Petroleum pipeline leakage detection method and system

Abstract

The invention relates to the technical field of petroleum pipelines, in particular to a petroleum pipeline leakage detection method and system, and the method comprises the following steps: S1, data collection and preprocessing: collecting the detection data of a petroleum pipeline in real time through a plurality of sensors and geographic information collection equipment; s2, feature extraction: performing feature extraction on the preprocessed detection data by adopting a multi-scale feature extraction algorithm; s3, model construction: constructing a petroleum pipeline state recognition model; s4, real-time detection: inputting detection data acquired in real time into the petroleum pipeline state recognition model; s5, leakage path prediction: predicting the diffusion path and range after petroleum leakage in real time; and S6, leakage alarm: when leakage of the petroleum pipeline is detected, an alarm mechanism is triggered, and related personnel are notified to take emergency measures. According to the method, the diffusion path and range after leakage are accurately predicted, and a reliable basis is provided for emergency response.

Description

Technical Field

[0001] The present invention relates to the technical field of oil pipelines, and in particular to an oil pipeline leakage detection method and system. Background Art

[0002] Oil pipelines are important infrastructure for oil transportation, carrying a large amount of oil flow. Once a leak occurs, it will not only cause huge economic losses, but may also cause serious environmental pollution and safety accidents. Therefore, how to efficiently and accurately detect oil pipeline leaks and respond quickly to emergencies has become an important issue in oil pipeline safety management. With the development of science and technology, modern oil pipeline leak detection technology has gradually adopted a variety of sensors and data analysis methods to improve the accuracy and timeliness of detection.

[0003] Existing oil pipeline leakage detection methods mainly rely on a single type of sensor, such as pressure sensors, flow sensors or acoustic wave sensors. These methods can usually only monitor certain specific parameters inside the pipeline and cannot fully reflect the operating status of the pipeline. In addition, the existing technology lacks effective means in data processing and feature extraction, resulting in insufficient detection sensitivity and accuracy. At the same time, traditional leakage path prediction methods and emergency response systems also have problems such as slow response and single treatment plan. They cannot meet the complex needs in practical applications and it is difficult to provide timely and effective decision support in leakage incidents.

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and provide an efficient, accurate and intelligent oil pipeline leakage detection method and system, which can significantly improve the efficiency and effectiveness of emergency response, thereby greatly improving the operational safety and reliability of oil pipelines. Summary of the invention

[0005] Based on the above objectives, the present invention provides a method and system for detecting oil pipeline leakage.

[0006] The oil pipeline leakage detection method comprises the following steps: S1, data collection and preprocessing: collect the detection data of the oil pipeline in real time through multiple sensors and geographic information collection equipment, including the pressure, flow and acoustic signal data inside the pipeline, as well as the terrain, meteorological conditions and soil properties around the pipeline, and preprocess the collected detection data; S2, feature extraction: a multi-scale feature extraction algorithm is used to extract features from the pre-processed detection data, including time features, frequency features and spatial features; S3, model construction: based on the extracted features, build an oil pipeline status recognition model; S4, real-time detection: input the real-time collected detection data into the oil pipeline state recognition model to perform real-time leakage detection to determine whether the oil pipeline has leaked, and adopt a dynamic threshold adjustment algorithm based on Bayesian reasoning to adapt to changes in pipeline operation status; S5, Leakage path prediction: Based on the detection data of oil pipelines and fluid dynamics models, the diffusion path and range of oil leaks are predicted in real time, including: S51, oil leakage simulation: Based on the collected oil pipeline detection data, the fluid dynamics model is used to simulate the oil flow in the pipeline, considering the physical properties, pressure distribution and flow changes of the oil in the pipeline, and simulating the flow behavior of the oil in the case of leakage; S52, real-time leak path prediction: using fluid dynamics models and real-time detection data to predict the diffusion path and range of oil leaks; S53, emergency decision support: based on the results of leakage path prediction, automatically generate emergency response plans, including leakage control, personnel evacuation and environmental protection measures, and adjust the emergency plans in real time; S6, Leakage Alarm: When a leak is detected in an oil pipeline, the alarm mechanism is triggered to notify relevant personnel to take emergency measures, and the diffusion path and scope of the oil leak are displayed through dynamic simulation and visualization tools.

[0007] Optionally, the data collection and preprocessing in S1 includes: S11, detection data collection: real-time collection of pressure, flow and acoustic wave signal data inside the oil pipeline through pressure sensors, flow sensors and acoustic wave sensors, and collection of terrain data, meteorological conditions and soil properties around the pipeline through environmental sensors and geographic information collection equipment; S12, data denoising: using moving average filtering to denoise the collected detection data; S13, data smoothing: using exponentially weighted moving average (EWMA) to smooth the denoised detection data to eliminate mutations and glitches in the data; S14, data normalization: normalizing the smoothed detection data to standardize the detection data of different sensors to the same scale range; S15, environmental factor compensation: perform environmental factor compensation on sensor data based on the terrain, meteorological and soil data acquired by geographic information acquisition equipment.

[0008] Optionally, the feature extraction in S2 includes: S21, time feature extraction: using sliding window statistics to extract features from the time series data in the preprocessed detection data; S22, frequency feature extraction: frequency domain analysis of the pre-processed detection data is performed using Fourier transform; S23, spatial feature extraction: extracting spatial features from the preprocessed detection data using spatial statistics; S24, multi-scale analysis: The detection data is decomposed into components at different scales through wavelet transform to capture the multi-level characteristics of the data.

