An early warning method for the stability of submarine scoured pipelines based on artificial intelligence

Abstract

The present invention discloses an artificial-intelligence-based early warning method for the stability of submarine scoured pipelines, which relates to the technical field of submarine pipeline protection. The historical re-survey data and monitoring data of submarine pipelines in the target area are collected and sorted out. Based on the Navier-Stokes equations and sediment transport models, a local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline is constructed. By integrating historical re-survey data, monitoring data and numerical model results, the present invention can comprehensively consider the coupling effects of various factors, and thus more comprehensively capture the dynamic characteristics of submarine scouring, so as to more accurately identify the key scouring factors affecting pipeline stability. At the same time, combining the scouring process and results simulated by the numerical model, it can more accurately analyze the stress distribution and deformation of the pipeline at different scouring depths, and then issue an early warning signal in advance, effectively reducing the risk of pipeline damage due to scouring and improving the reliability and accuracy of early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of submarine pipeline protection, and particularly relates to a method for warning the stability of submarine scoured pipelines based on artificial intelligence. Background Art

[0002] The estuary is the intersection area of rivers, seas, land and the atmosphere, and is an active place where the lithosphere, atmosphere, hydrosphere and biosphere interact. The delta has vast coastal wetlands with various wetland vegetation growing, and is an ideal habitat for migratory birds and fish migration, with rich species diversity. It is listed as the third largest ecosystem on the earth together with forests and oceans. The shallow submarine geology in the Yellow River subaqueous delta is complex, and submarine scouring is widely developed, seriously threatening the safety and stability of oil platforms, submarine pipelines and wharf embankments. Moreover, there are many submarine pipelines in this sea area, which are greatly affected by submarine scouring.

[0003] In the prior art, submarine scouring and pipeline stability are affected by the coupling of various environmental factors, and there are certain differences in geological conditions in different sea areas, resulting in difficulty in accurately analyzing the change trend of the stability of submarine scoured pipelines, and then accurately warning submarine scoured pipelines. Therefore, a method for warning the stability of submarine scoured pipelines based on artificial intelligence is proposed to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for warning the stability of submarine scoured pipelines based on artificial intelligence to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A method for warning the stability of submarine scoured pipelines based on artificial intelligence includes the following steps:

[0007] S1. Collect and sort out the historical re-survey data and monitoring data of submarine pipelines in the target area;

[0008] S2. Based on the Navier-Stokes equation and the sediment transport model, construct a local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, and input the monitoring data as boundary and constraint conditions into the model to simulate the dynamic process of submarine scouring;

[0009] S3. Use the historical re-survey data and the results of the numerical model to analyze the corresponding relationship between the pipeline occurrence state and submarine scouring, and then identify the key scouring factors and control mechanisms affecting pipeline stability;

[0010] S4. Train the AI large model by combining historical data and numerical model results, analyze the laws of pipeline scouring risks, and combine real-time monitoring data to judge the stress distribution and deformation conditions of the pipeline at different scouring depths;

[0011] S5. Based on the analysis results of pipeline scouring risks, set different levels of early warnings and match corresponding countermeasures.

[0012] A further improvement of the technical solution of the present invention is that: the historical re-survey data of the submarine pipeline is used to analyze the historical change laws of pipeline displacement and occurrence status (the occurrence status includes suspended position, suspended length, and burial depth);

[0013] The monitoring data includes the measured wave size, sea current size, and physical and mechanical property parameters of seabed sediments.

[0014] A further improvement of the technical solution of the present invention is that: the specific steps of S1 include:

[0015] Obtain the historical re-survey reports and records of the submarine pipeline in the target area from the database, ensure that the collected data is complete and accurate, covers a sufficient long time range for analyzing historical change laws, and classify and file the collected historical re-survey data;

[0016] For the classified historical re-survey data of the submarine pipeline, analyze the historical change laws of pipeline displacement and occurrence status. Among them, the pipeline displacement data is the displacement amount of the pipeline at different time points, including lateral displacement (the movement of the pipeline in the plane direction) and longitudinal displacement. Arrange the displacement data in chronological order, analyze its change trend, and calculate the displacement rate (the displacement amount per unit time). The pipeline occurrence status data includes suspended position, suspended length, and burial depth. Then, correspond the suspended position, suspended length, and burial depth data with time, analyze their change laws over time, and draw the change curves of suspended length and burial depth over time;

[0017] Deploy a sensor network and underwater monitoring equipment near the submarine pipeline in the target area to obtain monitoring data, monitor the wave size, sea current size, and physical and mechanical property parameters of seabed sediments, ensure the accuracy and reliability of the monitoring equipment, and collect the data transmitted by the sensor network and underwater monitoring equipment through the set data acquisition system;

[0018] Preprocess the collected data, including data cleaning and data standardization processing. Then, combine geographic information system technology to analyze the spatial distribution relationship between the pipeline occurrence status and seabed environmental factors;

[0019] Conduct trend analysis on the pipeline displacement. Taking time as the horizontal axis and the pipeline displacement as the vertical axis, plot the change curves of the horizontal and vertical displacements of the pipeline, and calculate the displacement rate of the pipeline in different time periods through curve fitting to analyze its acceleration or deceleration trend;

[0020] Conduct analysis on the variation law of the occurrence state, analyze the change trend of the suspended position, judge whether the suspended section is gradually expanding or shrinking, plot the change curve of the suspended length over time, calculate the change rate of the suspended length, analyze the change trend of the pipeline burial depth, judge whether the pipeline is gradually floating or subsiding, and combine the change laws of the pipeline displacement and the occurrence state to analyze their relationships with waves, ocean currents, and sediment characteristics to identify the main factors affecting the pipeline stability;

[0021] Sort out the historical change laws of the pipeline displacement and the occurrence state, as well as the results of the correlation analysis with environmental factors, and write a data collection and analysis report to store it in the database for subsequent analysis and processing.

