Island power transmission submarine cable risk dynamic management and control method based on space-time three-dimensional association

Abstract

The invention discloses an island power transmission submarine cable risk dynamic management and control method based on space-time three-dimensional association, and the method comprises the steps: collecting the multi-source heterogeneous data of an island power transmission submarine cable, carrying out the data pre-cleaning and space-time dimension alignment processing, and obtaining the pre-processed multi-source heterogeneous data, constructing a data cube structure of the multi-source heterogeneous data to carry out space-time cross correlation analysis to obtain a space-time correlation graph of the multi-source heterogeneous data, inputting the space-time correlation graph into a pre-constructed dynamic risk assessment model to carry out graph structure definition, and carrying out space-time convolution and feature fusion processing to output a risk probability, and carrying out risk grade division on the whole island power transmission submarine cable according to the risk probability, carrying out grading risk management and control strategy matching on the island power transmission submarine cable, and executing response processing. The method has the advantages that the active sensing capability, the precise early warning level and the intelligent operation and maintenance decision-making efficiency of the submarine cable system can be improved, and the operation toughness and the power supply reliability of an island power grid are fundamentally improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of submarine cable risk management, in particular to a dynamic management method for island power transmission submarine cable risk based on space-time stereoscopic correlation. BACKGROUND

[0002] At present, as the core energy artery connecting offshore islands and mainland power grids, the reliability of island power transmission submarine cables is directly related to the safety of marine strategy and the livelihood protection of border areas. In recent years, the rapid development of new marine economic forms such as deep-sea wind power grid connection and offshore resource development has made submarine cable systems face multi-dimensional and multi-scale compound risk challenges.

[0003] The existing submarine cable risk early warning method is usually based on static parameters to judge the submarine cable risk, and its system architecture and early warning strategy are relatively simple, but it cannot adapt to the changes in the complexity of the environment of the power transmission submarine cable, and has limitations in data linkage, real-time response and dynamic trend prediction, and it is difficult to meet the perception and prevention and control needs of the changing external risks of island power transmission submarine cables. The submarine cable risk assessment method based on machine learning is mostly used for binary classification problems under static data, and has not formed comprehensive judgment ability on multiple data dimensions of submarine cable risk related data, and still has problems such as poor real-time performance and limited generalization ability. SUMMARY

[0004] In view of the problems of multi-source data island, static risk assessment and lagging management and control in the existing submarine cable risk early warning method, the present application provides a dynamic management method for island power transmission submarine cable risk based on space-time stereoscopic correlation, which can fully integrate space-time multi-dimensional data features, realize dynamic modeling and prediction of risk state, and has intelligent response capability, which can improve the active perception ability, accurate early warning level and intelligent operation and maintenance decision efficiency of the submarine cable system, and fundamentally improve the operation resilience and power supply reliability of the island power grid.

[0005] In the first aspect, the above application aims to achieve the following technical solutions: A dynamic management method for island power transmission submarine cable risk based on space-time stereoscopic correlation, the method comprising: Collecting multi-source heterogeneous data of island power transmission submarine cables, performing data pre-cleaning and space-time dimension alignment processing on the multi-source heterogeneous data to obtain pre-processed multi-source heterogeneous data; Building a data cube structure of the multi-source heterogeneous data, and performing space-time cross-correlation analysis on the multi-source heterogeneous data to obtain a space-time correlation graph of the multi-source heterogeneous data; Inputting the space-time correlation graph into a pre-constructed dynamic risk assessment model for graph structure definition, and performing space-time convolution and feature fusion processing, and outputting the risk probability of the island power transmission submarine cable; Based on the aforementioned risk probabilities, the entire isolated island power transmission submarine cable is classified into risk levels, and risk management strategies are matched and response processes are executed according to the classified risk levels.

[0006] In a preferred embodiment, this application can be further configured as follows: after classifying the entire isolated island submarine power transmission cable into risk levels based on the risk probability, and matching and executing response processing of the graded risk management strategy for the isolated island submarine power transmission cable according to the classified risk levels, it further includes: The execution response effect of the hierarchical risk management strategy is obtained, and the corresponding risk management strategy is rewarded based on the execution response effect and a preset reward function, and the strategy priority of the risk management strategy is dynamically adjusted. The calculation expression for the reward function is as follows: (1) in, This represents the reward coefficient, used to adjust strategy priority. Indicates the amount of risk reduction. Indicates control costs, , This represents the weighting coefficient.

