Floating installation-oriented LMU load real-time inversion method

Abstract

The invention provides an LMU load real-time inversion method for floating support installation. The method comprises the following steps: constructing an initial physical diagram model of a floating support installation structure; collecting environmental data in real time, fusing the environmental data with sensor data, and generating a dynamic graph based on a time step t; designing a load inference model based on a graph neural network, and forming initial load distribution of a whole graph based on the inversion load of each node; designing physical constraints to correct and optimize the initial load distribution to obtain optimized load distribution; and based on a Bayesian inference method, carrying out uncertainty analysis on a load result according to the optimized load distribution and the dynamic graph based on the time step t. According to the method, high-precision real-time inversion, robustness improvement and credibility guarantee of the floating support installation load are realized, and a comprehensive and reliable technical solution is provided for floating support installation in a complex marine environment.

Description

Technical Field

[0001] The invention belongs to the field of floating installation, and in particular relates to a real-time inversion method of LMU load for floating installation. Background Art

[0002] Floating installation technology is an important engineering technology in the field of marine engineering, and is widely used in the installation process of platform foundations, conductor frames and floating structures. With the deepening of marine development, the complexity of floating installation operations is increasing. Especially in deep-sea environments, floating structures will be affected by the combined effects of multiple marine environmental loads such as wind, waves, and currents, which makes the real-time monitoring and inversion of loads during their installation process particularly critical. The distribution of loads not only affects the safety and stability of floating structures, but is also directly related to the efficiency of installation operations. However, due to the complexity and real-time changing characteristics of the marine environment, existing load inversion methods face the following challenges: During the floating installation process, environmental loads are extremely dynamic, such as waves, tidal currents, and wind speed changes. These dynamic changes will directly affect the mechanical properties of the floating structure, resulting in drastic changes in load distribution in time and space. However, traditional load inversion methods usually rely on static models or simplified dynamic models, which are difficult to accurately describe the dynamic load distribution in complex environments, resulting in poor accuracy and real-time performance of the inversion results.

[0003] Load inversion for floating installations requires the combination of multiple types of sensor data, such as stress, strain, acceleration, displacement, etc. However, these sensors are susceptible to noise interference in actual marine environments, such as instrument errors, signal attenuation, and environmental noise (such as wave noise), resulting in reduced data quality. Existing methods lack effective noise modeling and data fusion mechanisms when dealing with load inversion in multi-noise environments, and the credibility of the inversion results is difficult to guarantee.

[0004] During the floating installation process, real-time monitoring and inversion are important means to ensure the safety of the project. However, due to the high computational complexity of marine engineering, traditional load inversion methods often rely on offline analysis or centralized computing modes, which makes it difficult to meet real-time requirements. In particular, when facing sudden changes in the marine environment (such as large wave impacts), the response speed of existing methods is not enough to provide timely support for installation operations.

[0005] In complex marine environments, any load inversion results inevitably have a certain degree of uncertainty. However, traditional methods mostly focus on the output of a single result and lack uncertainty analysis and quantitative evaluation of the inversion results. This defect not only affects the reliability of the inversion results, but may also lead to potential risks being ignored during the installation process. Summary of the invention

[0006] The purpose of the present invention is to propose a real-time inversion method for LMU loads installed on a floating platform. Through a real-time inversion method for LMU (Load Monitoring Unit) installed on a floating platform based on dynamic weight physical drive and Bayesian inference, a systematic optimization is carried out on the core problems in the prior art.

