Deep and far sea wind power foundation platform and load integrated system under extreme environment condition

Abstract

The invention discloses a deep and far sea wind power foundation platform and load integrated system under extreme environment conditions, and belongs to the technical field of offshore wind power platforms. The system comprises a basic platform body, a multi-dimensional monitoring subsystem, an environment-load coupling analysis subsystem, an intelligent regulation and control execution subsystem, a safety early warning decision subsystem and a shore-based collaborative management subsystem. Through synchronous acquisition of extreme marine environment, basic platform structure load and operation state data, distributed multi-type sensor layout and high-frequency sampling, data acquisition comprehensiveness and timeliness are realized, monitoring data are fused, a coupling model is constructed, 24-hour load change is predicted, and the real-time performance of the system is improved. According to the method, the attitude of the platform is dynamically adjusted, bearing configuration is optimized, emergency protection is started, the limitation of traditional separation analysis and fixed regulation and control is broken through, and the risk disposal and operation and maintenance efficiency is improved through risk quantification and grading early warning, remote monitoring and intelligent scheduling of a shore-based collaborative management subsystem and four-level early warning accurate matching coping strategies.

Description

Technical Field

[0001] This invention relates to the field of offshore wind power platform technology, and particularly to deep-sea wind power foundation platforms and integrated load systems under extreme environmental conditions. Background Technology

[0002] Traditional basic platforms often adopt a fixed design, which makes it difficult to dynamically adapt to changes in the environment. This can easily lead to problems such as localized load concentration and structural fatigue damage, which in severe cases can cause platform tilting, equipment failure, or even safety accidents.

[0003] Currently, the management and control of deep-sea wind power infrastructure platforms largely rely on decentralized systems: environmental monitoring only collects single meteorological data, load analysis is disconnected from environmental parameters, control execution lacks forward-looking basis, and data interoperability between shore-based management and offshore platforms is poor, resulting in delayed emergency response. Furthermore, data acquisition is susceptible to interference in extreme environments, coupled analysis models lack accuracy, making it difficult to accurately predict load change trends. This leads to weak passive protection capabilities of the platform under extreme conditions such as strong winds and waves, high operation and maintenance costs, and significant safety risks, failing to meet the needs of large-scale, high-quality development of deep-sea wind power. Summary of the Invention

[0004] The purpose of this invention is to provide a deep-sea wind power foundation platform and load-integrated system for extreme environmental conditions, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a deep-sea wind power foundation platform and load-integrated system under extreme environmental conditions, including the foundation platform body, a multi-dimensional monitoring subsystem, an environment-load coupling analysis subsystem, an intelligent control and execution subsystem, a safety early warning decision-making subsystem, and a shore-based collaborative management subsystem;

[0006] The multi-dimensional monitoring subsystem is configured to simultaneously collect extreme marine environment data, basic platform structural load data, and operational status data.

[0007] The environment-load coupling analysis subsystem is configured to perform fusion processing and coupled modeling of data collected by the multi-dimensional monitoring subsystem.

[0008] The intelligent control and execution subsystem is configured to dynamically adjust the structural attitude and load-bearing configuration of the base platform based on the modeling results of the environment-load coupling analysis subsystem.

[0009] The safety early warning decision-making subsystem is configured to generate risk early warning information and push it to the shore-based collaborative management subsystem.

[0010] The multi-dimensional monitoring subsystem communicates unidirectionally with the environment-load coupling analysis subsystem. The environment-load coupling analysis subsystem communicates bidirectionally with the intelligent control and execution subsystem and the safety early warning decision subsystem. The safety early warning decision subsystem communicates bidirectionally with the shore-based collaborative management subsystem. The intelligent control and execution subsystem is connected to the basic platform body for execution.

[0011] Furthermore, the specific data acquisition process of the multi-dimensional monitoring subsystem includes:

[0012] Meteorological stations, wave radars, current meters, and water quality sensors are deployed on the top and around the basic platform to collect data on wind speed, wind direction, significant wave height, wave period, ocean current speed, ocean current direction, seawater temperature, salinity, and sediment content. Time-series environmental data are generated based on a 1-minute sampling interval.

[0013] Fiber optic strain sensors, piezoelectric pressure sensors, and displacement gauges are deployed at key locations such as the foundation platform pile leg welds, tower platform connection nodes, and deck load-bearing beams to collect structural strain, contact pressure, and node displacement, and simultaneously generate a multi-dimensional load data matrix.

[0014] Using high-definition cameras, vibration sensors, and tilt sensors, the system collects data on the positional offset of equipment on the platform deck, structural vibration acceleration, and overall tilt angle of the platform to generate an operational status dataset.

[0015] The preprocessed time-series environmental data sequence, multidimensional load data matrix, and operating status dataset are integrated to obtain a standardized monitoring dataset, which includes environmental data, load data, and status data.