[0009] Optionally, the oil pipeline state recognition model in S3 adopts a graph convolutional network (GCN) model, and the graph convolutional network (GCN) model includes: S31, constructing a graph structure: constructing a graph structure according to the physical location and connection relationship of the sensors, where nodes represent sensors and edges represent physical connections or mutual influences between sensors; S32, feature initialization: Initialize the extracted time features, frequency features and space features. The calculation formula is: ; in, is the initial feature of node v, , and They are time feature, frequency feature and space feature extraction functions respectively; S33, graph convolution layer: In each layer of graph convolution, the node’s own features and neighbor node features are combined to update the features. The calculation formula is: ; in, is the feature representation of node v at the k+1th layer, is the weight matrix of the kth layer, is the bias, is the ReLU activation function, is the set of neighbor nodes of node v, is the normalization coefficient, is the feature representation of node u at layer k, is the extended neighborhood set of node v; S34, dynamic weighting mechanism: Introduce a dynamic weighting mechanism to weight the aggregation process by calculating the importance coefficient of neighbor nodes. The calculation formula is: ; ; ; in, is the attention coefficient between node v and node u, is the feature representation of node v at layer k, is the normalized attention weight of node v to node u, is the normalized denominator, which is the sum of the exponential values ​​of all attention scores in the neighbor set of node v. is the attention weight vector, is a vector concatenation operation; S35, final feature representation and classification: After the last layer of graph convolution, the features of each node are summarized and classified using a fully connected layer to determine whether each node (sensor) detects a leak. The calculation formula is: ; in, is the classification result of node v, is the weight matrix of the final classification layer, is the bias term, is the classification activation function, is the feature representation of node v in the final layer.

[0010] Optionally, the real-time detection in S4 includes: S41, real-time data input: the detection data collected in real time is input into the oil pipeline state recognition model; S42, state recognition: The oil pipeline state recognition model processes the input real-time detection data and generates the state probability distribution of each node (sensor). The calculation formula is: ; in, is the classification result of node v, is the weight matrix of the final classification layer, is the bias term, is the classification activation function, is the feature representation of node v at the final layer.

[0011] S43, leakage judgment: judging whether the oil pipeline has leakage according to the result of state recognition; S44, dynamic threshold adjustment: adopt the dynamic threshold adjustment algorithm based on Bayesian reasoning to adjust the threshold of leakage judgment in real time. The calculation formula is: ; in, is the parameter under the given data D The posterior probability of is a parameter The likelihood of the data D is, is a parameter The prior probability of is the marginal probability of data D; S45, threshold update: calculate the new threshold based on real-time monitoring data ,like Exceed , then it is judged that the pipeline is leaking, and the calculation formula is: ; in, is the updated threshold, is the previous threshold, is the learning rate, is the posterior probability calculated based on Bayesian inference, is the target probability.

[0012] Optionally, the oil spill simulation in S51 includes: S511, Flow equation: The Navier-Stokes equation is used to describe the flow behavior of oil in the pipeline. The calculation formula is: ; in, is the density of oil, u is the velocity vector, p is the pressure, is the dynamic viscosity, f is the external force (such as gravity), is the rate of change of the velocity vector field with time t, is the pressure gradient force, is the Laplace operator of the velocity vector field; S512, Mass Conservation Equation: Use the continuity equation to ensure mass conservation, calculated as: ; S513, boundary conditions and initial conditions: set the pipeline boundary conditions and initial conditions, the boundary conditions include the no-slip condition of the pipeline wall and the flow conditions of the inlet and outlet; S514, Leakage simulation: Introduce a leakage point into the fluid dynamics model to simulate the flow behavior of oil in the event of a leakage. Consider the impact of pressure drop and flow change on the overall flow. Simulate the occurrence and propagation of leakage by adjusting the pressure and flow at the leakage point. The calculation formula is: ; in, is the leakage flow, is the flow coefficient at the leak point, is the leakage area, and are the pressures inside and outside the pipe respectively; S515, Numerical solution: The finite element method is used to solve the flow equation and the mass conservation equation to obtain the numerical solution of the oil flow in the pipeline.

[0013] Optionally, the real-time leakage path prediction in S52 includes: S521, real-time detection data input: input the real-time collected detection data into the fluid dynamics model to update the current flow state; S522, leakage point identification: Identify the location and time of leakage in the real-time detection data, determine the location and leakage amount of the leakage point, and calculate the formula as follows: ; S523, real-time path prediction: Based on the updated fluid dynamics model and leakage point information, the oil diffusion path and range after the leakage are calculated in real time, and the oil diffusion outside the pipeline is predicted. The calculation formula is: ; in, is the oil concentration at time t and location r, is the location of the leak, is the diffusion coefficient.

[0014] Optionally, the emergency decision support in S53 includes: S531, based on the results of leakage path prediction: using the results of real-time leakage path prediction, determine the diffusion path and scope of oil leakage, and identify the affected areas and key points; S532, automatically generate emergency treatment plan: according to the leakage path prediction result, automatically generate emergency treatment plan, specifically including: Leak control: Identify isolation and control measures for leaks, including closing pipeline valves, deploying containment booms and oil suction equipment; Evacuation: assess the risk to people in the affected area and plan evacuation routes and shelter locations; Environmental protection measures: Assess the impact of the leak on the environment and develop environmental protection measures, including laying oil absorbent mats, setting up pollution control barriers and emergency recovery of leaked oil; S533, real-time adjustment of emergency plan: During the emergency treatment process, real-time monitoring of leakage and diffusion is carried out, and the emergency treatment plan is dynamically adjusted in combination with the latest prediction results and on-site data. The calculation formula is: ; in, It is an adjusted emergency plan. It is the current emergency plan. is the adjustment factor, This is a forecast based on the latest data. is the target control level.