[0022] A further improvement of the technical solution of the present invention lies in that: the S2 specifically includes:

[0023] Based on the Navier-Stokes equation, establish a three-dimensional hydrodynamic model to describe the velocity field, pressure field, and density field of the water flow in the ocean, where the Navier-Stokes equation is the Navier-Stokes equation;

[0024] Combine with the sediment transport equation to establish a sediment transport model to analyze the movement laws of sediment particles in the water flow, including two movement forms of suspended load and bed load. Introduce the Rouse equation to calculate the vertical distribution of suspended sediment concentration, and use the Meyer-Peter formula to estimate the bed load transport rate. Then couple the three-dimensional hydrodynamic model with the sediment transport model to form a complete three-dimensional numerical model, that is, the local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-seabed pipeline;

[0025] According to the seabed topography and pipeline position of the target area, use the finite difference method or the finite volume method to divide the grid of the target area to ensure that the grid can cover the seabed pipeline and its surrounding key areas, and set boundary conditions including water level, flow velocity, and wave period according to the seabed topography, pipeline position, and water flow direction;

[0026] The wave size, current size, and physical and mechanical property parameters of seabed sediments of the monitoring data are input into a three-dimensional numerical model as boundary conditions and constraint conditions, and the initial conditions of the model are set, including the initial topography and the initial water flow velocity distribution. Among them, the boundary conditions include wave boundaries, current boundaries, and sediment boundaries. The wave size data is used as the upper boundary condition of the model to simulate the influence of waves on the seabed water flow. The current size data is used as the side boundary condition to simulate the effect of currents on sediment transport. The physical and mechanical property parameter data of seabed sediments are used as the bottom boundary condition to simulate the physical and mechanical properties of seabed sediments;

[0027] According to the dynamic characteristics of water flow and sediment, the dynamic process of seabed erosion is gradually simulated by means of time stepping. And within each time step, the Navier-Stokes equation and the sediment transport equation are solved simultaneously to simulate the water flow movement and sediment transport process, and at the same time, the pipeline displacement and occurrence state data during the simulation process are recorded.

[0028] A further improvement of the technical solution of the present invention lies in that: in the three-dimensional numerical model, the coupling equation of water flow movement and sediment transport is expressed as:

[0029] ;

[0030] ;

[0031] In the formula, is the fluid density, u is the fluid velocity field, p is the fluid pressure, is the fluid dynamic viscosity, f is the external force (such as gravity, wave force, etc.), is the feedback force of sediment on water flow, C is the sediment concentration, D is the sediment diffusion coefficient, is the sediment source-sink term, indicating the sediment settlement or resuspension;

[0032] The sediment settlement rate is calculated by the Rouse equation, and its expression is as follows:

[0033] ;

[0034] In the formula, is the sediment settlement rate, is the sediment settlement velocity, K is the Rouse number, k is the von Kármán constant (about 0.41), is the water flow shear velocity.

[0035] A further improvement of the technical solution of the present invention lies in that: the specific content of S3 includes:

[0036] Extract the data on the occurrence state of the pipeline from the historical re-survey report, including information such as the suspended position, suspended length, burial depth, and displacement, and organize them in a time series to form time series data for facilitating the analysis of its variation law over time;

[0037] Analyze the seabed scouring situation. According to the terrain change data in the historical re-survey materials, analyze the depth and scope of the scour pit, as well as the variation trend of the scour area over time, and distinguish the seabed scouring types, including clear-water scouring and movable-bed scouring. Among them, clear-water scouring occurs at a relatively low flow velocity, and only a scour pit is formed near the pipeline, while movable-bed scouring occurs at a relatively high flow velocity, and the entire seabed will be scoured, and the scoured materials may fill the scour pit under the pipeline;

[0038] Based on the local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, simulate the flow field changes around the submarine pipeline, and combine with the sediment transport equation to simulate the movement laws of seabed sediment under the action of the flow field (bed load and suspended load). Through the numerical model to simulate the scouring process, obtain the variation trend of the scouring depth and scope over time, and then compare the simulation results with the historical re-survey data to verify the accuracy of the model;

[0039] Combined with the output of the numerical model, analyze the flow field characteristics, sediment movement laws, and pipeline force conditions in different scouring stages, and identify the key parameters affecting the scouring depth, including flow velocity, sediment particle size, pipeline diameter, and suspended amount. Among them, the scouring stages include intermittent scouring, wake scouring, and equilibrium scouring;

[0040] Through the comprehensive analysis of the historical re-survey data and the numerical model results, identify the key scouring factors affecting the pipeline stability, including flow velocity, sediment particle size, pipeline diameter, suspended amount, as well as seabed topography and soil properties, and formulate corresponding control measures for the identified key scouring factors. Specifically: improve the anti-scouring ability of the pipeline by adjusting the pipeline laying depth, increasing the density and burial depth of the guide frame, adding reinforcement rings, etc.; reduce the impact of scouring on the pipeline by measures such as manually placing sandbags and constructing breakwaters.