[0016] Furthermore, the specific processing steps of the environment-load coupling analysis subsystem include:

[0017] Based on the improved DS evidence theory, the standardized monitoring dataset is fused, the basic probability assignment function of environmental data, load data and state data is defined, and the fused data is obtained through evidence synthesis.

[0018] Construct an environment-load dual-drive coupled model, using environmental parameters from the fused data as the input vector and load parameters as the output vector, and train the model using a deep belief network.

[0019] The input layer corresponds to environmental parameters; the number of neurons in each hidden layer is determined by particle swarm optimization algorithm, and is 16, 8, and 4 respectively; the output layer corresponds to load parameters.

[0020] The model accuracy was verified by the root mean square error (RMSE), and the final coupled model was determined when the RMSE was less than 5%.

[0021] Based on the final coupled model, the system takes into account the environmental forecast data for the next 24 hours, predicts the load forecast data for the corresponding time, and generates a load change trend curve.

[0022] Furthermore, the specific control process of the intelligent control execution subsystem includes:

[0023] Based on the overall platform tilt angle and ocean current direction, the hydraulic expansion joint at the pile legs is controlled to establish a tilt angle-expansion amount mapping relationship:

[0024]

[0025] in, Let i be the extension / retraction amount of the i-th pile leg. For the overall tilt angle of the platform, According to the direction of the ocean current, This is the tilt adjustment coefficient. Let be the azimuth angle of the i-th pile leg. The reference expansion / contraction amount is collected in real time via a displacement feedback sensor. PID algorithm closed-loop control is adopted to make Maintain within the range of [-0.5°, 0.5°];

[0026] Based on load prediction data, a genetic algorithm is used to optimize the layout of deck equipment. The objective function is to minimize the maximum structural strain, and the constraints are equipment weight and equipment spacing. The optimal layout scheme is output, and the deck robotic arm is controlled to adjust the position of the equipment.

[0027] Calculate the comprehensive environmental intensity index:

[0028]

[0029] in, , and As weight, For wind speed, Design limits for wind speed. For the effective wave height, The effective wave height is the design limit value. For ocean current speed, Design limit values ​​for ocean current speed; when the comprehensive environmental intensity index When necessary, activate the emergency mode, shut down non-essential equipment on the platform, activate the anti-lateral displacement damper, tighten the pile leg locking device, and reduce the platform's windward area.

[0030] Furthermore, the specific working process of the security early warning decision subsystem includes:

[0031] A risk assessment indicator system was constructed, including environmental risk, structural risk and operational risk. The weights were determined and the comprehensive risk index was calculated using the analytic hierarchy process.

[0032] Early warning levels are classified based on a comprehensive risk index, including Level 1, Level 2, Level 3, and Level 4.

[0033] A response decision database is generated for different early warning levels.

[0034] Furthermore, the specific functions of the shore-based collaborative management subsystem include:

[0035] The system receives real-time monitoring data, control command execution results, and early warning information uploaded by the basic platform via satellite communication and a maritime 5G dual-mode network.

[0036] Build a 3D visualization platform to render the 3D model of the basic platform in real time, overlay and display environmental parameters, load data and warning levels, and generate daily and weekly data reports.

[0037] When a Level 4 warning is received, the system automatically associates information on shore-based maintenance vessels, helicopters, and spare parts depots, plans the optimal rescue route based on the Dijkstra algorithm, generates a dispatch plan, and pushes it to the relevant maintenance units, while also recording the warning processing process.

[0038] Furthermore, the fiber optic strain sensor adopts a distributed deployment method, with one monitoring point deployed every 1 meter along the pile leg axis. Each monitoring point integrates three strain sensing units in different directions to collect axial strain, circumferential strain, and shear strain respectively. Distributed acquisition and positioning of strain data are achieved through optical time-domain reflectometry.

[0039] Furthermore, the training process for deep belief networks includes:

[0040] Randomly initialize the weights and biases of each layer, and set the learning rate and number of iterations;

[0041] An unsupervised learning approach is used to train each layer of the restricted Boltzmann machine to minimize the reconstruction error of the input data.

[0042] A supervised learning approach is adopted, with the error between the predicted and actual values ​​of load parameters as the objective function. The weights and biases of the entire network are adjusted through the backpropagation algorithm until the number of iterations or the error is less than a preset threshold is reached.

[0043] Furthermore, the anti-lateral displacement damper adopts a magnetorheological damper, which adjusts the damping force in real time based on the horizontal load in the load prediction data; the damping force is continuously adjustable by adjusting the magnetic field strength inside the damper through a current controller.