[0007] In a preferred embodiment, this application can be further configured as follows: the step of classifying the risk level of the entire isolated island power transmission submarine cable according to the risk probability, and matching and executing a graded risk management strategy for the isolated island power transmission submarine cable according to the classified risk level, specifically includes: The entire isolated island power transmission submarine cable is divided into grids using a pre-constructed data cube structure, and each grid is aligned in both spatiotemporal dimensions to obtain a spatiotemporally aligned submarine cable grid. Submarine cable risk heatmaps are aggregated according to the risk probability of each submarine cable grid. Based on the aggregation results, continuous submarine cable risk heatmaps are overlaid and predicted. Based on the prediction results, risk levels are classified. Based on the risk level classification results, a graded risk management strategy matching process is performed for each submarine cable grid, and the matched risk management strategy is executed and responded to in each submarine cable grid according to the matching results.

[0008] In a preferred embodiment, this application can be further configured as follows: the process of aggregating submarine cable risk heatmaps according to the risk probability of each submarine cable grid, performing overlay prediction processing on continuous submarine cable risk heatmaps based on the aggregation results, and classifying risk levels based on the prediction results, further includes: The overlay prediction of the submarine cable risk heat map is used for submarine cable risk prediction during extreme weather events.

[0009] In a preferred embodiment, this application can be further configured as follows: the step of inputting the spatiotemporal correlation graph into a pre-constructed dynamic risk assessment model for graph structure definition, and performing spatiotemporal convolution and feature fusion processing to output the risk probability of isolated submarine power transmission cables specifically includes: The grid attribute vectors in the spatiotemporal correlation graph are defined as graph nodes, and the reciprocal of the grid spatial distance in the spatiotemporal correlation graph is defined as the edge weight, thus obtaining the graph structure definition parameters of the spatiotemporal correlation graph. The spatial features and temporal dependency features of the spatiotemporal correlation map are extracted through the dynamic risk assessment model, and spatiotemporal convolution is performed according to the spatial features and the temporal dependency features to obtain the spatiotemporal features. The spatiotemporal features are sequentially concatenated according to the time dimension and input into the fully connected network layer of the dynamic risk assessment model to output the risk probability of the isolated submarine power transmission cable.

[0010] In a preferred example, this application can be further configured as follows: the training process of inputting the spatiotemporal correlation graph into a pre-constructed dynamic risk assessment model for graph structure definition, and performing spatiotemporal convolution and feature fusion processing to output the risk probability of isolated submarine power transmission cables, specifically includes: Historical monitoring data containing real accidents involving isolated submarine power transmission cables were acquired. A dynamic risk assessment model was obtained by constructing a neural convolutional network using cross-entropy and spatiotemporal smoothing terms as loss functions and the AdamW optimization algorithm combined with a custom learning rate.

[0011] In a preferred embodiment, this application can be further configured as follows: the construction of the data cube structure of the multi-source heterogeneous data, and the performance of spatiotemporal cross-correlation analysis on the multi-source heterogeneous data to obtain a spatiotemporal correlation map of the multi-source heterogeneous data, specifically includes: Three-dimensional modeling is performed using time, space, and attribute dimensions to construct a data cube structure for the multi-source heterogeneous data. The spatiotemporal lag conditions of the multi-source heterogeneous data are then analyzed based on the data cube structure. Calculate the correlation coefficient matrix of the multi-source heterogeneous data under spatiotemporal lag conditions, and identify key lag times; The Granger causality test is used to determine the variable causal relationship between the key lag time and the corresponding multi-source heterogeneous data, and the spatiotemporal cross-correlation is judged based on the variable causal relationship. Spatiotemporal correlation maps are constructed for multi-source heterogeneous data with spatiotemporal cross-correlation, resulting in spatiotemporal correlation maps of multi-source heterogeneous data.

[0012] In a preferred embodiment, this application can be further configured as follows: the process of collecting multi-source heterogeneous data from isolated submarine power transmission cables, pre-cleaning the multi-source heterogeneous data, and performing spatiotemporal alignment processing to obtain pre-processed multi-source heterogeneous data specifically includes: Calculate the average value of the historical data sliding window of the isolated submarine power transmission cable, and set a dynamic threshold range according to the average value of the historical data sliding window to remove multi-source heterogeneous data outside the dynamic threshold range; The multi-source heterogeneous data was processed by cubic spline interpolation to repair missing values, resulting in pre-cleaned multi-source heterogeneous data. The multi-source heterogeneous data is aligned in time dimension according to a preset unified time format, and then aligned in spatial dimension using inverse distance weighted interpolation to obtain preprocessed multi-source heterogeneous data.