[0007] In order to achieve the above object, the present invention provides a real-time inversion method for LMU loads installed in a floating manner, the method comprising: Constructing an initial physical graph model of the floating installation structure, wherein the initial physical graph model is formed by modeling each sensor position in the floating structure as a graph node, modeling the physical mechanical relationship as a graph edge, and generating an initial physical graph based on graph node features and edge weights, and performing physical verification on the initial physical graph to generate a complete initial physical graph; Collect environmental data in real time, fuse the environmental data with sensor data, dynamically update the complete initial physical map, and generate a dynamic map based on time step t; wherein the environmental data includes current velocity, wave pressure and wind speed; Design a load inference model based on graph neural network, use the load inference model to perform load inversion calculation on the dynamic graph, obtain the inversion load of each graph node, and form a preliminary load distribution of the whole graph based on the inversion load of each graph node; Designing physical constraints to modify and optimize the preliminary load distribution to obtain an optimized load distribution; wherein the physical constraints include global force balance, local smoothness, and time dynamic consistency; Based on the Bayesian inference method, uncertainty analysis of the load results is performed according to the optimized load distribution and the dynamic graph based on time step t, and the load mean and variance of each graph node are obtained to describe the credibility of the load inversion results. At the same time, the uncertainty distribution map of the entire graph is generated in combination with the graph node position, and high-risk areas are marked to provide engineering optimization and decision-making for floating installation.

[0008] Furthermore, the initial physical graph model is constructed as follows: Define each sensor location in the floating structure as a graph node of the physical graph model , graph node The initial eigenvector of Contains the following data: initial measurement value of sensor i, physical position coordinates of sensor i, theoretical load value calculated according to the initial mechanical model of the floating structure; the theoretical load value calculated according to the initial mechanical model of the floating structure It is obtained by calculating the theoretical load value applied to the nodes of the graph; side Obtained through modeling of physical and mechanical relationships; According to the graph node and edge Definition of , generating the initial physical diagram , where the node set Including all sensor locations, each graph node has the feature ; Edge Set Represents a physical connection, each edge has an initial weight ; represents the initial edge weight set; Adjacency Matrix Represents the initial graph structure, where the edge weights of the adjacency matrix .

[0009] Furthermore, the overall equilibrium error is calculated through error analysis based on the total connection force and theoretical load value of the graph nodes; If the overall balance error exceeds the preset threshold, the initial weights are optimized according to the following rules: : Increase the connection weights between graph nodes associated with high force values; Adjust the connectivity of isolated graph nodes so that they are connected to the nearest high-weight graph nodes.

[0010] Furthermore, after collecting environmental data, we use the graph node Location And the spatial distribution of environmental data, environmental force mapping is performed to determine the local environmental force of each graph node .

[0011] Furthermore, the environmental data is fused with the sensor data, and the complete initial physical map is dynamically updated to generate a dynamic map based on the time step t, specifically including: The real-time interaction force and the physical regularization term are integrated into the edge weight to update the edge weight; wherein the physical regularization term is used to limit the deviation between the edge weight and the actual environment force; Among them, during the edge weight update process, the dynamically adjusted weight balance coefficient is calibrated by fitting the noise level and historical data; in, Each graph node feature is updated under the influence of the dynamic environment, which consists of two parts: basic feature enhancement and neighborhood aggregation: Basic feature enhancement: Modify the graph node features of the initial graph nodes in combination with environmental forces; Neighborhood aggregation update: aggregate neighborhood graph node features, with weights determined by updated edge weights; The final output time step Dynamic image ,in, is a node set, For edge sets, is the edge weight set.

[0012] Furthermore, before designing a load inference model based on a graph neural network, a node information propagation mechanism is designed to correct and propagate updates of graph node information, including: Graph node feature propagation is achieved through dynamic neighborhood aggregation; At the same time, during the process of graph node information propagation, a residual correction term is designed based on the error between the total environmental force and the total connection force of the real-time graph nodes to correct the graph node feature update.

[0013] Furthermore, the load inference model based on graph neural network is constructed as follows: Load inference is achieved through a multi-layer perceptron, which uses the corrected graph node features to output the load value of the current graph node; wherein the multi-layer perceptron includes a two-layer fully connected network, and its parameters are trained through a supervised learning method; At the same time, a total force balance constraint and a local gradient smoothing constraint are designed to ensure the physical consistency of the inversion results. The total force balance constraint forces the total load of the graph node to be consistent with the environmental force, and the local gradient smoothing constraint limits the drastic changes between the loads of adjacent graph nodes.