[0041] A further improvement of the technical solution of the present invention lies in: in the S4, the process of analyzing the law of the pipeline scouring risk includes:

[0042] Extract the data on seabed scouring and the occurrence state of the pipeline from the historical re-survey data, and extract the data on wave size, sea current size, and physical and mechanical property parameters of seabed sediments from the historical monitoring data;

[0043] Extract the simulation results from the established local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, including the scour depth, scope, flow field characteristics, and sediment movement law. Then, integrate the historical re-survey data, historical monitoring data, and numerical model results to obtain a complete input data set;

[0044] Combine the historical re-survey data and historical monitoring data to analyze historical scour events and pipeline damage conditions, determine the critical scour velocity and critical scour depth as the annotation basis, and annotate the input data set to form an annotated data set. Select the convolutional neural network model as the basic architecture of the AI large model to analyze the pipeline scour risk. Among them, the convolutional neural network model includes an input layer, a hidden layer, and an output layer. The input features of the input layer include flow velocity, sediment particle size, pipeline diameter, and suspension amount. The hidden layer uses multiple LSTMs to capture the long-term dependencies in time series data. Add a Dropout layer to the hidden layer to prevent overfitting. The output layer outputs the pipeline scour risk index;

[0045] Divide the annotated data set into a training set and a validation set. Use the training set to input the convolutional neural network model to train the AI large model. Adopt the Adam optimizer, set appropriate learning rates and decay strategies, evaluate the model performance through the validation set, and adjust the hyperparameters. During the training process, record the loss value of each epoch and the accuracy on the validation set to prevent overfitting. Adopt the K-fold cross-validation method to ensure the generalization ability of the model. According to the performance evaluation results on the validation set, adjust the hyperparameters of the model, including the learning rate, the number of hidden layer units, and the Dropout ratio. After the AI large model is trained, save the trained AI large model as a model file and deploy it to the server;

[0046] The AI large model combines and identifies the key parameters affecting the scour depth, including flow velocity, sediment particle size, pipeline diameter, and suspension amount, outputs the pipeline scour risk index, quantitatively evaluates the risk degree of the pipeline under different scour conditions, analyzes the law of pipeline scour risk, and judges the stress distribution and deformation conditions of the pipeline at different scour depths, and evaluates whether the pipeline is at risk of local instability or overall failure;

[0047] Comprehensively analyze the law of pipeline scour risk and real-time monitoring data to analyze the stability of the submarine scoured pipeline.

[0048] A further improvement of the technical solution of the present invention lies in: the process of obtaining the pipeline scour risk index is as follows:

[0049] Obtain the water flow velocity and scour depth at each time step from the numerical model, obtain the median sediment particle size through laboratory analysis, obtain the pipeline diameter and total length from the design document, and at the same time obtain the suspension length from the historical re-survey data, and integrate to obtain the scour risk sequence;

[0050] Calculate the critical scour velocity based on sediment particle size and soil type, and determine the critical scour depth based on pipeline burial depth and design requirements;

[0051] For each time step, the risk value is calculated by combining the parameters of the scour risk sequence with the critical scour velocity and critical scour depth. ;

[0052] The risk values of all time steps are averaged to obtain the final pipeline scour risk index.

[0053] A further improvement of the technical solution of the present invention is that: S5 specifically includes:

[0054] Combining historical data and analyzed pipeline scour risk rules, different warning levels are set based on historical pipeline scour risk indicators, namely low warning level, medium warning level and high warning level, and corresponding warning thresholds are assigned to each warning level;

[0055] Based on the sensor network and underwater monitoring equipment, real-time monitoring data near the submarine pipeline in the target area is obtained to monitor the wave size, current size and physical and mechanical characteristic parameters of the seabed sediment, and then the pipeline scour risk index is calculated to analyze the stability of the submarine scour pipeline;

[0056] Compare the pipeline scour risk index calculated through real-time monitoring data with the warning threshold to determine the current warning level and generate the corresponding warning signal;

[0057] The warning signal is delivered to the relevant responsible persons and management departments through multiple channels (SMS, email, monitoring system), and the content of the warning information is clarified, including the warning level, scope of impact and possible consequences, etc. According to the warning level, the corresponding response measures are matched to ensure the safe operation of the pipeline.

[0058] Due to the adoption of the above technical solution, the present invention has the following technical advances compared with the prior art:

[0059] The present invention provides an artificial intelligence-based submarine scour pipeline stability early warning method. By integrating historical resurvey data, monitoring data and numerical model results, it can comprehensively consider the coupling effect of multiple factors, and then more comprehensively capture the dynamic characteristics of submarine scour, so as to more accurately identify the key scour factors affecting pipeline stability. At the same time, combined with the scour process and results simulated by the numerical model, it can more accurately analyze the stress distribution and deformation of the pipeline at different scour depths, and then issue an early warning signal in advance, so as to buy more time for pipeline maintenance and repair, effectively reduce the risk of pipeline damage due to scour, and improve the reliability and accuracy of the early warning.