[0015] Optionally, the leakage alarm in S6 includes: S61, leakage detection: when a leak is detected in the oil pipeline, confirm the leak event according to a preset leak determination threshold; S62, triggering the alarm mechanism: if the leakage event is confirmed, the alarm mechanism is immediately triggered to generate an alarm signal, which is sent through a variety of means, including SMS, email, phone and emergency broadcast; S63, Notify relevant personnel: notify relevant emergency response personnel and management personnel through pre-set communication channels and provide information on the leak, including the location, volume and initial spread of the leak; S64, Dynamic Simulation and Visualization: Using dynamic simulation tools, based on real-time detection data and leakage path prediction results, simulate the diffusion path and range of oil spills; S65, Visualization Display: Through visualization tools, dynamic simulation results are displayed on a map or 3D model to intuitively show the diffusion path and affected area of ​​the oil spill. The visualization display includes the leakage point, diffusion path, affected area and key facilities.

[0016] The oil pipeline leakage detection system is used to implement the above-mentioned oil pipeline leakage detection method, and includes the following modules: Data acquisition and preprocessing module: collects the detection data of the oil pipeline in real time through multiple sensors and geographic information acquisition equipment, including the pressure, flow and acoustic signal data inside the pipeline, as well as the terrain, meteorological conditions and soil properties around the pipeline, and preprocesses the collected detection data; Feature extraction module: uses a multi-scale feature extraction algorithm to extract features from the pre-processed detection data, including time features, frequency features, and spatial features; Model building module: Based on the extracted features, the oil pipeline status recognition model is constructed; Real-time detection module: inputs the real-time collected detection data into the oil pipeline status recognition model to perform real-time leakage detection, determine whether the oil pipeline has leaked, and adopts a dynamic threshold adjustment algorithm based on Bayesian reasoning to adapt to changes in pipeline operation status; Leakage path prediction module: Based on the detection data of oil pipelines and fluid dynamics models, it can predict the diffusion path and range of oil leakage in real time; Leakage alarm module: When a leak is detected in an oil pipeline, the alarm mechanism is triggered to notify relevant personnel to take emergency measures, and the diffusion path and scope of the oil leak are displayed through dynamic simulation and visualization tools.

[0017] Beneficial effects of the present invention: The present invention performs in-depth feature analysis on data through a multi-scale feature extraction algorithm, which significantly enhances the detection capability of different leakage types, thereby improving the sensitivity and accuracy of leakage detection, ensuring the high quality and reliability of data, and laying a solid foundation for subsequent state identification and leakage prediction.

[0018] The present invention adopts a graph convolutional network model, combines the node's own characteristics and neighbor node characteristics to update the characteristics, and introduces a dynamic weighting mechanism to improve the robustness and recognition accuracy of the model. The real-time leakage detection system combines the dynamic threshold adjustment algorithm of Bayesian reasoning to adapt to changes in pipeline operation status in a timely manner and ensure the accuracy of leakage judgment. The real-time leakage path prediction system uses fluid dynamics models and real-time detection data to accurately predict the diffusion path and range after leakage, providing a reliable basis for emergency response.

[0019] The present invention notifies relevant personnel to take emergency measures by triggering an alarm mechanism, and displays the leakage diffusion path and affected scope through dynamic simulation and visualization tools, thereby helping emergency response personnel to quickly understand the leakage situation and formulate response measures. By automatically generating and adjusting emergency treatment plans in real time, including leakage control, personnel evacuation and environmental protection measures, the system significantly improves the efficiency and effectiveness of emergency response, ensures that oil pipeline leakage incidents can be handled quickly and effectively, and thus greatly improves the safety and reliability of pipeline operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 A schematic diagram of a detection method flow chart of an embodiment of the present invention; Figure 2 Schematic diagram of system function modules according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0023] like Figure 1 As shown, the oil pipeline leakage detection method comprises the following steps: S1, data collection and preprocessing: multiple sensors and geographic information collection equipment are used to collect the detection data of the oil pipeline in real time, including the pressure, flow and acoustic signal data inside the pipeline, as well as the terrain, meteorological conditions and soil properties around the pipeline, and the collected detection data are preprocessed to obtain high-quality detection data; S2, feature extraction: a multi-scale feature extraction algorithm is used to extract features from the pre-processed detection data, including time features, frequency features and spatial features, to enhance the detection capability of different leakage types; S3, model construction: based on the extracted features, build an oil pipeline status recognition model; S4, real-time detection: input the real-time collected detection data into the oil pipeline state recognition model to perform real-time leakage detection to determine whether the oil pipeline has leaked, and adopt a dynamic threshold adjustment algorithm based on Bayesian reasoning to adapt to changes in pipeline operation status; S5, Leakage path prediction: Based on the detection data of oil pipelines and fluid dynamics models, the diffusion path and range of oil leaks are predicted in real time, including: S51, oil leakage simulation: Based on the collected oil pipeline detection data, the fluid dynamics model is used to simulate the oil flow in the pipeline, considering the physical properties, pressure distribution and flow changes of the oil in the pipeline, and simulating the flow behavior of the oil in the case of leakage; S52, real-time leak path prediction: using fluid dynamics models and real-time detection data to predict the diffusion path and range of oil leaks; S53, emergency decision support: based on the results of leakage path prediction, automatically generate emergency response plans, including leakage control, personnel evacuation and environmental protection measures, and adjust the emergency plans in real time to improve the efficiency and effectiveness of emergency response; S6, Leakage Alarm: When a leak is detected in an oil pipeline, the alarm mechanism is triggered to notify relevant personnel to take emergency measures, and the diffusion path and scope of the oil leak are displayed through dynamic simulation and visualization tools to help pipeline managers respond in a timely manner; Through the above steps, efficient, accurate and intelligent leak detection and emergency response are achieved. By comprehensively utilizing multi-source data, the sensitivity of detection and the efficiency of emergency treatment are significantly improved, thereby greatly improving the operational safety and reliability of oil pipelines.