[0014] Furthermore, the design physical constraints modify and optimize the preliminary load distribution to obtain an optimized load distribution, specifically including: The global force balance check correction: Inverse load value for each graph node Should be equal to the total environmental force on the current graph node If they are consistent, then the balance error is defined ,when Exceeds the set tolerance threshold When the inverse load value is Error correction is performed by combining dynamically adjusted step-size weighted balancing errors; The local gradient smoothing check correction: Adjacent graph nodes and The load gradient satisfies ,in is the maximum allowed gradient, Adjacent graph nodes and The dynamic correlation weight of When , the neighborhood weighted average method is used for correction; in, The inverted load value of each graph node is updated using the gradient descent method; Iterate the optimization until the following convergence conditions are met: Balanced error of all graph nodes satisfy ; The rate of change of the optimization objective function is below the threshold.

[0015] Furthermore, based on the Bayesian inference method, uncertainty analysis is performed on the load results according to the optimized load distribution and the dynamic graph based on the time step t to obtain the load mean and variance of each node, specifically: Assume that the graph node The inversion load value Normal distribution ,in represents the mean value of the load, is the variance, indicating the uncertainty degree of the graph nodes, For graph nodes Inverted load values ​​in the optimized load distribution, where the posterior distribution is given by: ; in, is the joint probability of the load mean and variance given the optimized inversion load, is the likelihood function, reflecting the credibility of the inversion results; is a prior distribution, used to introduce background knowledge of mechanics; in, , represents the prior mean, represents the uncertainty of the mean; ,in The prior parameter representing the variance reflects the credibility of the mechanics; Define the likelihood function , combining the global balance error and the local gradient constraint: ; Among them, the first Describing Graph Nodes Inverse load values ​​in the optimized load distribution Deviation from the mean The second item Describing the gradients of adjacent graph nodes The rationality of Represents the dynamic weight and enhances the influence of gradient on the edges with high mechanical relevance.

[0016] Furthermore, by combining the prior distribution and the likelihood function, the posterior distribution is solved by the variational Bayesian inference method. The beneficial technical effects of the present invention are at least as follows: In view of the defect that traditional methods are difficult to adapt to dynamic environments, this paper introduces a dynamic weight mechanism based on physical drive. By combining real-time environmental loads (such as waves, tides, wind force) and sensor measurements, the key weight parameters in the load inversion model are dynamically adjusted to enable it to respond to environmental changes in real time. This mechanism effectively improves the adaptability of the inversion results to dynamic environments and solves the problem of low static modeling accuracy of traditional methods.

[0017] To solve the problem of sensor data being disturbed by noise in marine environments, this paper uses multimodal data fusion technology to jointly process multiple data sources such as stress, strain, acceleration, etc., and uses variational Bayesian inference to model and separate sensor noise. Through quantitative evaluation and optimization of data quality, the robustness of inversion results in complex environments is significantly improved.

[0018] In order to solve the problem of poor real-time performance of traditional methods, this paper adopts Graph Neural Network (GNN) to globally model the floating structure and designs a lightweight real-time computing framework. GNN uses the graph model of the floating structure to efficiently transmit and calculate information, thereby realizing real-time inversion of load distribution, effectively solving the problem of slow response of traditional centralized computing.

[0019] In order to solve the problem of lack of result uncertainty assessment in traditional methods, the present invention combines variational Bayesian inference technology to quantitatively evaluate the credible interval of the inversion results and provide uncertainty analysis of load distribution. The evaluation results can provide intuitive risk warnings and decision support for operators of floating installation. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0021] Figure 1 The present invention is a flow chart of a real-time inversion method of LMU load for floating installation. DETAILED DESCRIPTION

[0022] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0023] like Figure 1As shown, an embodiment of the present invention provides a real-time inversion method for LMU loads for floating installation, the method comprising the following steps S1-S5: S1. Construct an initial physical graph model of the floating installation structure. The initial physical graph model is constructed by modeling each sensor position in the floating structure as a graph node, modeling the physical mechanical relationship as a graph edge, and generating an initial physical graph based on node features and edge weights. At the same time, the initial physical graph is physically verified to generate a complete initial physical graph.