[0060] The present invention provides an artificial intelligence-based early warning method for the stability of submarine scoured pipelines. By deploying a sensor network and underwater monitoring equipment near the submarine pipeline, key data such as the size of waves, the size of ocean currents, and the physical and mechanical property parameters of seabed sediments can be obtained, the current scouring situation can be quickly analyzed, and compared with a preset early warning threshold to timely determine whether the stability of the pipeline is threatened, so as to issue an early warning signal earlier, enabling relevant departments and personnel to quickly take countermeasures to avoid or reduce pipeline damage and accidents caused by scouring and ensure the safe operation of the submarine pipeline. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0062] Figure 1 It is a schematic flowchart of the method of the present invention;

[0063] Figure 2 It is a schematic flowchart of the working process for analyzing the law of pipeline scouring risk of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0065] Embodiment 1, as Figure 1 shown, the present invention provides an artificial intelligence-based early warning method for the stability of submarine scoured pipelines, including the following steps:

[0066] S1. Collect and organize the historical re-survey data and monitoring data of the submarine pipeline in the target area. The historical re-survey data of the submarine pipeline is used to analyze the historical variation laws of pipeline displacement and occurrence status (the occurrence status includes the suspended position, suspended length, and burial depth). The monitoring data includes the measured wave height, sea current magnitude, and physical and mechanical property parameters of the seabed sediment. Obtain the historical re-survey reports and records of the submarine pipeline in the target area from the database to ensure that the collected data is complete, accurate, and covers a sufficient long time range for analyzing the historical variation laws. Classify and file the collected historical re-survey data of the submarine pipeline. For the classified historical re-survey data of the submarine pipeline, analyze the historical variation laws of pipeline displacement and occurrence status. Among them, the pipeline displacement data is the displacement amount of the pipeline at different time points, including lateral displacement (the movement of the pipeline in the plane direction) and longitudinal displacement. Arrange the displacement data in chronological order, analyze its change trend, and calculate the displacement rate (the displacement amount per unit time). The pipeline occurrence status data includes the suspended position, suspended length, and burial depth. The suspended position is the specific position coordinates of the suspended pipeline recorded. The suspended length is the measured and recorded length of the suspended section. The burial depth is the burial depth data of the pipeline at different positions recorded, including the distance from the top of the pipeline to the seabed surface. Then, correspond the suspended position, suspended length, and burial depth data with time, analyze their variation laws with time, and draw the variation curves of the suspended length and burial depth with time. Deploy a sensor network and underwater monitoring equipment near the submarine pipeline in the target area to obtain monitoring data, monitor the wave height, sea current magnitude, and physical and mechanical property parameters of the seabed sediment, ensure the accuracy and reliability of the monitoring equipment, and collect the data transmitted by the sensor network and underwater monitoring equipment through the set data acquisition system. Among them, the wave height is collected by the wave monitoring buoy installed in the target area, record the period, wave height (the distance from the wave crest to the wave trough), and wavelength of the wave, and statistically analyze the average value, maximum value, and minimum value of the wave height to analyze the seasonal variation and extreme events of the wave. The sea current magnitude is collected by a current meter (acoustic Doppler current profiler), record the velocity magnitude and flow direction of the sea current to analyze the spatio-temporal variation laws of the sea current, and then draw the distribution map of the sea current velocity with time and space. Obtain the physical and mechanical property parameters of the seabed sediment through geological sampling and laboratory tests. The physical properties include the particle size distribution, density, porosity, etc. of the sediment, and the mechanical properties include the shear strength, internal friction angle, compression modulus, etc. of the sediment. Correspond the physical and mechanical property parameters of the sediment with the sampling position and time, analyze their spatial distribution laws and variation with time. Preprocess the collected data, including data cleaning and data standardization processing. Among them, conduct a quality check on the collected data, remove the obvious error or abnormal data points, and fill in the missing data by interpolation method. Convert all data to a unified unit for subsequent analysis. Normalize the data so that its range is between 0 and 1 to avoid the influence of data with different dimensions on the analysis results. Then, combine with geographic information system technology,Analyze the spatial distribution relationship between the pipeline occurrence state and seabed environmental factors, conduct a trend analysis of pipeline displacement. With time as the horizontal axis and pipeline displacement as the vertical axis, plot the change curves of the pipeline's horizontal and vertical displacements, calculate the displacement rate of the pipeline in different time periods through curve fitting, analyze its acceleration or deceleration trend, analyze the variation law of the occurrence state, analyze the variation trend of the suspended position, judge whether the suspended section is gradually expanding or shrinking, plot the change curve of the suspended length over time, calculate the change rate of the suspended length, analyze the variation trend of the pipeline burial depth, judge whether the pipeline is gradually floating or sinking. Combine the change laws of pipeline displacement and occurrence state, analyze their relationships with waves, ocean currents, and sediment characteristics, identify the main factors affecting pipeline stability, sort out the historical change laws of pipeline displacement and occurrence state, as well as the correlation analysis results with environmental factors, and write a data collection and analysis report, which is stored in the database for subsequent analysis and processing;

[0067] S2. Based on the Navier-Stokes equation and the sediment transport model, a local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline is constructed. The monitoring data is input into the model as boundary and constraint conditions to simulate the dynamic process of seabed scouring. Based on the Navier-Stokes equation, a three-dimensional hydrodynamic model is established to describe the velocity field, pressure field, and density field of the water flow in the ocean. Among them, the Navier-Stokes equation is the Navier-Stokes equation. Combining with the sediment transport equation, a sediment transport model is established to analyze the movement law of sediment particles in the water flow, including two movement forms of suspended sediment and bed load. The Rouse equation is introduced to calculate the vertical distribution of suspended sediment concentration, and the Meyer-Peter formula is used to estimate the bed load transport rate. Then, the three-dimensional hydrodynamic model and the sediment transport model are coupled to form a complete three-dimensional numerical model, which is the local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline. By solving the Navier-Stokes equation and the sediment transport equation, the dynamic process of seabed scouring is simulated, including water flow movement, sediment transport, and terrain change. According to the seabed topography and pipeline position in the target area, the target area is meshed using the finite difference method or the finite volume method to ensure that the mesh can cover the submarine pipeline and its surrounding key areas. And according to the seabed topography, pipeline position, and water flow direction, boundary conditions including water level, flow velocity, and wave period are set. The wave size, sea current size, and physical and mechanical property parameters of seabed sediments in the monitoring data are input into the three-dimensional numerical model as boundary conditions and constraint conditions, and the initial conditions of the model are set, including the initial topography and the initial water flow velocity distribution. Among them, the boundary conditions include wave boundary, sea current boundary, and sediment boundary. The wave size data is used as the upper boundary condition of the model to simulate the influence of waves on the seabed water flow. The sea current size data is used as the side boundary condition to simulate the effect of sea currents on sediment transport. The physical and mechanical property parameter data of seabed sediments are used as the bottom boundary condition to simulate the physical and mechanical properties of seabed sediments. According to the dynamic characteristics of water flow and sediment, the dynamic process of seabed scouring is gradually simulated by the time-stepping method. And in each time step, the Navier-Stokes equation and the sediment transport equation are solved simultaneously to simulate the water flow movement and sediment transport process, and at the same time, the pipeline displacement and storage state data during the simulation process are recorded;