[0024] Data collection and preprocessing in S1 include: S11, detection data collection: real-time collection of pressure, flow and acoustic wave signal data inside the oil pipeline through pressure sensors, flow sensors and acoustic wave sensors, and collection of terrain data, meteorological conditions and soil properties around the pipeline through environmental sensors and geographic information collection equipment; S12, data denoising: Use moving average filtering to denoise the collected detection data. The calculation formula is: ; in, is the filtered data, is the detection data, N is the size of the filter window, For parameters Represents the original detection data value traced back i time steps at the current time t; S13, data smoothing: Use exponentially weighted moving average (EWMA) to smooth the denoised detection data to eliminate mutations and glitches in the data. The calculation formula is: ; in, is the smoothing coefficient, usually in the range of , is the smoothed detection data, is the filtered data, is the smoothing result of the previous moment t-1; S14, data normalization: normalize the smoothed detection data to standardize the detection data of different sensors to the same scale range for subsequent processing and analysis. The calculation formula is: ; in, is the detection data after smoothing. and are the minimum and maximum values ​​of the data, respectively. is the normalized data; S15, environmental factor compensation: Based on the terrain, meteorological and soil data obtained by the geographic information acquisition equipment, the sensor data is compensated for environmental factors to correct the measurement errors caused by environmental changes. The calculation formula is: ; in, is the normalized data, is the environmental compensation coefficient, is environmental data (such as temperature, humidity, etc.), is the data after compensation;` Through the above steps, key detection data inside and around the oil pipeline are obtained in real time, and through processing such as denoising, smoothing, normalization and environmental factor compensation, the high quality and accuracy of the data are ensured, thereby improving the sensitivity and accuracy of detection and enhancing the adaptability and reliability of the system.

[0025] Feature extraction in S2 includes: S21, time feature extraction: Use sliding window statistics to extract features from the time series data in the preprocessed detection data. The calculation formula is: ; ; in, is the mean value within the sliding window, is the standard deviation within the sliding window, W is the window size, is the time series data value of i time steps forward from the current time t; S22, frequency feature extraction: Fourier transform is used to perform frequency domain analysis on the pre-processed detection data. The calculation formula is: ; in, is the frequency domain signal, is the time domain signal, N is the signal length, and f is the frequency; S23, spatial feature extraction: spatial feature extraction is performed on the pre-processed detection data using spatial statistics, and the calculation formula is: ; in, is the spatial covariance, and is the position data of two different sensors, and are their means, respectively, and M is the total number of data points used to calculate the spatial covariance; S24, multi-scale analysis: The detection data is decomposed into components at different scales through wavelet transform to capture the multi-level characteristics of the data. The calculation formula is: ; in, are the wavelet coefficients, is the signal, wavelet function The scale is s and the translation is t; Through the above steps, multi-level features in time, frequency and space can be extracted from the preprocessed detection data, enhancing the detection capability of different leakage types, accurately capturing the changes in the operating status of the oil pipeline, and improving the sensitivity and accuracy of leakage detection.

[0026] The oil pipeline status recognition model in S3 uses a graph convolutional network (GCN) model, which includes: S31, constructing a graph structure: constructing a graph structure according to the physical location and connection relationship of the sensors, where nodes represent sensors and edges represent physical connections or mutual influences between sensors; S32, feature initialization: Initialize the extracted time features, frequency features and space features. The calculation formula is: ; in, is the initial feature of node v, , and They are time feature, frequency feature and space feature extraction functions respectively; S33, graph convolution layer: In each layer of graph convolution, the node’s own features and neighbor node features are combined to update the features. The calculation formula is: ; in, is the feature representation of node v at the k+1th layer, is the weight matrix of the kth layer, is the bias, is the ReLU activation function, is the set of neighbor nodes of node v, is the normalization coefficient, is the feature representation of node u at layer k, is the extended neighborhood set of node v; The normalization coefficient uses the symmetric normalized Laplace matrix, and the calculation formula is: ; in, and are the degrees of node v and node u respectively; S34, dynamic weighting mechanism: Introduce a dynamic weighting mechanism to weight the aggregation process by calculating the importance coefficient of neighbor nodes, and enhance the focus on the characteristics of important neighbor nodes. The calculation formula is: ; ; ; in, is the attention coefficient between node v and node u, is the feature representation of node v at layer k, is the normalized attention weight of node v to node u, is the normalized denominator, which is the sum of the exponential values ​​of all attention scores in the neighbor set of node v. is the attention weight vector, is a vector concatenation operation; S35, final feature representation and classification: After the last layer of graph convolution, the features of each node are summarized and classified using a fully connected layer to determine whether each node (sensor) detects a leak. The calculation formula is: ; in, is the classification result of node v, is the weight matrix of the final classification layer, is the bias term, is the classification activation function, is the feature representation of node v in the final layer; Classification Activation Function The calculation formula is: ; in, is the input score of the ith category, K is the total number of categories, is the normalization term; Through the above, the complex dependencies in the sensor network can be captured more accurately, the sensitivity and accuracy of leak detection can be improved, and thus the overall performance and robustness of oil pipeline status identification can be enhanced.