[0024] Specifically, this step aims to construct an initial physical graph model of the floating installation structure to provide accurate basic input for subsequent dynamic weight updates and load inversion. By modeling the sensor positions as graph nodes, the physical and mechanical relationships as graph edges, and generating a complete initial graph based on node features and edge weights, , achieving an accurate description of the complex mechanical relationships of the floating structure. This step specifically targets the multimodal data and physical models in the floating installation scenario, and designs a method for graph generation and verification to ensure that the model has engineering applicability and physical consistency.

[0025] Each sensor location in the floating structure is defined as a node , the initial eigenvector of the node Contains the following data: : Initial measurement values ​​of the sensor, including raw data of stress, strain and acceleration.

[0026] : The physical position coordinates of the node, describing the position relationship of the node in the three-dimensional floating structure.

[0027] : The theoretical load value calculated based on the initial mechanical model of the floatation structure, using the static formula: ; in, is the stiffness coefficient of the node, is the current displacement of the node, is the initial equilibrium position of the node.

[0028] side Connecting nodes that have a physical mechanical transmission relationship, initial edge weight Calculated according to the following formula: ; in, represents the three-dimensional Euclidean distance between nodes, It is the force value between nodes calculated according to the fluid mechanics formula.

[0029] Preferably, an initial physical graph is generated based on the definition of nodes and edges ,in: Node Set Includes all sensor locations, each node has characteristics ; Edge Set Represents a physical connection, each edge has a weight ; Adjacency Matrix represents the initial graph structure, where .

[0030] Preferably, to ensure that Mechanical rationality, calculation of overall balance error : ; in, Is a node The total connection force, is the theoretical load value.

[0031] if Exceeding the preset threshold , then optimize according to the following rules : Increase the connection weights of nodes associated with high force values.

[0032] Adjust the connectivity of isolated nodes so that they are connected to the nearest high-weight nodes.

[0033] Preferably, the initial physical graph model is output: Completely generated Contains the following structure information: The initial characteristics of each node .

[0034] The initial weight of each edge .

[0035] Adjacency Matrix and graph structure .

[0036] The initial physical map generated by this step It will serve as the basic input for dynamic weight update and real-time load inversion in subsequent steps.

[0037] S2. Collect environmental data in real time, fuse the environmental data with sensor data, dynamically update the complete initial physical map, and generate a dynamic map based on time step t; wherein the environmental data includes current velocity, wave pressure and wind speed.

[0038] Specifically, this step integrates environmental data and sensor data in real time to generate the initial physical map Perform dynamic updates to generate time steps Dynamic image , to reflect the mechanical characteristics of the floating structure in a dynamic environment. In order to adapt to the complex marine environment (such as wind speed, wave height, tidal current, etc.) in the patent scenario, this step innovatively designs a dynamic edge weight update method with physical regularization terms and a dynamic enhancement mechanism for specific node features. The final generated Provide accurate and high-quality input for real-time load inversion.

[0039] Furthermore, input data: Initial physical map , including node features and initial edge weights .

[0040] Real-time environmental data ,include: : ocean current speed, unit .

[0041] : Wave pressure, unit .

[0042] : Wind speed, unit .

[0043] Real-time sensor data , providing the node displacement and acceleration.

[0044] Preferably, the environmental force map: Using node location and the spatial distribution of environmental data, and calculate the local environmental force at each node : ; in, is a weighting factor determined by the floatation material properties and geometry, Representation Node The total environmental force.

[0045] Preferably, in order to adapt to the dynamic environment, the initial edge weight Based on the real-time interactive and innovative physical regularization terms , update the edge weights: ; in, is the generated initial edge weight. For Node and nodes Real-time interactive force, according to the environment force and node displacement difference calculate: ; in, It is the weight factor of the displacement difference, which indicates the influence of the mechanical action between nodes. is an innovative physical regularization term used to limit the deviation of edge weights from actual environmental forces: ; in, is the regularization strength coefficient, ensuring that the sum of the node connection forces is consistent with the local environmental forces.

