to generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; The spatiotemporal feature matrix is ​​input into the deep Gaussian process model to perform probabilistic modeling on the abnormal features in the operation status of the pipeline network and output a risk feature vector with confidence. According to the risk characteristic vector, a three-dimensional risk concentration distribution field is constructed, and the propagation path and diffusion trend of the risk concentration distribution field are predicted based on the graph neural network; According to the prediction results of the risk concentration distribution field, an adaptive warning threshold generation model based on the generative adversarial network is constructed. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level warning threshold that changes with the operating status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the warning threshold, the risk-aware warning of the operating status of the oil and gas pipeline network is realized.

2. The method according to claim 1, characterized in that According to the real-time operation data of the oil and gas pipeline network, the multi-dimensional features of the pipeline network operation status are integrated to generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information, including: According to the real-time operation data collected by multi-source sensors of the oil and gas pipeline network, timestamp alignment and missing value filling processing are performed, and a multi-source data alignment algorithm based on dynamic time warping is used to eliminate the difference in sensor sampling frequency and obtain a time-synchronized multi-source data sequence, wherein the real-time operation data includes pressure, temperature, flow and vibration data; The multi-source data sequence is input into a spatiotemporal convolutional neural network, wherein a graph convolutional network is used in the spatial dimension to capture the topological structure characteristics of the pipeline network, and a one-dimensional convolutional kernel is used in the temporal dimension to extract the time series characteristics, thereby obtaining a preliminary spatiotemporal feature representation; For the extracted spatiotemporal feature representation, the multi-head self-attention mechanism is used to calculate the correlation weights between different sensor data, and the features are weighted fused according to the weights to highlight the contribution of key sensor data to risk perception, thus obtaining the weighted spatiotemporal feature representation; The weighted spatiotemporal features are input into the sparse autoencoder, and the dimension is reduced through nonlinear mapping to remove redundant information, retain key features, and generate a low-dimensional spatiotemporal feature matrix containing pressure, temperature, flow and vibration information.

3. The method according to claim 2, characterized in that The spatiotemporal feature matrix is ​​input into the deep Gaussian process model, the abnormal features in the operation status of the pipeline network are probabilistically modeled, and the risk feature vector with confidence is output, including: Input the spatiotemporal feature matrix into a multi-layer perceptron, and map the original features to a high-dimensional latent space through multi-layer nonlinear transformation to obtain a deep feature representation, wherein each layer uses an activation function with a residual connection to avoid the gradient vanishing problem and enhance the feature expression capability; For the deep feature representation, a probability model based on Gaussian process regression is constructed, and a linear combination of radial basis kernel function and periodic kernel function is used to capture the nonlinear relationship and periodic change law in the operation state of the pipeline network, and obtain a preliminary probabilistic feature representation; According to the probabilistic feature representation, the predicted mean and variance of each feature point are calculated, and the uncertainty of the model is quantified through a variational inference algorithm to obtain a probabilistic feature distribution with a confidence interval; For the probability feature distribution, an anomaly detection algorithm based on Mahalanobis distance is used to calculate the distance between each feature point and the normal operating condition distribution, and the abnormal feature points are screened out according to a preset confidence threshold to generate a risk feature vector with confidence.

4. The method according to claim 3, characterized in that The three-dimensional risk concentration distribution field is constructed according to the risk characteristic vector, and the propagation path and diffusion trend of the risk concentration distribution field are predicted based on the graph neural network, including: According to the risk characteristic vector, combined with the topological structure information of the oil and gas pipeline network, each risk characteristic point is mapped to the three-dimensional spatial coordinates of the pipeline network, and the Kriging interpolation algorithm is used to perform spatial interpolation on the discrete risk characteristic points to generate a preliminary three-dimensional risk concentration distribution field; Based on the preliminary three-dimensional risk concentration distribution field, combined with the time dimension information, the spatiotemporal Kriging interpolation algorithm is used, combined with the historical risk propagation data, to dynamically correct the risk concentration distribution field and obtain the three-dimensional risk concentration distribution field that changes with time; The three-dimensional risk concentration distribution field that changes with time is input into the graph neural network, and a graph model is constructed according to the topological structure of the pipeline network. The nodes of the graph model represent the key positions of the pipeline network, and the edges of the graph model represent the connection relationship of the pipeline sections. The graph attention mechanism is used to capture the risk propagation dependency between nodes to obtain the initial prediction result of risk propagation. According to the initial prediction results of risk propagation, a sequence prediction model based on time convolutional network is adopted, combined with historical risk propagation path data, to predict the risk propagation path and diffusion trend in future time steps, and generate the propagation prediction results of dynamic risk concentration distribution field.