[0027] Real-time detection in S4 includes: S41, real-time data input: the detection data collected in real time is input into the oil pipeline state recognition model; S42, state recognition: The oil pipeline state recognition model processes the input real-time detection data and generates the state probability distribution of each node (sensor). The calculation formula is: ; in, is the classification result of node v, is the weight matrix of the final classification layer, is the feature vector of node v, is the bias term; S43, leakage judgment: judging whether the oil pipeline has leakage according to the result of state recognition, if the leakage state probability of a node exceeds a preset threshold, it is judged that leakage has occurred at the node; S44, dynamic threshold adjustment: In order to adapt to changes in pipeline operation status, a dynamic threshold adjustment algorithm based on Bayesian reasoning is used to adjust the threshold of leakage judgment in real time. The calculation formula is: ; in, is the parameter under the given data D The posterior probability of is a parameter The likelihood of the data D is, is a parameter The prior probability of is the marginal probability of data D; S45, threshold update: calculate the new threshold based on real-time monitoring data ,like Exceed , then it is judged that the pipeline is leaking, and the calculation formula is: ; in, is the updated threshold, is the previous threshold, is the learning rate, is the posterior probability calculated based on Bayesian inference, is the target probability; Through the above steps, the real-time leakage detection method for oil pipelines can timely and accurately identify pipeline leakage, and adapt to changes in pipeline operation status by dynamically adjusting thresholds, thereby improving the sensitivity and accuracy of leakage detection.

[0028] Oil spill simulation in S51 includes: S511, Flow equation: The Navier-Stokes equation is used to describe the flow behavior of oil in the pipeline. The calculation formula is: ; in, is the density of oil, u is the velocity vector, p is the pressure, is the dynamic viscosity, f is the external force (such as gravity), is the rate of change of the velocity vector field with time t, is the pressure gradient force, is the Laplace operator of the velocity vector field; S512, Mass Conservation Equation: Use the continuity equation to ensure mass conservation, calculated as: ; S513, boundary conditions and initial conditions: set the pipeline boundary conditions and initial conditions, the boundary conditions include the no-slip condition of the pipeline wall and the flow conditions of the inlet and outlet; The boundary conditions and initial conditions specifically include: S5131, initial conditions: Initial conditions refer to the state of the fluid in the pipeline at the beginning of the simulation, including the initial velocity field and pressure field. The calculation formula is: ; ; in, is the velocity distribution at the initial moment, is the pressure distribution at the initial moment, is the spatial position at time t=0 The fluid velocity vector, is the spatial position at time t=0 Fluid pressure; S5132, Boundary conditions: Boundary conditions include no-slip conditions on the pipe wall and flow conditions at the inlet and outlet, specifically including: No-slip condition on the pipe wall: u=0 on the pipe wall. The no-slip condition means that the velocity of the fluid is zero at the pipe wall, that is, the fluid adheres to the wall. Inlet conditions: At the inlet of the pipe, set the flow rate or flow rate of the fluid, expressed as: At the entrance; At the entrance; in, is the velocity at the entrance, is the pressure at the inlet; Outlet conditions: At the outlet of the pipeline, set the pressure or flow of the fluid, expressed as: At the exit; At the exit; in, is the component of the velocity at the outlet in the normal direction, n is the normal unit vector, is the flow velocity at the outlet, is the pressure at the outlet; S5133, Comprehensive Application: In practical applications, initial conditions and boundary conditions are set according to specific circumstances, including: At the inlet of the pipeline, a constant inlet flow rate or inlet pressure is set; At the outlet of the pipeline, a constant outlet pressure is set; At the pipe wall, no-slip conditions are always set; S514, Leakage simulation: Introduce a leakage point into the fluid dynamics model to simulate the flow behavior of oil in the event of a leakage. Consider the impact of pressure drop and flow change on the overall flow. Simulate the occurrence and propagation of leakage by adjusting the pressure and flow at the leakage point. The calculation formula is: ; in, is the leakage flow, is the flow coefficient at the leak point, is the leakage area, and are the pressures inside and outside the pipe respectively; S515, Numerical solution: The finite element method is used to solve the flow equation and the mass conservation equation to obtain the numerical solution of the oil flow in the pipeline; Finite element methods include: S5151, weak form: Convert the Navier-Stokes equations and the continuity equation into weak form, and the calculation formula is: ; ; in, is the solution area, v and q are the test functions of velocity and pressure respectively, is the convection acceleration of the fluid, For control The tiny volume unit in S5152, Discretization: Use linear interpolation function to discretize the region Divided into small finite elements, the calculation formula is: ; ; in, is the interpolation function, and are the nodal velocities and pressures, and N is the total number of interpolation functions; S5153, Matrix equation: The discretized weak form equation is converted into a matrix equation. The calculation formula is: ; ; Where M is the mass matrix, is the convection matrix, K is the stiffness matrix, F is the external force matrix, B is the matrix corresponding to the continuity equation, B is the matrix corresponding to the continuity equation, and U is the node velocity vector; S5154, numerical solution: The implicit Euler method is used for time discretization, and the discretized matrix equation is solved to obtain the numerical solution of velocity and pressure. The calculation formula is: ; in, is the velocity vector of the node at the n+1th time step, is the velocity vector of the node at the nth time step, is the time step; Through the above content, detailed leakage prediction and assessment can be provided, helping to identify potential leakage points and leakage diffusion paths, supporting timely and effective emergency response decisions, thereby greatly improving the safety and reliability of oil pipeline operations.