[0046] During the update process, and is a dynamically adjusted weight balance coefficient, which is calibrated by fitting the noise level and historical data to meet .

[0047] Preferably, each node feature Updated under the influence of dynamic environment, it consists of two parts: basic feature enhancement and neighborhood aggregation: Basic feature enhancement: Modify the initial node features in combination with environmental forces: ; in, It is the sensitivity parameter of node characteristics to environmental forces, calibrated by empirical data.

[0048] Neighborhood aggregation update: Aggregate neighborhood node features, and weights are updated by Decide: ; in, Is a node The set of neighbors.

[0049] Preferably, output the dynamic graph model: Output time step Dynamic image : Node Features ,The dynamic update is completed, including the current environmental characteristics of the node and the neighborhood aggregation information.

[0050] Dynamic edge weights , reflecting the real-time changes of the mechanical relationship between nodes.

[0051] Adjacency Matrix , updated by edge weights.

[0052] Through the above steps, the dynamic graph generated It accurately depicts the real-time mechanical changes of the floating structure in a complex marine environment, and provides a dynamically adaptive graph input for the next step of load inversion.

[0053] S3. Design a load inference model based on graph neural network, perform load inversion calculation on the dynamic graph through the load inference model, obtain the inversion load of each node, and form a preliminary load distribution of the whole graph based on the inversion load of each node.

[0054] Specifically, this step uses the dynamic graph generated in the previous step , the real-time load inversion of floating structures is realized through Graph Neural Network (GNN). In view of the complex dynamic mechanical relationship and marine environmental impact in the floating installation scenario, this step combines neighborhood aggregation, residual correction and physical regularization constraints to design an innovative node information propagation and load inference mechanism to ensure that the inversion results are not only highly accurate but also meet the mechanical constraints.

[0055] The goal of this step is to use , calculate the load of each node through information propagation and inference and ensure that the inversion results satisfy the physical constraints.

[0056] Preferably, node feature propagation is achieved through dynamic neighborhood aggregation: node The feature update formula is: ; in Representation Node The neighbor set of is the number of neighbors, is the edge weight.

[0057] Preferably, considering the nonlinear effect of the dynamic environment, a residual correction term is designed during the node information propagation process. To correct the node feature update: ; in, is the total node connectivity predicted by edge weights. is the correction intensity parameter.

[0058] Preferably, the corrected node features are: ; Preferably, the load inference model is designed as follows: Load inference is achieved through a multi-layer perceptron (MLP) using the corrected node features Output node load values : ; in, It is a two-layer fully connected network whose parameters are trained by supervised learning methods.

[0059] Preferably, to ensure the physical consistency of the inversion results, the following regularization constraints are designed: Total force balance constraint: The total load on the node is forced to be consistent with the environmental force, and the formula is: ; Local gradient smoothing constraint: Limits drastic changes between adjacent node loads: ; Optimization objective function for load inference: ; in, represents the mean square error between the predicted value and the true value. and is the weight of the regularization term.

[0060] Understandably, output the load of each node , which constitutes the load distribution of the whole graph .

[0061] In this step, a load inversion method that adapts to complex dynamic environments is constructed through innovative node propagation mechanisms and physical regularization, providing scientific support for real-time monitoring of floating installations.

[0062] S4. Design physical constraints to modify and optimize the preliminary load distribution to obtain an optimized load distribution; wherein the physical constraints include global force balance, local smoothness and time dynamic consistency.

[0063] Specifically, the input data includes the preliminary load distribution output from step 3 ,in Representation Node The inverse load value of . Dynamic graph ,in Representation Node and Dynamic correlation weights between Provides nodes External force information, load distribution in the previous time step Reference for time dynamic constraints. The goal is to Perform verification and optimization to ensure that: Global Force Balance: Nodal Loads Forces from the external environment Be consistent.