[0013] In a preferred embodiment, this application can be further configured as follows: the step of collecting multi-source heterogeneous data from isolated submarine power transmission cables, performing data pre-cleaning and spatiotemporal dimension alignment on the multi-source heterogeneous data to obtain pre-processed multi-source heterogeneous data, further includes: The multi-source heterogeneous data includes submarine cable temperature data, submarine cable electrical data, marine environmental data, seabed geological activity data, and ship activity data.

[0014] Secondly, the above-mentioned inventive objective of this application is achieved through the following technical solutions: A dynamic risk management system for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation, wherein the system is applied to the aforementioned dynamic risk management method for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation, and the system includes: The front-end sensing layer is equipped with industrial-grade Hall effect voltage sensors, Rogowski coil current sensors, dedicated temperature sensors, and capacitive humidity sensors to collect multi-source heterogeneous data from isolated submarine power transmission cables. The edge computing layer is equipped with high-performance FPGA chips, high-performance DSP modules, PCIe and Ethernet interfaces, and is used for data pre-cleaning and spatiotemporal dimension alignment of multi-source heterogeneous data. The cloud-based analytics layer is equipped with high-performance servers for modeling and risk assessment of dynamic risk assessment models for isolated submarine power transmission cables. The control and execution layer is used to match the corresponding hierarchical risk control strategies based on the risk assessment results and to deploy the corresponding control equipment in accordance with the risk control strategies.

[0015] In summary, this application includes at least one of the following beneficial technical effects: 1. Achieve efficient fusion of multi-source heterogeneous data: By integrating multi-source heterogeneous data through a spatiotemporal data cube, the data of submarine earthquakes, temperature, strain, electrical parameters and other multi-source data are uniformly spatiotemporally aligned and fused, solving the data silo problem in traditional technologies, significantly improving the risk identification coverage, and enhancing the ability to fully perceive risk signs. 2. Improve the accuracy of dynamic risk assessment: By constructing a spatiotemporal data cube and introducing algorithms such as STCCA (Spatiotemporal Cross-Correlation Analysis) and spatiotemporal graph convolutional networks, the correlation lag effects of different risk factors in time and space can be accurately identified; by constructing a graph structure reflecting the spatial location and time series along the submarine cable through a spatiotemporal graph neural network, the change process of the submarine cable's operating status with time and geographical location can be dynamically modeled to achieve accurate prediction. 3. Excellent adaptability to remote scenarios and feasibility of engineering implementation: By combining 5G private network, Beidou short message and maritime satellite communication technologies, stable data transmission in near and far sea areas is achieved, ensuring that the system can continue to operate under special conditions such as extreme weather and far from land, providing a reliable guarantee for the long-term stable operation of the submarine power transmission cable in isolated islands; 4. Optimize the dynamic management and control system with a graded response mechanism: Set risk levels based on model prediction results and trigger different response strategies (such as scheduling, load limiting, and alarms) accordingly to ensure operational safety and response efficiency; dynamically match multi-level response strategies according to risk levels, from daily report generation and drone inspection to emergency power outage and ROV intervention, to achieve a closed-loop mechanism from "risk discovery" to "proactive intervention". Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0017] Figure 1 This is a flowchart illustrating the implementation of the dynamic risk management method for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation in this embodiment.

[0018] Figure 2 This is a flowchart of the dynamic threshold setting process for the risk dynamic management method of isolated submarine power transmission cables in this embodiment.

[0019] Figure 3 This is a schematic diagram of the data processing of the dynamic risk assessment model in this embodiment.

[0020] Figure 4 This is a risk assessment diagram of the dynamic risk management method for isolated submarine power transmission cables in this embodiment.

[0021] Figure 5This is an overall data flow diagram of the dynamic risk management method for isolated submarine power transmission cables in this embodiment.