[0029] Real-time leak path prediction in S52 includes: S521, real-time detection data input: input the real-time collected detection data into the fluid dynamics model to update the current flow state; S522, leakage point identification: Identify the location and time of leakage in the real-time detection data, determine the location and leakage amount of the leakage point, and calculate the formula as follows: ; S523, real-time path prediction: Based on the updated fluid dynamics model and leakage point information, the oil diffusion path and range after the leakage are calculated in real time, and the oil diffusion outside the pipeline is predicted. The calculation formula is: ; in, is the oil concentration at time t and location r, is the location of the leak, is the diffusion coefficient; Through the above steps, fluid dynamics models and real-time detection data can be used to accurately predict the diffusion path and range of oil spills, providing a reliable basis for emergency response and leakage control, thereby improving the safety of pipeline operations and the efficiency of responding to leakage incidents.

[0030] Emergency decision support in S53 includes: S531, based on the results of leakage path prediction: using the results of real-time leakage path prediction, determine the diffusion path and scope of oil leakage, and identify the affected areas and key points; S532, automatically generate emergency treatment plan: according to the leakage path prediction result, automatically generate emergency treatment plan, specifically including: Leak control: Identify isolation and control measures for leaks, including closing pipeline valves, deploying containment booms and oil suction equipment; Evacuation: Assess the risk to personnel in the affected area, develop evacuation routes and shelters, and ensure safe evacuation of personnel; Environmental protection measures: Assess the impact of the leak on the environment and develop environmental protection measures, including laying oil absorbent mats, setting up pollution control barriers and emergency recovery of leaked oil; S533, real-time adjustment of emergency plan: During the emergency treatment process, real-time monitoring of leakage and diffusion is carried out, and the emergency treatment plan is dynamically adjusted in combination with the latest prediction results and on-site data. The calculation formula is: ; in, It is an adjusted emergency plan. It is the current emergency plan. is the adjustment factor, This is a forecast based on the latest data. is the target control level; Through the above steps, a comprehensive emergency response plan can be automatically generated based on the results of leakage path prediction, including leakage control, personnel evacuation and environmental protection measures, and the plan can be adjusted in real time during the emergency response process, thereby improving the efficiency and effectiveness of emergency response and ensuring the rapid and effective handling of oil spill incidents.

[0031] Leak alarms in S6 include: S61, leakage detection: when a leak is detected in the oil pipeline, confirm the leak event according to a preset leak determination threshold; S62, triggering the alarm mechanism: if a leakage event is confirmed, the alarm mechanism is immediately triggered to generate an alarm signal, which is sent through a variety of means, including SMS, email, phone and emergency broadcast, to ensure that relevant personnel can receive the notification in time; S63, Notify relevant personnel: notify relevant emergency response personnel and management personnel through pre-set communication channels and provide information on the leak, including the location, volume and initial spread of the leak; S64, Dynamic Simulation and Visualization: Using dynamic simulation tools, based on real-time detection data and leakage path prediction results, simulate the diffusion path and range of oil spills; S65, Visualization: Through visualization tools, dynamic simulation results are displayed on a map or 3D model to intuitively show the diffusion path and affected area of ​​the oil spill. The visualization includes the leak point, diffusion path, affected area and key facilities, helping emergency responders to quickly understand the leak and take appropriate measures. Key facilities include: Pipeline valves: key components for controlling the flow of oil. Closing the valves can limit the scope of leakage. Pumping station: A facility used to transport oil, where leakage may affect its normal operation; Storage tanks: Facilities where oil is stored, it is vital to prevent leaks from spreading to the tank area; Monitoring equipment: Sensors and instruments used to monitor the status of pipelines. Ensuring their proper operation is critical to detecting and managing leaks. Emergency Response Center: A facility that coordinates emergency measures and ensures that its communications and command functions are functioning properly; Power and communications infrastructure: Power supply and communications networks that maintain the normal operation of pipelines and related facilities; Environmental protection facilities: facilities used to prevent and mitigate the impact of leaks on the environment, such as oil absorption booms, oil-water separators, etc.; Public safety facilities: including evacuation routes and shelters to ensure the safety of people; Through the above steps, when an oil pipeline leak is detected, the alarm mechanism can be triggered quickly to notify relevant personnel to take emergency measures, and the diffusion path and scope of the oil leak can be displayed through dynamic simulation and visualization tools, thereby improving the timeliness and effectiveness of the emergency response and ensuring that the leakage incident is quickly controlled and handled.

[0032] like Figure 2 As shown, the oil pipeline leakage detection system is used to implement the above-mentioned oil pipeline leakage detection method, including the following modules: Data acquisition and preprocessing module: collects the detection data of the oil pipeline in real time through multiple sensors and geographic information acquisition equipment, including the pressure, flow and acoustic signal data inside the pipeline, as well as the terrain, meteorological conditions and soil properties around the pipeline, and preprocesses the collected detection data; Feature extraction module: uses a multi-scale feature extraction algorithm to extract features from the pre-processed detection data, including time features, frequency features, and spatial features; Model building module: Based on the extracted features, the oil pipeline status recognition model is constructed; Real-time detection module: inputs the real-time collected detection data into the oil pipeline status recognition model to perform real-time leakage detection, determine whether the oil pipeline has leaked, and adopts a dynamic threshold adjustment algorithm based on Bayesian reasoning to adapt to changes in pipeline operation status; Leakage path prediction module: Based on the detection data of oil pipelines and fluid dynamics models, it can predict the diffusion path and range of oil leakage in real time; Leakage alarm module: When a leak is detected in an oil pipeline, the alarm mechanism is triggered to notify relevant personnel to take emergency measures, and the diffusion path and scope of the oil leak are displayed through dynamic simulation and visualization tools.