[0068] In the three-dimensional numerical model, the coupling equations of water flow movement and sediment transport are expressed as:

[0069] ;

[0070] ;

[0071] In the formula, is the fluid density, u is the fluid velocity field, p is the fluid pressure, is the hydrodynamic viscosity, and the dynamic viscosity of seawater is approximately 1.002×10⁻³ Pa·s. f is the external force (such as gravity, wave force, etc.). is the feedback force of sediment on the water flow. Considering the influence of sediment movement on the water flow, C is the sediment concentration, and the suspended sediment concentration ranges from 10⁻³ - 10 kg / m³. D is the sediment diffusion coefficient, which describes the diffusion ability of sediment in the water flow and is between 10⁻ 4 -10⁻² m² / s. The specific value depends on the water flow turbulence intensity. In the high-turbulence region, the diffusion coefficient is larger. is the sediment source-sink term, representing the sediment settlement or resuspension;

[0072] The sediment settlement rate is calculated by the Rouse equation, and its expression is as follows:

[0073] ;

[0074] In the formula, is the sediment settlement rate, is the sediment settlement velocity, which is between 10⁻ 4 -10⁻¹ m / s. K is the Rouse number, and k is the von Kármán constant (about 0.41). is the water flow shear velocity;

[0075] S3. Utilize historical re-survey data and numerical model results to analyze the corresponding relationship between the pipeline occurrence state and seabed scouring, and then identify the key scouring factors and control mechanisms affecting pipeline stability. Extract the pipeline occurrence state data from the historical re-survey report, including information such as the suspended position, suspended length, burial depth, and displacement, and organize them in a time series to form time series data for facilitating the analysis of their variation patterns over time. Analyze the seabed scouring situation. Based on the terrain change data in the historical re-survey data, analyze the depth and extent of the scour pit, as well as the variation trend of the scouring area over time, and distinguish the types of seabed scouring, including clear-water scouring and mobile-bed scouring. Among them, clear-water scouring occurs at relatively low flow velocities and only forms a scour pit near the pipeline, while mobile-bed scouring occurs at relatively high flow velocities and the entire seabed will be scoured, and the scoured materials may fill the scour pit under the pipeline. Based on the local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, simulate the flow field changes around the submarine pipeline, and combine with the sediment transport equation to simulate the movement laws of seabed sediment under the action of the flow field (bed load and suspended load). Through the numerical model to simulate the scouring process, obtain the variation trend of the scouring depth and extent over time, and then compare the simulation results with the historical re-survey data to verify the accuracy of the model. Combine the output of the numerical model to analyze the flow field characteristics, sediment movement laws, and pipeline stress conditions in different scouring stages, and identify the key parameters affecting the scouring depth, including flow velocity, sediment particle size, pipeline diameter, and suspended amount. Among them, the scouring stages include intermittent scouring, wake scouring, and equilibrium scouring. Intermittent scouring means that scouring mainly occurs in the gap area between the pipeline and the seabed. Wake scouring means that a wake area is formed downstream of the pipeline, resulting in local scouring. Equilibrium scouring means that the scouring reaches a dynamic equilibrium and the scouring depth and extent tend to be stable. Flow velocity indicates that the water flow velocity is one of the key factors affecting the scouring depth, and high flow velocity will intensify scouring. Sediment particle size indicates that the smaller the sediment particle size, the easier it is to be carried away by the water flow, resulting in intensified scouring. Pipeline diameter indicates that the pipeline diameter affects the flow field distribution around it, and thus affects the scouring depth. Suspended amount means that the pipeline suspension will change the water flow direction and velocity, intensifying local scouring. Through the comprehensive analysis of historical re-survey data and numerical model results, identify the key scouring factors affecting pipeline stability, including flow velocity, sediment particle size, pipeline diameter, suspended amount, as well as seabed topography and soil properties, and formulate corresponding control measures for the identified key scouring factors. Specifically: improve the anti-scouring ability of the pipeline by adjusting the pipeline laying depth, increasing the density and burial depth of the guide frame, adding reinforcement rings, etc.; reduce the impact of scouring on the pipeline by measures such as manually placing sandbags and building breakwaters;

[0076] S4. Combine historical data and numerical model results to train the AI large model, analyze the laws of pipeline scouring risks, and combine real-time monitoring data to judge the stress distribution and deformation conditions of the pipeline at different scouring depths;

[0077] S5. Based on the analysis results of the pipeline scouring risk, set different levels of early warnings and match corresponding countermeasures.