[0022] Figure 6 This is a structural block diagram of the island power transmission submarine cable risk dynamic management and control system based on spatiotemporal three-dimensional correlation in this embodiment. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0025] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0027] In one embodiment, this application discloses a method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation, such as... Figure 1 As shown, the specific steps include the following: S10: Collect multi-source heterogeneous data of isolated island power transmission submarine cables, perform data pre-cleaning and spatiotemporal dimension alignment processing on the multi-source heterogeneous data to obtain pre-processed multi-source heterogeneous data.

[0028] Specifically, multi-source heterogeneous data of the isolated island power transmission submarine cable is collected through industrial-grade Hall effect voltage sensors, Rogowski coil current sensors, dedicated temperature sensors, and capacitive humidity sensors deployed at the front end. The multi-source heterogeneous data includes submarine cable temperature data, submarine cable electrical data, marine environmental data, seabed geological activity data, and ship activity data.

[0029] Specifically, submarine cable temperature data is collected by a distributed optical fiber sensing system (DTS), with parameters including temperature and strain values; submarine cable electrical parameters are obtained through current transformers (CT) and voltage transformers (PT), with parameters including current, voltage, and resistance; marine environmental data is collected using an acoustic Doppler current profiler (ADCP) and a pressure wave height meter, with parameters including ocean current velocity and wave height; seabed geological activity data is collected by seabed seismometers (OBS) and satellite InSAR, with parameters including seismic intensity and deformation; and ship activity data is collected using an automatic identification system (AIS) and radar, with parameters including ship speed and distance from the submarine cable, and the collection frequency is real-time tracking.

[0030] The data pre-cleaning and spatiotemporal dimension alignment processing in step S10 includes: S101: Calculate the historical sliding window mean of the isolated submarine power transmission cable, and set a dynamic threshold range according to the historical sliding window mean to remove multi-source heterogeneous data outside the dynamic threshold range.

[0031] Specifically, using historical data from isolated submarine power transmission cables, the average of the historical data sliding window is calculated. In this embodiment, the sliding window is set to 7 days, meaning the average of historical data over seven days is calculated. This sets the dynamic threshold range for data sampling. In this embodiment, the threshold range is set to the mean ± 3 times the standard deviation (3σ), automatically eliminating heterogeneous data points from multiple sources outside the dynamic threshold range. The flowchart for setting the dynamic threshold in this embodiment is as follows: Figure 2 As shown.

[0032] S102: The missing value repair process of multi-source heterogeneous data is performed by cubic spline interpolation to obtain pre-cleaned multi-source heterogeneous data.

[0033] Specifically, the missing value repair process of multi-source heterogeneous data is carried out by cubic spline interpolation to complete the data pre-cleaning and obtain the pre-cleaned multi-source heterogeneous data.

[0034] S103: Perform time dimension alignment processing on the multi-source heterogeneous data according to the preset unified time format, and perform spatial dimension alignment processing on the multi-source heterogeneous data through the inverse distance weight interpolation method to obtain the preprocessed multi-source heterogeneous data.

[0035] Specifically, multi-source heterogeneous data with different sampling frequencies are unified into a 10-minute time slice according to a preset unified time format. For example, electrical parameters sampled 10 times per second are downsampled to a 10-minute average. The submarine cable path is divided into a 100m×100m grid. Inverse distance weighted interpolation (IDW) is used to map buoy and satellite data to the grid center point, and the spatial dimensions of the multi-source heterogeneous data are aligned to obtain preprocessed multi-source heterogeneous data.

[0036] S20: Construct a data cube structure for multi-source heterogeneous data, and perform spatiotemporal cross-correlation analysis on the multi-source heterogeneous data to obtain a spatiotemporal correlation map of the multi-source heterogeneous data.

[0037] Specifically, step S20 includes: S201: Perform three-dimensional modeling using time, space, and attribute dimensions to construct a data cube structure for multi-source heterogeneous data, and analyze the spatiotemporal lag conditions of multi-source heterogeneous data based on the data cube structure.

[0038] Specifically, a three-dimensional model is constructed using time (T-axis), space (S-axis), and attributes (A-axis) to build a data cube. The T-axis is divided into 10-minute, 1-hour, and 1-day segments, supporting sliding time window analysis. Each grid on the S-axis contains attributes such as latitude and longitude coordinates, water depth, and distance to the submarine cable. The A-axis contains over 30 parameters across four categories, including submarine cable temperature gradient (ΔT / 100m), vessel density (vessels / km²), and cumulative microseismic energy (J). The spatiotemporal lag conditions of multi-source heterogeneous data are analyzed based on the data cube.