[0064] Local smoothness: The load gradient between adjacent nodes remains within a reasonable range.

[0065] Time dynamic consistency: The load of the node changes continuously over time without drastic fluctuations.

[0066] Preferably, each node Inversion load Should interact with its environment Definition of balance error .when Exceeds the set tolerance threshold When , it is corrected by the following formula: ; in, The step size is adjusted dynamically to avoid excessive or insufficient correction. After verification, the balance error of each node will be controlled within a reasonable range to ensure global physical consistency.

[0067] Preferably, adjacent nodes and The load gradient satisfies ,in is the maximum allowed gradient. When the gradient condition is not met, the neighborhood weighted average method is used for correction: ; in For Node The neighbor set of For Node and This optimization step ensures local smoothness and avoids sudden changes in loading values.

[0068] Preferably, a time dynamic consistency regularization term is designed to limit the load change of the same node in consecutive time steps to be too large: ; This regularization term provides constraints on time continuity in the optimization objective to avoid unreasonable fluctuations in a dynamic environment.

[0069] Preferably, the above checks and constraints are combined to define the optimization objective function: ; The first term ensures global force balance, the second term achieves smoothness through local gradient constraints, and the third term constrains the temporal variation of the load through a temporal dynamic consistency regularization term. and is the weight that controls the priority of the regular term.

[0070] Update using gradient descent : ; in is the learning rate. Iterate and optimize until the following convergence conditions are met: Balance error of all nodes satisfy .

[0071] Optimizing the objective function The rate of change is below the threshold .

[0072] Understandably, the output is optimized for load distribution , the results have the following characteristics: Satisfy the global force balance constraints to ensure that the node loads are consistent with the external forces.

[0073] Achieve gradient smoothing between neighborhoods and avoid local load mutations.

[0074] It complies with the requirements of time dynamic consistency and ensures the time continuity of load distribution.

[0075] This step combines physical verification, neighborhood optimization, and smoothing regularization to improve the rationality of the load inversion results for complex floating installation scenarios, providing high-quality input for subsequent analysis.

[0076] S5. Based on the Bayesian inference method, uncertainty analysis is performed on the load results according to the optimized load distribution and the dynamic diagram based on time step t, and the load mean and variance of each node are obtained to describe the credibility of the load inversion results. At the same time, the uncertainty distribution map of the whole map is generated in combination with the node position, and the high-risk areas are marked to provide engineering optimization and decision-making for floating installation.

[0077] Specifically, the input data is the load distribution after optimization in step 4 , each node Indicates the load value after inversion optimization. In addition, the dynamic graph Provides node adjacency and weight , and the local gradients between nodes The goal of this step is to use the Bayesian inference method to quantify the uncertainty of the load results and output the mean load of each node. and variance , helping to assess the credibility and risk of load inversion results.

[0078] Preferably, the Bayesian model is constructed: Assume Node Load value Normal distribution ,in represents the mean value of the load, is the variance, indicating the degree of node uncertainty. The posterior distribution is given by the following formula: ; in is the likelihood function, reflecting the credibility of the inversion results; is a prior distribution, which is used to introduce background knowledge of mechanics.

[0079] Prior distribution setting: , represents the prior mean, represents the uncertainty in the mean.

[0080] ,in The prior parameter representing the variance reflects the credibility of the mechanics.

[0081] Preferably, define the likelihood function , combining the global balance error and the local gradient constraint: ; The first item describes the node Inversion load Deviation from the mean degree.

[0082] The second term describes the gradient of the adjacent nodes The rationality of Represents the dynamic weight and enhances the influence of gradient on the edges with high mechanical relevance.

[0083] To enhance physical consistency, a gradient regularization term is innovatively introduced , controlling the anomaly of the gradient distribution.

[0084] Preferably, the prior distribution and the likelihood function are combined, and the posterior distribution is solved by the variational Bayesian inference method to optimize the following objective function: ; in is the logarithmic posterior loss function.