5. The method according to claim 4, characterized in that Based on the prediction results of the risk concentration distribution field, an adaptive warning threshold generation model based on a generative adversarial network is constructed. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level warning threshold that changes with the operation status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the warning threshold, the risk perception warning of the operation status of the oil and gas pipeline network is realized, including: Based on historical normal operation data, a normal operating condition generator based on a generative adversarial network is trained to generate simulated data that conforms to the normal operation characteristics of the pipeline network. The generator uses a variational autoencoder with conditional constraints to ensure the diversity and authenticity of the generated data. Construct a real-time discriminator based on a deep residual network, input the prediction results of the risk concentration distribution field and the normal operating condition data generated by the generator, learn the boundary characteristics of normal and abnormal states through dynamic game, and output the discrimination results and their confidence; According to the discrimination results, an adaptive clustering algorithm is used to divide the discrimination results into multiple levels, and combined with the dynamic changes of the risk concentration distribution field, a multi-level warning threshold that changes with the operation status of the pipeline network is generated. The multi-level warning threshold includes low risk, medium risk and high risk thresholds; The real-time risk characteristics are compared with the multi-level warning thresholds, and a fuzzy logic-based decision model is used to dynamically adjust the warning level according to the confidence level of the risk characteristics and the degree of threshold deviation, thereby realizing risk-aware warning of the operating status of the oil and gas pipeline network.

6. An oil and gas pipeline network operation status risk perception and early warning system, characterized in that: The system comprises: The fusion module is used to fuse the multi-dimensional characteristics of the pipeline network operation status according to the real-time operation data of the oil and gas pipeline network, and generate a spatiotemporal feature matrix containing pressure, temperature, flow and vibration information; The modeling module is used to input the spatiotemporal feature matrix into the deep Gaussian process model, perform probabilistic modeling on the abnormal features in the operation status of the pipeline network, and output a risk feature vector with confidence; A construction module is used to construct a three-dimensional risk concentration distribution field according to the risk characteristic vector, and predict the propagation path and diffusion trend of the risk concentration distribution field based on a graph neural network; The early warning module is used to build an adaptive early warning threshold generation model based on the generative adversarial network according to the prediction results of the risk concentration distribution field. Through the dynamic game between the normal operating condition generator and the real-time discriminator, a multi-level early warning threshold that changes with the operation status of the pipeline network is generated. Based on the comparison results of the real-time risk characteristics and the early warning threshold, the risk perception early warning of the operation status of the oil and gas pipeline network is realized.

7. The system according to claim 6, characterized in that The fusion module is specifically used for: According to the real-time operation data collected by multi-source sensors of the oil and gas pipeline network, timestamp alignment and missing value filling processing are performed, and a multi-source data alignment algorithm based on dynamic time warping is used to eliminate the difference in sensor sampling frequency and obtain a time-synchronized multi-source data sequence, wherein the real-time operation data includes pressure, temperature, flow and vibration data; The multi-source data sequence is input into a spatiotemporal convolutional neural network, wherein a graph convolutional network is used in the spatial dimension to capture the topological structure characteristics of the pipeline network, and a one-dimensional convolutional kernel is used in the temporal dimension to extract the time series characteristics, thereby obtaining a preliminary spatiotemporal feature representation; For the extracted spatiotemporal feature representation, the multi-head self-attention mechanism is used to calculate the correlation weights between different sensor data, and the features are weighted fused according to the weights to highlight the contribution of key sensor data to risk perception, thus obtaining the weighted spatiotemporal feature representation; The weighted spatiotemporal features are input into the sparse autoencoder, and the dimension is reduced through nonlinear mapping to remove redundant information, retain key features, and generate a low-dimensional spatiotemporal feature matrix containing pressure, temperature, flow and vibration information.

8. The system according to claim 7, characterized in that The modeling module is specifically used for: Input the spatiotemporal feature matrix into a multi-layer perceptron, and map the original features to a high-dimensional latent space through multi-layer nonlinear transformation to obtain a deep feature representation, wherein each layer uses an activation function with a residual connection to avoid the gradient vanishing problem and enhance the feature expression capability; For the deep feature representation, a probability model based on Gaussian process regression is constructed, and a linear combination of radial basis kernel function and periodic kernel function is used to capture the nonlinear relationship and periodic change law in the operation state of the pipeline network, and obtain a preliminary probabilistic feature representation; According to the probabilistic feature representation, the predicted mean and variance of each feature point are calculated, and the uncertainty of the model is quantified through a variational inference algorithm to obtain a probabilistic feature distribution with a confidence interval; For the probability feature distribution, an anomaly detection algorithm based on Mahalanobis distance is used to calculate the distance between each feature point and the normal operating condition distribution, and the abnormal feature points are screened out according to a preset confidence threshold to generate a risk feature vector with confidence.

9. A storage medium, characterized in that: The storage medium stores a computer program, wherein the computer program is configured to execute the method according to any one of claims 1 to 5 when executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method according to any one of claims 1 to 5.

Images