[0033] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0034] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for detecting leakage in a petroleum pipeline, characterized in that: The following steps are involved: S1, data collection and preprocessing: collect the detection data of the oil pipeline in real time through multiple sensors and geographic information collection equipment, including the pressure, flow and acoustic signal data inside the pipeline, as well as the terrain, meteorological conditions and soil properties around the pipeline, and preprocess the collected detection data; S2, feature extraction: a multi-scale feature extraction algorithm is used to extract features from the pre-processed detection data, including time features, frequency features and spatial features; S3, model construction: based on the extracted features, build an oil pipeline status recognition model; S4, real-time detection: input the real-time collected detection data into the oil pipeline state recognition model to perform real-time leakage detection to determine whether the oil pipeline has leaked, and adopt a dynamic threshold adjustment algorithm based on Bayesian reasoning to adapt to changes in pipeline operation status; S5, Leakage path prediction: Based on the detection data of oil pipelines and fluid dynamics models, the diffusion path and range of oil leaks are predicted in real time, including: S51, oil leakage simulation: Based on the collected oil pipeline detection data, the fluid dynamics model is used to simulate the oil flow in the pipeline, considering the physical properties, pressure distribution and flow changes of the oil in the pipeline, and simulating the flow behavior of the oil in the case of leakage; S52, real-time leak path prediction: using fluid dynamics models and real-time detection data to predict the diffusion path and range of oil leaks; S53, emergency decision support: based on the results of leakage path prediction, automatically generate emergency response plans, including leakage control, personnel evacuation and environmental protection measures, and adjust the emergency plans in real time; S6, Leakage Alarm: When a leak is detected in an oil pipeline, the alarm mechanism is triggered to notify relevant personnel to take emergency measures, and the diffusion path and scope of the oil leak are displayed through dynamic simulation and visualization tools; The oil pipeline state recognition model in S3 adopts a graph convolutional network model, which includes: S31, constructing a graph structure: constructing a graph structure according to the physical location and connection relationship of the sensors, where nodes represent sensors and edges represent physical connections or mutual influences between sensors; S32, feature initialization: Initialize the extracted time features, frequency features and space features. The calculation formula is: ； in, is the initial feature of node v, , and They are time feature, frequency feature and space feature extraction functions respectively; S33, graph convolution layer: In each layer of graph convolution, the node’s own features and neighbor node features are combined to update the features. The calculation formula is: ； in, is the feature representation of node v at the k+1th layer, is the weight matrix of the kth layer, is the bias, is the ReLU activation function, is the set of neighbor nodes of node v, is the normalization coefficient, is the feature representation of node u at layer k, is the extended neighborhood set of node v; S34, dynamic weighting mechanism: Introduce a dynamic weighting mechanism to weight the aggregation process by calculating the importance coefficient of neighbor nodes. The calculation formula is: ；. ； ； in, is the attention coefficient between node v and node u, is the feature representation of node v at layer k, is the normalized attention weight of node v to node u, is the normalized denominator, which is the sum of the exponential values ​​of all attention scores in the neighbor set of node v. is the attention weight vector, is a vector concatenation operation; S35, final feature representation and classification: After the last layer of graph convolution, the features of each node are summarized and classified using a fully connected layer to determine whether a leak is detected at each node. The calculation formula is: ； in, is the classification result of node v, is the weight matrix of the final classification layer, is the bias term, is the classification activation function, is the feature representation of node v in the final layer; The real-time detection in S4 includes: S41, real-time data input: the detection data collected in real time is input into the oil pipeline state recognition model; S42, state recognition: The oil pipeline state recognition model processes the input real-time detection data and generates the state probability distribution of each node. The calculation formula is: ； in, is the classification result of node v, is the weight matrix of the final classification layer, is the feature vector of node v, is the bias term; S43, leakage judgment: judging whether the oil pipeline has leakage according to the result of state recognition; S44, dynamic threshold adjustment: adopt the dynamic threshold adjustment algorithm based on Bayesian reasoning to adjust the threshold of leakage judgment in real time. The calculation formula is: ； in, is the parameter under the given data D The posterior probability of is a parameter The likelihood of the data D is, is a parameter The prior probability of is the marginal probability of data D; S45, threshold update: calculate the new threshold based on real-time monitoring data ,like Exceed , then it is judged that the pipeline is leaking, and the calculation formula is: ； in, is the updated threshold, is the previous threshold, is the learning rate, is the posterior probability calculated based on Bayesian inference, is the target probability.

2. The oil pipeline leakage detection method according to claim 1, characterized in that: The data collection and preprocessing in S1 include: S11, detection data collection: real-time collection of pressure, flow and acoustic wave signal data inside the oil pipeline through pressure sensors, flow sensors and acoustic wave sensors, and collection of terrain data, meteorological conditions and soil properties around the pipeline through environmental sensors and geographic information collection equipment; S12, data denoising: using moving average filtering to denoise the collected detection data; S13, data smoothing: using exponentially weighted moving average to smooth the denoised detection data to eliminate mutations and glitches in the data; S14, data normalization: normalizing the smoothed detection data to standardize the detection data of different sensors to the same scale range; S15, environmental factor compensation: perform environmental factor compensation on sensor data based on the terrain, meteorological and soil data acquired by geographic information acquisition equipment.

3. The oil pipeline leakage detection method according to claim 1, characterized in that: The feature extraction in S2 includes: S21, time feature extraction: using sliding window statistics to extract features from the time series data in the preprocessed detection data; S22, frequency feature extraction: frequency domain analysis of the pre-processed detection data is performed using Fourier transform; S23, spatial feature extraction: extracting spatial features from the preprocessed detection data using spatial statistics; S24, multi-scale analysis: The detection data is decomposed into components at different scales through wavelet transform to capture the multi-level characteristics of the data.