[0044] Furthermore, the process of determining weights using the analytic hierarchy process includes:

[0045] Construct a judgment matrix, calculate the largest eigenvalue and corresponding eigenvector of the judgment matrix, normalize the eigenvector to obtain the weight vector;

[0046] Perform a consistency check, calculate the consistency index, and find the average random consistency index. When the ratio of the consistency index to the average random consistency index is less than 0.1, the judgment matrix meets the consistency requirement and the weights are effective.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] 1. The multi-dimensional monitoring subsystem of this invention synchronously collects data on extreme marine environments, basic platform structural loads, and operational status. After preprocessing, the data is integrated into a standardized dataset. Through distributed multi-type sensor layout and high-frequency sampling, the system achieves comprehensive and timely data collection. Combined with error elimination, it solves the problems of high data interference and low accuracy in extreme environments, providing a high-quality data foundation for subsequent coupled analysis, ensuring the reliability of data analysis and decision-making throughout the system, and avoiding control deviations caused by data defects.

[0049] 2. The environment-load coupling analysis subsystem of this invention integrates monitoring data and constructs a coupling model to predict 24-hour load changes. The intelligent control and execution subsystem dynamically adjusts the platform attitude, optimizes the load configuration, and initiates emergency protection based on this. It breaks through the limitations of traditional separate analysis and fixed control. The model accurately associates the environment and load. PID control and genetic algorithms achieve attitude stability and load optimization. The magnetorheological damper effectively resists impact, greatly improving the platform's adaptability to extreme environments, avoiding post-event response lag, and reducing the risk of structural damage.

[0050] 3. The safety early warning decision-making subsystem of this invention quantifies risks and provides graded early warnings, while the shore-based collaborative management subsystem provides remote monitoring and intelligent scheduling. Safety early warnings are scientifically weighted using the analytic hierarchy process, and the four-level early warning system accurately matches response strategies. Shore-based dual-mode communication ensures data transmission, and three-dimensional visualization intuitively presents the status. In emergencies, it automatically plans rescue routes, solving the problems of subjective early warnings and lagging shore-based scheduling, improving the efficiency of risk handling and operation and maintenance, and providing support for the platform's full lifecycle security. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the real-time load control method for deep-sea wind power foundation platforms under extreme environments according to the present invention. Detailed Implementation

[0052] 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 embodiments of the present invention, and not all embodiments. 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.

[0053] Please see Figure 1 The present invention provides the following technical solutions:

[0054] The deep-sea wind power foundation platform and load-integrated system under extreme environmental conditions includes the foundation platform itself, a multi-dimensional monitoring subsystem, an environment-load coupling analysis subsystem, an intelligent control and execution subsystem, a safety early warning decision-making subsystem, and a shore-based collaborative management subsystem.

[0055] The multi-dimensional monitoring subsystem is configured to simultaneously collect extreme marine environment data, basic platform structural load data, and operational status data.

[0056] The environment-load coupling analysis subsystem is configured to perform fusion processing and coupled modeling of data collected by the multi-dimensional monitoring subsystem.

[0057] The intelligent control and execution subsystem is configured to dynamically adjust the structural attitude and load-bearing configuration of the base platform based on the modeling results of the environment-load coupling analysis subsystem.

[0058] The safety early warning decision-making subsystem is configured to generate risk early warning information and push it to the shore-based collaborative management subsystem to achieve full life-cycle safety management and control of deep-sea wind power foundation platforms in extreme environments;

[0059] The multi-dimensional monitoring subsystem communicates unidirectionally with the environment-load coupling analysis subsystem. The environment-load coupling analysis subsystem communicates bidirectionally with the intelligent control and execution subsystem and the safety early warning decision-making subsystem. The safety early warning decision-making subsystem communicates bidirectionally with the shore-based collaborative management subsystem. The intelligent control and execution subsystem is connected to the basic platform body for execution.

[0060] The specific data collection process of the multi-dimensional monitoring subsystem includes:

[0061] Meteorological stations, wave radars, current meters, and water quality sensors are deployed on the top and around the basic platform to collect data on wind speed, wind direction, significant wave height, wave period, ocean current speed, ocean current direction, seawater temperature, salinity, and sediment content. Time-series environmental data are generated based on a 1-minute sampling interval.

[0062] Fiber optic strain sensors, piezoelectric pressure sensors, and displacement gauges are deployed at key locations such as the foundation platform pile leg welds, tower platform connection nodes, and deck load-bearing beams to collect structural strain, contact pressure, and node displacement, and simultaneously generate a multi-dimensional load data matrix.

[0063] The fiber optic strain sensor adopts a distributed deployment method, with one monitoring point deployed every 1 meter along the pile leg axis. Each monitoring point integrates three strain sensing units in different directions to collect axial strain, circumferential strain and shear strain respectively. Distributed acquisition and positioning of strain data are realized through optical time domain reflection technology.