[0039] S202: Calculate the correlation coefficient matrix of multi-source heterogeneous data under spatiotemporal lag conditions and identify key lag times.

[0040] Specifically, by using STCCA (Spatiotemporal Cross-Correlation Analysis) and spatiotemporal graph convolutional networks, the correlation lag effects of different risk factors in time and space are identified. Then, the correlation coefficient matrix of multi-source heterogeneous data under spatiotemporal lag conditions is calculated, and key lag times are identified, such as the lag time of the impact of ocean current velocity changes on submarine cable temperature being 3 hours.

[0041] The correlation coefficient matrix expression in this embodiment is as follows: (1) in, Represents the correlation coefficient matrix. Indicates the weighting coefficient. , These represent the temporal correlation matrix and spatial correlation matrix of multi-source heterogeneous data, respectively, which are obtained through the data coupling relationship of multi-source heterogeneous data.

[0042] S203: Determine the causal relationship between key lag times and corresponding multi-source heterogeneous data through Granger causality tests, and make spatiotemporal cross-correlation judgments based on the causal relationship between variables.

[0043] Specifically, based on Granger causality tests to determine the causal relationship between key lag times and corresponding multi-source heterogeneous data, a vector autoregression (VAR) model is constructed. The VAR model is used to test the causal relationship between variables. If a causal relationship exists, it is determined that there is spatiotemporal cross-correlation; if no causal relationship exists, it is determined that there is no spatiotemporal cross-correlation.

[0044] S204: Construct a spatiotemporal correlation map for multi-source heterogeneous data with spatiotemporal cross-correlation to obtain a spatiotemporal correlation map of multi-source heterogeneous data.

[0045] Specifically, a spatiotemporal correlation map is constructed for multi-source heterogeneous data that exhibits spatiotemporal cross-correlation, resulting in a spatiotemporal correlation map of the multi-source heterogeneous data.

[0046] S30: Input the spatiotemporal correlation graph into the pre-built dynamic risk assessment model to define the graph structure, and perform spatiotemporal convolution and feature fusion processing to output the risk probability of isolated submarine power transmission cables.

[0047] Specifically, step S30 includes: S301: Define the grid attribute vectors in the spatiotemporal correlation graph as graph nodes, and define the reciprocal of the grid spatial distance in the spatiotemporal correlation graph as the edge weight, to obtain the graph structure definition parameters of the spatiotemporal correlation graph.

[0048] Specifically, the grid attribute vectors in the spatiotemporal correlation graph are defined as graph nodes, and the reciprocal of the grid spatial distance in the spatiotemporal correlation graph is defined as the edge weight, thus obtaining the graph structure definition parameters of the spatiotemporal correlation graph. The grid attribute vectors include ocean current velocity, temperature, ship density, etc. The closer the grid spatial distance, the higher the edge weight.

[0049] S302: Extract spatial and temporal dependency features of the spatiotemporal correlation map through a dynamic risk assessment model, and perform spatiotemporal convolution processing according to the spatial and temporal dependency features to obtain spatiotemporal features.

[0050] Specifically, spatial features of the spatiotemporal correlation graph are extracted through the spatial convolutional layer in the dynamic risk assessment model, and temporal dependency features are extracted through the temporal convolutional layer. Spatiotemporal convolution processing is performed according to the spatial features and temporal dependency features to obtain the spatiotemporal features of the spatiotemporal correlation graph. Among them, the spatial convolutional layer uses Chebyshev polynomial approximation graph convolution to extract spatial features, and the temporal convolutional layer uses one-dimensional dilated CNN to capture long temporal dependencies.

[0051] S303: The spatiotemporal features are sequentially spliced ​​according to the time dimension and input into the fully connected network layer of the dynamic risk assessment model to output the risk probability of isolated submarine power transmission cables.

[0052] Specifically, the spatiotemporal features are sequentially spliced ​​according to the time dimension and input into the fully connected network layer of the dynamic risk assessment model for data fusion, outputting the risk probability of isolated submarine power transmission cables.

[0053] The training process of the dynamic risk assessment model in this embodiment specifically includes: Historical monitoring data containing real accidents involving isolated submarine power transmission cables were acquired. A dynamic risk assessment model was obtained by constructing a neural convolutional network using cross-entropy and spatiotemporal smoothing terms as loss functions and the AdamW optimization algorithm combined with a custom learning rate.