[0078] Example 2. As Figure 2 shown, based on Example 1, the present invention provides a technical solution: Preferably, in S4, the process of analyzing the law of pipeline scouring risk includes:

[0079] Extract data on seabed scouring and pipeline occurrence status from historical re-survey data, extract data on wave size, sea current size, and physical and mechanical property parameters of seabed sediments from historical monitoring data, and extract simulation results from the established local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, including scouring depth, range, flow field characteristics, and sediment movement law. Then, integrate the historical re-survey data, historical monitoring data, and numerical model results to obtain a complete input data set. Combine the historical re-survey data and historical monitoring data to analyze historical scouring events and pipeline damage conditions, determine the critical scouring velocity and critical scouring depth as the annotation basis, and annotate the input data set to form an annotated data set. Select a convolutional neural network model as the basic architecture of the AI large model to analyze the pipeline scouring risk. Among them, the convolutional neural network model includes an input layer, a hidden layer, and an output layer. The input features of the input layer include flow velocity, sediment particle size, pipeline diameter, and suspension amount. The hidden layer uses multiple LSTMs to capture long-term dependencies in time series data. Add a Dropout layer to the hidden layer to prevent overfitting. The output layer outputs the pipeline scouring risk index. Divide the annotated data set into a training set and a validation set. Use the training set to input the convolutional neural network model to train the AI large model. Adopt the Adam optimizer, set appropriate learning rates and decay strategies, evaluate the model performance through the validation set, and adjust the hyperparameters. During the training process, record the loss value of each epoch and the accuracy on the validation set to prevent overfitting. Adopt the K-fold cross-validation method to ensure the generalization ability of the model. According to the performance evaluation results on the validation set, adjust the hyperparameters of the model, including learning rate, number of hidden layer units, and Dropout ratio. After the AI large model training is completed, save the trained AI large model as a model file and deploy it to the server. The AI large model combines and identifies the key parameters affecting the scouring depth, including flow velocity, sediment particle size, pipeline diameter, and suspension amount, outputs the pipeline scouring risk index, quantitatively evaluates the risk degree of the pipeline under different scouring conditions, analyzes the law of pipeline scouring risk, and judges the stress distribution and deformation conditions of the pipeline at different scouring depths, evaluates the risk of local instability or overall failure of the pipeline, and comprehensively analyzes the stability of the seabed scouring pipeline based on the law of pipeline scouring risk and real-time monitoring data;

[0080] The process of obtaining the pipeline scouring risk index is as follows:

[0081] Obtain the water flow velocity and scour depth for each time step from the numerical model, obtain the median sediment grain size through laboratory analysis, obtain the pipeline diameter and total length from the design documents, and obtain the suspended length from the historical re-survey data. Integrate to obtain the scour risk sequence. Calculate the critical scour velocity based on the sediment grain size and soil type, and determine the critical scour depth through the pipeline burial depth and design requirements. For each time step, calculate the risk value by integrating the parameters of the scour risk sequence, the critical scour velocity, and the critical scour depth. ;

[0082] Average the risk values for all time steps to obtain the final pipeline scour risk index;

[0083] The expression of the pipeline scour risk index is as follows:

[0084] ;

[0085] ;

[0086] In the formula, R is the pipeline scour risk index, used to evaluate the risk degree of the pipeline under different scour conditions. is the risk value, N is the number of time steps. is the water flow velocity at the i-th time step. is the critical scour velocity, that is, the minimum velocity at which scour begins. is the median sediment grain size, representing the representative value of the sediment grain size. L is the pipeline diameter, which affects the flow field distribution around the pipeline. is the suspended length, that is, the length of the suspended part of the pipeline. is the total length of the pipeline, used to normalize the suspended length. is the scour depth at the i-th time step. is the critical scour depth, that is, the scour depth at which significant risks begin to appear in the pipeline. The value of ranges from 0.5 - 3.0 m / s. Depends on the sediment grain size and soil type, generally in the range of 0.5 - 1.5 m / s. The higher the flow velocity, the greater the scour risk. The value of ranges from 0.05 - 2.0 mm. The smaller the grain size, the higher the scour risk. L is in the range of 0.3 - 1.2 m. The larger the pipeline diameter, the relatively lower the scour risk. The higher the value of R, the greater the risk of the pipeline under the current scour conditions.

[0087] S5 specifically includes:

[0088] Combined with the law of pipeline scouring risk obtained from historical data and analysis, different warning levels are set based on historical pipeline scouring risk indicators, namely low warning level, medium warning level and high warning level, and corresponding warning thresholds are assigned to each warning level. Based on the sensor network and underwater monitoring equipment, real-time monitoring data near the submarine pipeline in the target area are obtained, the wave size, sea current size and physical and mechanical property parameters of seabed sediments are monitored, and then the pipeline scouring risk indicator is calculated to analyze the stability of the scoured submarine pipeline. The pipeline scouring risk indicator calculated from the real-time monitoring data is compared with the warning threshold to determine the current warning level, and then the corresponding warning signal is generated. The warning signal is transmitted to the relevant responsible persons and management departments through multiple channels (text messages, emails, monitoring systems), and the content of the warning information is clarified, including the warning level, the affected range and possible consequences, etc. According to the warning level, corresponding response measures are matched to ensure the safe operation of the pipeline. For the low warning level, strengthen daily inspections and regularly check the pipeline scouring situation. For the medium warning level, increase the inspection frequency, reinforce the pipeline, increase the pipeline wall thickness, set up casing, etc. For the high warning level, immediately start the emergency plan, close the control valves upstream and downstream of the pipeline, organize the repair team to quickly rush to the scene for emergency repair.