[0085] Preferably, update the rules: initialization .

[0086] Iterative update using gradient descent: ; in is the learning rate.

[0087] Iterate until convergence, and get and final estimate.

[0088] It is understandable that the final output is the mean load of each node and variance , which is used to describe the credibility of the load inversion results. At the same time, the uncertainty distribution map of the whole map is generated in combination with the node position, and the high-risk areas are marked to support the engineering optimization and decision-making of the floating installation.

[0089] This step quantifies the uncertainty of load distribution by capturing the global balance and local gradient changes through the Bayesian inference method, providing a reliable risk analysis tool.

[0090] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0091] In addition, for technical details that are not described in detail in this embodiment, reference can be made to the parameter operation method provided in any embodiment of the present invention, and will not be repeated here.

[0092] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0093] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory / random access memory, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0094] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A real-time inversion method for LMU loads for floating installation, characterized in that: The method comprises: Constructing an initial physical graph model of the floating installation structure, wherein the initial physical graph model is formed by modeling each sensor position in the floating structure as a graph node, modeling the physical mechanical relationship as a graph edge, and generating an initial physical graph based on graph node features and edge weights, and performing physical verification on the initial physical graph to generate a complete initial physical graph; Collect environmental data in real time, fuse the environmental data with sensor data, dynamically update the complete initial physical map, and generate a dynamic map based on time step t; wherein the environmental data includes current velocity, wave pressure and wind speed; Design a load inference model based on graph neural network, use the load inference model to perform load inversion calculation on the dynamic graph, obtain the inversion load of each graph node, and form a preliminary load distribution of the whole graph based on the inversion load of each graph node; Designing physical constraints to modify and optimize the preliminary load distribution to obtain an optimized load distribution; wherein the physical constraints include global force balance, local smoothness, and time dynamic consistency; Based on the Bayesian inference method, uncertainty analysis of the load results is performed according to the optimized load distribution and the dynamic graph based on time step t, and the load mean and variance of each graph node are obtained to describe the credibility of the load inversion results. At the same time, the uncertainty distribution map of the entire graph is generated in combination with the graph node position, and high-risk areas are marked to provide engineering optimization and decision-making for floating installation.

2. A real-time inversion method for LMU loads for floating installation according to claim 1, characterized in that: The initial physical graph model is constructed as follows: Define each sensor location in the floating structure as a graph node of the physical graph model , graph node The initial eigenvector of Contains the following data: initial measurement value of sensor i, physical position coordinates of sensor i, theoretical load value calculated according to the initial mechanical model of the floating structure; The theoretical load value calculated based on the initial mechanical model of the floating structure It is obtained by calculating the theoretical load value applied to the nodes of the graph; side Obtained through modeling of physical and mechanical relationships; According to the graph node and edge Definition of , generating the initial physical diagram , where the node set Including all sensor locations, each graph node has the feature ; Edge Set Represents a physical connection, each edge has an initial weight ; represents the initial edge weight set; Adjacency Matrix Represents the initial graph structure, where the edge weights of the adjacency matrix .

3. A real-time inversion method for LMU loads for floating installation according to claim 2, characterized in that: Calculate the overall balance error through error analysis based on the total connection force and theoretical load value of the graph nodes; If the overall balance error exceeds the preset threshold, the initial weights are optimized according to the following rules: : Increase the connection weights between graph nodes associated with high force values; Adjust the connectivity of isolated graph nodes so that they are connected to the nearest high-weight graph nodes.

4. The real-time inversion method for LMU loads for floating installation according to claim 1 is characterized in that: After collecting environmental data, use the graph node Location And the spatial distribution of environmental data, environmental force mapping is performed to determine the local environmental force of each graph node .