[0054] Specifically, in this embodiment, the model training samples use historical monitoring data from the past three years containing real cable accident events. The optimization algorithm is AdamW, the loss function is cross-entropy plus a spatiotemporal smoothing term, and the learning rate parameter is customized through grid search. A neural convolutional network is constructed using these settings to obtain the dynamic risk assessment model. The dynamic risk assessment model in this embodiment is a spatiotemporal graph convolutional network (ST-GCN) model. A schematic diagram of the data processing for the dynamic risk assessment model is shown below. Figure 3 As shown.

[0055] S40: Classify the risk level of the entire isolated island power transmission submarine cable according to the risk probability, and match the graded risk management strategy for the isolated island power transmission submarine cable according to the classified risk level and execute the response processing.

[0056] Specifically, step S40 includes: S401: The entire isolated submarine power transmission cable is divided into grids using a pre-built data cube structure, and each grid is aligned in both spatiotemporal dimensions to obtain a spatiotemporally aligned submarine cable grid.

[0057] Specifically, the entire isolated island power transmission submarine cable is divided into grids using a pre-constructed data cube structure, and each grid is aligned in both time and space dimensions to obtain a spatiotemporally aligned submarine cable grid.

[0058] S402: Aggregate the submarine cable risk heatmap according to the risk probability of each submarine cable grid, overlay and predict the continuous submarine cable risk heatmaps based on the aggregation results, and classify the risk level based on the prediction results.

[0059] Specifically, the risk probability of each submarine cable grid is calculated and aggregated to generate a risk heat map of the entire submarine cable. The risk heat map is used to determine whether the risk probabilities of three consecutive grids fall within the preset risk value of the corresponding risk level, and then the corresponding risk level is predicted.

[0060] In this embodiment, the risk levels are categorized as low risk, medium risk, and high risk. Specifically, when the risk probability of three consecutive grids is less than 0.3, it is classified as low risk; between 0.3 and 0.7, it is classified as medium risk; and greater than or equal to 0.7, it is classified as high risk.

[0061] S403: Based on the risk level classification results, perform hierarchical risk management strategy matching for each submarine cable grid, and execute and respond to the matched risk management strategy in each submarine cable grid according to the matching results.

[0062] Specifically, based on the risk level classification, a tiered risk management strategy is matched for each submarine cable grid. The matched risk management strategy is then executed and responded to in each grid. Low-risk strategies include generating daily reports and conducting inspections every 6 hours, with the local monitoring terminal as the linkage system. Medium-risk strategies include drone patrols, anti-corrosion current injection, and AIS broadcast alerts, with the drone control platform and AIS system as the linkage system. High-risk strategies include power cut-off, dispatching ROVs for maintenance, and maritime control. The system responds to the corresponding risk management strategy and executes the appropriate control actions based on the risk management strategy matching results.

[0063] The submarine cable risk heatmap overlay prediction in this embodiment is also used for submarine cable risk prediction during extreme weather events. A risk assessment diagram of this embodiment is shown below. Figure 4 As shown.

[0064] This embodiment further includes the following after step S40: Obtain the execution response effect of the tiered risk management strategy, reward the corresponding risk management strategy based on the execution response effect and a preset reward function, and dynamically adjust the strategy priority of the risk management strategy.

[0065] The calculation expression for the reward function is as follows: (1) in, This represents the reward coefficient, used to adjust strategy priority. Indicates the amount of risk reduction. Indicates control costs, , This represents the weighting coefficient.

[0066] Specifically, by recording the execution effects of each control strategy, such as risk reduction rate and fault avoidance rate, a reward coefficient is calculated for the corresponding risk control strategy using a preset reward function. The strategy priority is then dynamically adjusted based on the reward coefficient and a deep reinforcement learning algorithm. The overall data flow diagram of the dynamic risk control method for isolated submarine power transmission cables in this embodiment is shown below.Figure 5 As shown.

[0067] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0068] In one embodiment, a dynamic risk management system for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation is provided. This system corresponds one-to-one with the dynamic risk management method for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation described in the above embodiments. Figure 6 As shown, this dynamic risk management system for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation includes a front-end perception layer, an edge computing layer, a cloud analysis layer, and a management execution layer. Detailed descriptions of each functional module are as follows: The front-end sensing layer is equipped with industrial-grade Hall effect voltage sensors, Rogowski coil current sensors, dedicated temperature sensors, and capacitive humidity sensors to collect multi-source heterogeneous data from isolated submarine power transmission cables. These sensors have strong anti-interference capabilities, especially waterproof, moisture-proof, and corrosion-resistant properties, enabling them to adapt to the harsh environmental conditions of isolated submarine power transmission cables.