[0089] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all of them should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An early warning method for the stability of submarine scoured pipelines based on artificial intelligence, characterized in that, It includes the following steps: S1. Collect and collate the historical re-survey data and monitoring data of the submarine pipeline in the target area; S2. Based on the Navier-Stokes equations and the sediment transport model, construct a local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, and input the monitoring data as boundary and constraint conditions into the model to simulate the dynamic process of seabed scouring. Specifically, S2 includes: Based on the Navier-Stokes equations, establish a three-dimensional hydrodynamic model to describe the velocity field, pressure field and density field of the water flow in the ocean. Among them, the Navier-Stokes equations are the Navier-Stokes equations; Combine the sediment transport equations to establish a sediment transport model, analyze the movement laws of sediment particles in the water flow, including two movement forms of suspended sediment and bed load, introduce the Rouse equation to calculate the vertical distribution of suspended sediment concentration, and use the Meyer-Peter formula to estimate the bed load sediment transport rate. Then couple the three-dimensional hydrodynamic model with the sediment transport model to form a complete three-dimensional numerical model, that is, the local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline; According to the seabed topography and pipeline position in the target area, divide the grid of the target area, and set boundary conditions including water level, flow velocity and wave period according to the seabed topography, pipeline position and water flow direction; Input the wave size, sea current size and physical and mechanical property parameters of seabed sediments in the monitoring data as boundary conditions and constraint conditions into the three-dimensional numerical model, and set the initial conditions of the model, including the initial topography and the initial water flow velocity distribution. Among them, the boundary conditions include wave boundary, sea current boundary and sediment boundary. Use the wave size data as the upper boundary condition of the model to simulate the influence of waves on the seabed water flow, use the sea current size data as the side boundary condition to simulate the effect of sea currents on sediment transport, and use the physical and mechanical property parameter data of seabed sediments as the bottom boundary condition to simulate the physical and mechanical properties of seabed sediments; According to the dynamic characteristics of water flow and sediment, gradually simulate the dynamic process of seabed scouring by means of time stepping. And in each time step, solve the Navier-Stokes equations and the sediment transport equations simultaneously to simulate the water flow movement and sediment transport process, and record the pipeline displacement and storage state data during the simulation process; S3. Use the historical re-survey data and the results of the numerical model to analyze the corresponding relationship between the pipeline storage state and seabed scouring, and then identify the key scouring factors and control mechanisms affecting pipeline stability; S4. Train the AI large model by combining historical data and the results of the numerical model, analyze the laws of pipeline scouring risks, and combine real-time monitoring data to judge the stress distribution and deformation conditions of the pipeline at different scouring depths; S5. Based on the analysis results of pipeline scouring risks, set different levels of early warnings and match corresponding countermeasures.

2. The method for warning the stability of a submarine scouring pipeline based on artificial intelligence according to claim 1, wherein: The historical re-survey data of the submarine pipeline is used to analyze the historical change laws of pipeline displacement and storage state; The monitoring data includes the measured wave size, sea current size and physical and mechanical property parameters of seabed sediments.

3. The method for warning the stability of a submarine scouring pipeline based on artificial intelligence according to claim 2, wherein: Specifically, S1 includes: Obtain the historical re-survey reports and records of the submarine pipeline in the target area from the database, and classify and file the collected historical re-survey materials; For the classified historical re-survey materials of submarine pipelines, analyze the historical variation laws of pipeline displacement and occurrence status. Among them, the pipeline displacement data is the displacement amount of the pipeline at different time points, including lateral displacement and longitudinal displacement. Arrange the displacement data in chronological order, analyze its variation trend, calculate the displacement rate. The pipeline occurrence status data includes the suspended position, suspended length and burial depth. Then, correspond the suspended position, suspended length and burial depth data with time, analyze their variation laws with time, and draw the variation curves of suspended length and burial depth with time; Deploy a sensor network and underwater monitoring equipment near the submarine pipeline in the target area to obtain monitoring data, monitor the wave size, sea current size and physical and mechanical property parameters of seabed sediments, and collect the data transmitted by the sensor network and underwater monitoring equipment through the set data acquisition system; Preprocess the collected data, including data cleaning and data standardization processing. Then, combined with geographic information system technology, analyze the spatial distribution relationship between the pipeline occurrence status and seabed environmental factors; Conduct a trend analysis of pipeline displacement. Take time as the horizontal axis and pipeline displacement amount as the vertical axis, draw the variation curves of pipeline horizontal displacement and vertical displacement, and calculate the displacement rate of the pipeline in different time periods through curve fitting, and analyze its acceleration or deceleration trend; Conduct a variation law analysis of the occurrence status, analyze the variation trend of the suspended position, judge whether the suspended section is gradually expanding or shrinking, draw the variation curve of the suspended length with time, calculate the variation rate of the suspended length, analyze the variation trend of the pipeline burial depth, judge whether the pipeline is gradually floating or sinking, and combine the variation laws of pipeline displacement and occurrence status to analyze their relationships with waves, sea currents and sediment characteristics, and identify the main factors affecting pipeline stability; Sort out the historical variation laws of pipeline displacement and occurrence status, as well as the results of correlation analysis with environmental factors, and write a data collection and analysis report, which is stored in the database.