5. A real-time inversion method for LMU loads for floating installation according to any one of claims 3 or 4, characterized in that: The environmental data is integrated with the sensor data, and the complete initial physical graph is dynamically updated to generate a dynamic graph based on the time step t, specifically including: The real-time interaction force and the physical regularization term are integrated into the edge weight to update the edge weight; wherein the physical regularization term is used to limit the deviation between the edge weight and the actual environment force; Among them, during the edge weight update process, the dynamically adjusted weight balance coefficient is calibrated by fitting the noise level and historical data; in, Each graph node feature is updated under the influence of the dynamic environment, which consists of two parts: basic feature enhancement and neighborhood aggregation: Basic feature enhancement: Modify the graph node features of the initial graph nodes in combination with environmental forces; Neighborhood aggregation update: aggregate neighborhood graph node features, with weights determined by updated edge weights; The final output time step Dynamic graph ,in, is a node set, For edge sets, is the edge weight set.

6. A real-time inversion method for LMU loads for floating installation according to claim 5, characterized in that: Before designing a load inference model based on graph neural network, a node information propagation mechanism is designed to correct and propagate the update of graph node information, including: Graph node feature propagation is achieved through dynamic neighborhood aggregation; At the same time, during the process of graph node information propagation, a residual correction term is designed based on the error between the total environmental force and the total connection force of the real-time graph nodes to correct the graph node feature update.

7. A real-time inversion method for LMU loads for floating installation according to claim 6, characterized in that: The load inference model based on graph neural network is constructed as follows: Load inference is achieved through a multi-layer perceptron, which uses the corrected graph node features to output the load value of the current graph node; wherein the multi-layer perceptron includes a two-layer fully connected network, and its parameters are trained through a supervised learning method; At the same time, a total force balance constraint and a local gradient smoothing constraint are designed to ensure the physical consistency of the inversion results. The total force balance constraint forces the total load of the graph node to be consistent with the environmental force, and the local gradient smoothing constraint limits the drastic changes between the loads of adjacent graph nodes.

8. A real-time inversion method for LMU loads for floating installation according to claim 7, characterized in that: The design physical constraints modify and optimize the preliminary load distribution to obtain an optimized load distribution, specifically including: The global force balance check correction: Inverse load value for each graph node Should be equal to the total environmental force on the current graph node If they are consistent, then the balance error is defined ,when Exceeds the set tolerance threshold When the inverse load value is Error correction is performed by combining dynamically adjusted step-size weighted balancing errors; The local gradient smoothing check correction: Adjacent graph nodes and The load gradient satisfies ,in is the maximum allowed gradient, Adjacent graph nodes and The dynamic correlation weight of When , the neighborhood weighted average method is used for correction; in, The inverted load value of each graph node is updated using the gradient descent method; Iterate the optimization until the following convergence conditions are met: Balanced error of all graph nodes satisfy ; The rate of change of the optimization objective function is below the threshold.

9. A real-time inversion method for LMU loads for floating installation according to claim 8, characterized in that: Based on the Bayesian inference method, uncertainty analysis is performed on the load results according to the optimized load distribution and the dynamic graph based on the time step t to obtain the load mean and variance of each node, specifically: Assume that the graph node The inversion load value Normal distribution ,in represents the mean value of the load, is the variance, indicating the uncertainty degree of the graph nodes, For graph nodes Inverted load values ​​in the optimized load distribution, where the posterior distribution is given by: ； in, is the joint probability of the load mean and variance given the optimized inversion load, is the likelihood function, reflecting the credibility of the inversion results; is a prior distribution, used to introduce background knowledge of mechanics; in, , represents the prior mean, represents the uncertainty of the mean; ,in The prior parameter representing the variance reflects the credibility of the mechanics; Define the likelihood function , combining the global balance error and the local gradient constraint: ； Among them, the first Describing Graph Nodes Inverse load values ​​in the optimized load distribution Deviation from the mean The second item Describing the gradients of adjacent graph nodes The rationality of Represents the dynamic weight and enhances the influence of gradient on the edges with high mechanical relevance.

10. A real-time inversion method for LMU loads for floating installation according to claim 9, characterized in that: Combining the prior distribution with the likelihood function, the posterior distribution is solved by the variational Bayesian inference method.

Images