[0069] The edge computing layer is equipped with high-performance FPGA chips, such as the Xilinx Spartan-6 series, which have high-performance DSP modules, PCIe and Ethernet interfaces, support parallel computing, and can efficiently process large-scale data. It is used for data pre-cleaning and spatiotemporal dimension alignment of multi-source heterogeneous data.

[0070] The cloud-based analytics layer is equipped with high-performance servers deployed on shore or in central stations to perform deep learning and classification tasks, used for modeling and risk assessment of dynamic risk assessment models for isolated submarine power transmission cables.

[0071] The control and execution layer is used to match the corresponding hierarchical risk control strategies based on the risk assessment results and deploy the corresponding control equipment according to the risk control strategies, such as local monitoring terminals, drone control platforms, AIS systems, SCADA systems and maritime VTS systems.

[0072] Specific limitations regarding the dynamic risk management system for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation can be found in the limitations of the dynamic risk management method for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation mentioned above, and will not be repeated here. Each module in the aforementioned dynamic risk management system for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0073] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application of the technical solution and the constraints involved. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0074] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0075] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0076] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation, characterized in that, The method includes: Collect multi-source heterogeneous data of isolated island power transmission submarine cables, perform data pre-cleaning and spatiotemporal dimension alignment processing on the multi-source heterogeneous data to obtain pre-processed multi-source heterogeneous data. A data cube structure of the multi-source heterogeneous data is constructed, and spatiotemporal cross-correlation analysis is performed on the multi-source heterogeneous data to obtain a spatiotemporal correlation map of the multi-source heterogeneous data; The spatiotemporal correlation graph is input into a pre-constructed dynamic risk assessment model for graph structure definition, and spatiotemporal convolution and feature fusion processing are performed to output the risk probability of isolated submarine power transmission cables. Based on the aforementioned risk probabilities, the entire isolated island power transmission submarine cable is classified into risk levels, and risk management strategies are matched and response processes are executed according to the classified risk levels.

2. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 1, characterized in that, After classifying the entire isolated island power transmission submarine cable into risk levels based on the risk probability, and matching and executing response processing of the graded risk management strategy for the isolated island power transmission submarine cable according to the classified risk levels, the process further includes: The execution response effect of the hierarchical risk management strategy is obtained, and the corresponding risk management strategy is rewarded based on the execution response effect and a preset reward function, and the strategy priority of the risk management strategy is dynamically adjusted. The calculation expression for the reward function is as follows: （1） in, This represents the reward coefficient, used to adjust strategy priority. Indicates the amount of risk reduction. Indicates control costs, , This represents the weighting coefficient.

3. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 1, characterized in that, The process of classifying the risk level of the entire isolated island submarine power transmission cable according to the risk probability, and matching and executing a graded risk management strategy for the isolated island submarine power transmission cable according to the classified risk level, specifically includes: The entire isolated island power transmission submarine cable is divided into grids using a pre-constructed data cube structure, and each grid is aligned in both spatiotemporal dimensions to obtain a spatiotemporally aligned submarine cable grid. Submarine cable risk heatmaps are aggregated according to the risk probability of each submarine cable grid. Based on the aggregation results, continuous submarine cable risk heatmaps are overlaid and predicted. Based on the prediction results, risk levels are classified. Based on the risk level classification results, a graded risk management strategy matching process is performed for each submarine cable grid, and the matched risk management strategy is executed and responded to in each submarine cable grid according to the matching results.

4. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 3, characterized in that, The process of aggregating submarine cable risk heatmaps according to the risk probability of each submarine cable grid, performing overlay prediction processing on continuous submarine cable risk heatmaps based on the aggregation results, and classifying risk levels based on the prediction results also includes: The overlay prediction of the submarine cable risk heat map is also used for submarine cable risk prediction during extreme weather events.

5. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 1, characterized in that, The process of inputting the spatiotemporal correlation graph into a pre-constructed dynamic risk assessment model for graph structure definition, and performing spatiotemporal convolution and feature fusion processing to output the risk probability of isolated submarine power transmission cables specifically includes: The grid attribute vectors in the spatiotemporal correlation graph are defined as graph nodes, and the reciprocal of the grid spatial distance in the spatiotemporal correlation graph is defined as the edge weight, thus obtaining the graph structure definition parameters of the spatiotemporal correlation graph. The spatial features and temporal dependency features of the spatiotemporal correlation map are extracted through the dynamic risk assessment model, and spatiotemporal convolution is performed according to the spatial features and the temporal dependency features to obtain the spatiotemporal features. The spatiotemporal features are sequentially concatenated according to the time dimension and input into the fully connected network layer of the dynamic risk assessment model to output the risk probability of the isolated submarine power transmission cable.

6. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 5, characterized in that, The process of inputting the spatiotemporal correlation graph into a pre-constructed dynamic risk assessment model for graph structure definition, and performing spatiotemporal convolution and feature fusion processing to output the risk probability of isolated submarine power transmission cables, specifically includes: Historical monitoring data containing real accidents involving isolated submarine power transmission cables were acquired. A dynamic risk assessment model was obtained by constructing a neural convolutional network using cross-entropy and spatiotemporal smoothing terms as loss functions and the AdamW optimization algorithm combined with a custom learning rate.

7. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 1, characterized in that, The process of constructing a data cube structure for the multi-source heterogeneous data and performing spatiotemporal cross-correlation analysis on the multi-source heterogeneous data to obtain a spatiotemporal correlation map of the multi-source heterogeneous data specifically includes: Three-dimensional modeling is performed using time, space, and attribute dimensions to construct a data cube structure for the multi-source heterogeneous data. The spatiotemporal lag conditions of the multi-source heterogeneous data are then analyzed based on the data cube structure. Calculate the correlation coefficient matrix of the multi-source heterogeneous data under spatiotemporal lag conditions, and identify key lag times; The Granger causality test is used to determine the variable causal relationship between the key lag time and the corresponding multi-source heterogeneous data, and the spatiotemporal cross-correlation is judged based on the variable causal relationship. Spatiotemporal correlation maps are constructed for multi-source heterogeneous data with spatiotemporal cross-correlation, resulting in spatiotemporal correlation maps of multi-source heterogeneous data.

8. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 1, characterized in that, The process of collecting multi-source heterogeneous data from isolated submarine power transmission cables involves pre-cleaning and spatiotemporal alignment of the multi-source heterogeneous data to obtain pre-processed multi-source heterogeneous data, specifically including: Calculate the average value of the historical data sliding window of the isolated submarine power transmission cable, and set a dynamic threshold range according to the average value of the historical data sliding window to remove multi-source heterogeneous data outside the dynamic threshold range; The multi-source heterogeneous data was processed by cubic spline interpolation to repair missing values, resulting in pre-cleaned multi-source heterogeneous data. The multi-source heterogeneous data is aligned in time dimension according to a preset unified time format, and then aligned in spatial dimension using inverse distance weighted interpolation to obtain preprocessed multi-source heterogeneous data.

9. The method for dynamic risk management of isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation according to claim 8, characterized in that, The process of collecting multi-source heterogeneous data from isolated submarine power transmission cables, pre-cleaning the multi-source heterogeneous data and performing spatiotemporal alignment processing to obtain pre-processed multi-source heterogeneous data, further includes: The multi-source heterogeneous data includes submarine cable temperature data, submarine cable electrical data, marine environmental data, seabed geological activity data, and ship activity data.

10. A dynamic risk management system for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation, characterized in that, The system is applied to the dynamic risk management method for isolated submarine power transmission cables based on spatiotemporal three-dimensional correlation as described in any one of claims 1-9, and the system includes: The front-end sensing layer is equipped with industrial-grade Hall effect voltage sensors, Rogowski coil current sensors, dedicated temperature sensors, and capacitive humidity sensors to collect multi-source heterogeneous data from isolated submarine power transmission cables. The edge computing layer is equipped with high-performance FPGA chips, high-performance DSP modules, PCIe and Ethernet interfaces, and is used for data pre-cleaning and spatiotemporal dimension alignment of multi-source heterogeneous data. The cloud-based analytics layer is equipped with high-performance servers for modeling and risk assessment of dynamic risk assessment models for isolated submarine power transmission cables. The control and execution layer is used to match the corresponding hierarchical risk control strategies based on the risk assessment results and to deploy the corresponding control equipment in accordance with the risk control strategies.

Images