4. A method for warning the stability of a submarine scouring pipeline based on artificial intelligence according to claim 3, characterized in that: In the three-dimensional numerical model, the coupling equation of water flow movement and sediment transport is expressed as: ； ； In the formula, is the fluid density, u is the fluid velocity field, p is the fluid pressure, is the dynamic viscosity of the fluid, f is the external force, is the feedback force of sediment on the water flow, C is the sediment concentration, D is the sediment diffusion coefficient, is the sediment source-sink term; The sediment settlement rate is calculated by the Rouse equation, and its expression is as follows: ; In the formula, is the sediment settlement rate, is the sediment settlement velocity, K is the Rouse number, is the von Kármán constant, is the flow shear velocity.

5. The stability warning method for submarine scouring pipelines based on artificial intelligence according to claim 4, characterized in that: The specific content of S3 includes: Extract the occurrence status data of the pipeline from the historical re-survey report, including the suspended position, suspended length, burial depth and displacement information, and sort it in time series to form time series data; Analyze the seabed scouring situation. According to the terrain change data in the historical re-survey materials, analyze the depth and scope of the scour pit, as well as the variation trend of the scour area with time, and distinguish the seabed scour types, including clear water scour and movable bed scour; Based on the local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, simulate the flow field change around the submarine pipeline, and combined with the sediment transport equation, simulate the movement law of seabed sediment under the action of the flow field. Through the numerical model to simulate the scouring process, obtain the variation trend of the scouring depth and scope with time; Combined with the output of the numerical model, analyze the flow field characteristics, sediment movement laws, and pipeline stress conditions in different scouring stages, and identify the key parameters affecting the scouring depth, including flow velocity, sediment particle size, pipeline diameter, and suspension amount. Among them, the scouring stages include intermittent scouring, wake scouring, and equilibrium scouring; Through the comprehensive analysis of historical re-survey data and numerical model results, identify the key scouring factors affecting pipeline stability, including flow velocity, sediment particle size, pipeline diameter, suspension amount, as well as seabed topography and soil properties, and formulate corresponding control measures for the identified key scouring factors.

6. The method for warning the stability of a submarine scouring pipeline based on artificial intelligence according to claim 5, wherein: In S4, the process of analyzing the law of pipeline scouring risk includes: Extract data on seabed scouring and pipeline occurrence status from historical re-survey data, and extract data on wave size, sea current size, and physical and mechanical property parameters of seabed sediments from historical monitoring data; Extract simulation results from the established local fine three-dimensional hydrodynamic-sediment numerical model of the seabed-submarine pipeline, including scouring depth, range, flow field characteristics, and sediment movement laws, and then integrate the historical re-survey data, historical monitoring data, and numerical model results to obtain a complete input data set; Combined with historical re-survey data and historical monitoring data, analyze historical scouring events and pipeline damage conditions, determine the critical scouring velocity and critical scouring depth as the annotation basis, and annotate the input data set to form an annotated data set. Select the convolutional neural network model as the basic architecture of the AI large model to analyze the pipeline scouring risk. Among them, the convolutional neural network model includes an input layer, a hidden layer, and an output layer. The input features of the input layer include flow velocity, sediment particle size, pipeline diameter, and suspension amount, and the output layer outputs the pipeline scouring risk index; Divide the annotated data set into a training set and a validation set, use the training set to input the convolutional neural network model to train the AI large model, adjust the hyperparameters of the model according to the performance evaluation results on the validation set. After the AI large model is trained, save the trained AI large model as a model file and deploy it to the server; The AI large model combines the identified key parameters affecting the scouring depth, including flow velocity, sediment particle size, pipeline diameter, and suspension amount, outputs the pipeline scouring risk index, quantitatively evaluates the risk degree of the pipeline under different scouring conditions, analyzes the law of pipeline scouring risk, and judges the stress distribution and deformation conditions of the pipeline at different scouring depths, and evaluates the risk of local instability or overall failure of the pipeline; Comprehensively analyze the law of pipeline scouring risk and real-time monitoring data to analyze the stability of the scoured submarine pipeline.

7. The method for early warning of the stability of a submarine scouring pipeline based on artificial intelligence according to claim 6, characterized in that: The process of obtaining the pipeline scouring risk index is: Obtain the water flow velocity and scouring depth at each time step from the numerical model, obtain the median sediment particle size through laboratory analysis, obtain the pipeline diameter and total length from the design document, and obtain the suspension length from the historical re-survey data at the same time, and integrate to obtain the scouring risk sequence; Calculate the critical scouring velocity according to the sediment particle size and soil type, and determine the critical scouring depth through the pipeline burial depth and design requirements; For each time step, calculate the risk value by synthesizing the parameters of the scour risk sequence, the critical scour velocity, and the critical scour depth ; Average the risk values of all time steps to obtain the final pipeline scouring risk index.

8. The method for warning the stability of a submarine scoured pipeline based on artificial intelligence according to claim 7, characterized in that: S5 specifically includes: Combined with the laws of pipeline scour risk obtained from historical data and analysis, different warning levels are set based on historical pipeline scour risk indicators, namely low warning level, medium warning level, and high warning level, and corresponding warning thresholds are assigned to each warning level; Based on the sensor network and underwater monitoring equipment, real-time monitoring data near the submarine pipeline in the target area is obtained, the wave size, sea current size, and physical and mechanical property parameters of the seabed sediment are monitored, and then the pipeline scour risk indicator is calculated to analyze the stability of the scoured submarine pipeline; The pipeline scour risk indicator calculated from the real-time monitoring data is compared with the warning threshold to determine the current warning level, and then the corresponding warning signal is generated; The warning signal is transmitted to the relevant responsible persons and management departments through multiple channels, the content of the warning information is clarified, including the warning level, the affected area, and possible consequences, etc., and corresponding response measures are matched according to the warning level.

Images