4. The oil pipeline leakage detection method according to claim 1, characterized in that: The oil spill simulation in S51 includes: S511, Flow equation: The Navier-Stokes equation is used to describe the flow behavior of oil in the pipeline. The calculation formula is: ； in, is the density of oil, u is the velocity vector, p is the pressure, is the dynamic viscosity, f is the external force, is the rate of change of the velocity vector field with time t, is the pressure gradient force, is the Laplace operator of the velocity vector field; S512, Mass Conservation Equation: Use the continuity equation to ensure mass conservation, calculated as: ； S513, boundary conditions and initial conditions: set the pipeline boundary conditions and initial conditions, the boundary conditions include the no-slip condition of the pipeline wall and the flow conditions of the inlet and outlet; S514, Leakage simulation: Introduce a leakage point into the fluid dynamics model to simulate the flow behavior of oil in the event of a leakage. Consider the impact of pressure drop and flow change on the overall flow. Simulate the occurrence and propagation of leakage by adjusting the pressure and flow at the leakage point. The calculation formula is: ； in, is the leakage flow, is the flow coefficient at the leak point, is the leakage area, and are the pressures inside and outside the pipe respectively; S515, Numerical solution: The finite element method is used to solve the flow equation and the mass conservation equation to obtain the numerical solution of the oil flow in the pipeline.

5. The oil pipeline leakage detection method according to claim 4, characterized in that: The real-time leakage path prediction in S52 includes: S521, real-time detection data input: input the real-time collected detection data into the fluid dynamics model to update the current flow state; S522, leakage point identification: Identify the location and time of leakage in the real-time detection data, determine the location of the leakage point and the leakage amount, and calculate the formula as follows: ； S523, real-time path prediction: Based on the updated fluid dynamics model and leakage point information, the oil diffusion path and range after the leakage are calculated in real time, and the oil diffusion outside the pipeline is predicted. The calculation formula is: ； in, is the oil concentration at time t and location r, is the location of the leak, is the diffusion coefficient.

6. The oil pipeline leakage detection method according to claim 5, characterized in that: The emergency decision support in S53 includes: S531, based on the results of leakage path prediction: using the results of real-time leakage path prediction, determine the diffusion path and scope of oil leakage, and identify the affected areas and key points; S532, automatically generate emergency treatment plan: according to the leakage path prediction result, automatically generate emergency treatment plan, specifically including: Leak control: Identify isolation and control measures for leaks, including closing pipeline valves, deploying containment booms and oil suction equipment; Evacuation: assess the risk to people in the affected area and plan evacuation routes and shelter locations; Environmental protection measures: Assess the impact of the leak on the environment and develop environmental protection measures, including laying oil absorbent mats, setting up pollution control barriers and emergency recovery of leaked oil; S533, real-time adjustment of emergency plan: During the emergency treatment process, real-time monitoring of leakage and diffusion is carried out, and the emergency treatment plan is dynamically adjusted in combination with the latest prediction results and on-site data. The calculation formula is: ； in, It is an adjusted emergency plan. It is the current emergency plan. is the adjustment factor, This is a forecast based on the latest data. is the target control level.

7. The oil pipeline leakage detection method according to claim 1, characterized in that: The leakage alarm in S6 includes: S61, leakage detection: when a leak is detected in the oil pipeline, confirm the leak event according to a preset leak determination threshold; S62, triggering the alarm mechanism: if the leakage event is confirmed, the alarm mechanism is immediately triggered to generate an alarm signal, which is sent through a variety of means, including SMS, email, phone and emergency broadcast; S63, Notify relevant personnel: notify relevant emergency response personnel and management personnel through pre-set communication channels and provide information on the leak, including the location, volume and initial spread of the leak; S64, Dynamic Simulation and Visualization: Using dynamic simulation tools, based on real-time detection data and leakage path prediction results, simulate the diffusion path and range of oil spills; S65, Visualization Display: Through visualization tools, dynamic simulation results are displayed on a map or 3D model to intuitively show the diffusion path and affected area of ​​the oil spill. The visualization display includes the leakage point, diffusion path, affected area and key facilities.

8. A petroleum pipeline leakage detection system, used to implement the petroleum pipeline leakage detection method according to any one of claims 1 to 7, characterized in that: Includes the following modules: Data acquisition and preprocessing module: collects the detection data of the oil pipeline in real time through multiple sensors and geographic information acquisition equipment, including the pressure, flow and acoustic signal data inside the pipeline, as well as the terrain, meteorological conditions and soil properties around the pipeline, and preprocesses the collected detection data; Feature extraction module: uses a multi-scale feature extraction algorithm to extract features from the pre-processed detection data, including time features, frequency features, and spatial features; Model building module: Based on the extracted features, the oil pipeline status recognition model is constructed; Real-time detection module: inputs the real-time collected detection data into the oil pipeline status recognition model to perform real-time leakage detection, determine whether the oil pipeline has leaked, and adopts a dynamic threshold adjustment algorithm based on Bayesian reasoning to adapt to changes in pipeline operation status; Leakage path prediction module: Based on the detection data of oil pipelines and fluid dynamics models, it can predict the diffusion path and range of oil leakage in real time; Leakage alarm module: When a leak is detected in an oil pipeline, the alarm mechanism is triggered to notify relevant personnel to take emergency measures, and the diffusion path and scope of the oil leak are displayed through dynamic simulation and visualization tools.

Images