[0064] Using high-definition cameras, vibration sensors, and tilt sensors, the system collects data on the positional offset of equipment on the platform deck, structural vibration acceleration, and overall tilt angle of the platform to generate an operational status dataset.

[0065] The preprocessed time-series environmental data sequence, multidimensional load data matrix, and operational status dataset are integrated to obtain a standardized monitoring dataset, which includes environmental data, load data, and status data.

[0066] In the above embodiments, the multi-dimensional monitoring subsystem achieves comprehensiveness, accuracy and effectiveness of data acquisition in extreme deep-sea environments through multi-type, distributed sensor layout and standardized data preprocessing. It simultaneously covers three core dimensions: environment, load and operating status. High-frequency sampling ensures data timeliness, and the distributed fiber optic strain sensor's positioning accuracy and multi-directional strain acquisition capability can accurately capture subtle structural changes in key parts of the basic platform.

[0067] The specific processing steps of the environment-load coupling analysis subsystem include:

[0068] Based on the improved DS evidence theory, the standardized monitoring dataset is fused, the basic probability assignment function of environmental data, load data and state data is defined, and the fused data is obtained through evidence synthesis.

[0069] A dual-drive coupled environment-load model is constructed, with environmental parameters from the fused data as the input vector and load parameters as the output vector. The model is trained using a deep belief network.

[0070] The training process of a deep belief network includes:

[0071] Randomly initialize the weights and biases of each layer, and set the learning rate and number of iterations;

[0072] An unsupervised learning approach is used to train each layer of the restricted Boltzmann machine to minimize the reconstruction error of the input data.

[0073] A supervised learning approach is adopted, with the error between the predicted and actual values ​​of load parameters as the objective function. The weights and biases of the entire network are adjusted through the backpropagation algorithm until the number of iterations or the error is less than a preset threshold is reached.

[0074] The input layer corresponds to environmental parameters; the number of neurons in each hidden layer is determined by particle swarm optimization algorithm, and is 16, 8, and 4 respectively; the output layer corresponds to load parameters.

[0075] The model accuracy was verified by the root mean square error (RMSE), and the final coupled model was determined when the RMSE was less than 5%.

[0076] Based on the final coupled model, the system takes into account the environmental forecast data for the next 24 hours, predicts the load forecast data for the corresponding time, and generates a load change trend curve.

[0077] In the above embodiments, the environment-load coupling analysis subsystem overcomes the limitations of traditional separate analysis of environment and load by improving the DS evidence theory data fusion and deep belief network coupling modeling, realizing the dynamic correlation and accurate prediction between the two. The improved DS evidence theory can effectively integrate multi-source heterogeneous monitoring data, solve the problem of data uncertainty, and improve data credibility. The DBN model optimizes the number of hidden layer nodes through particle swarm optimization algorithm, and combines unsupervised pre-training and supervised fine-tuning to enable the model to have a strong nonlinear fitting ability. The load prediction results can be accurately output by inputting environmental parameters. The 24-hour load trend prediction based on the model provides a forward-looking basis for intelligent regulation, avoids the lag of traditional post-event response, and significantly improves the basic platform's ability to predict and actively protect against extreme environments.

[0078] The specific control process of the intelligent control execution subsystem includes:

[0079] Based on the overall platform tilt angle and ocean current direction, the hydraulic expansion joint at the pile legs is controlled to establish a tilt angle-expansion amount mapping relationship:

[0080]

[0081] in, Let i be the extension / retraction amount of the i-th pile leg. For the overall tilt angle of the platform, According to the direction of the ocean current, This is the tilt adjustment coefficient. Let be the azimuth angle of the i-th pile leg. The reference expansion / contraction amount is collected in real time via a displacement feedback sensor. PID algorithm closed-loop control is adopted to make Maintain within the range of [-0.5°, 0.5°];

[0082] Based on load prediction data, a genetic algorithm is used to optimize the layout of deck equipment. The objective function is to minimize the maximum structural strain, and the constraints are equipment weight and equipment spacing. The optimal layout scheme is output, and the deck robotic arm is controlled to adjust the position of the equipment.

[0083] Calculate the comprehensive environmental intensity index:

[0084]

[0085] in, , and As weight, For wind speed, Design limits for wind speed. For the effective wave height, The effective wave height is the design limit value. For ocean current speed, Design limit values ​​for ocean current speed; when the comprehensive environmental intensity index When necessary, activate the emergency mode, shut down non-essential equipment on the platform, activate the anti-lateral displacement damper, tighten the pile leg locking device, and reduce the platform's windward area;

[0086] The anti-lateral displacement damper uses a magnetorheological damper, which adjusts the damping force in real time based on the horizontal load in the load prediction data; the damping force is continuously adjustable by adjusting the magnetic field strength inside the damper through a current controller.

[0087] In the above embodiments, the intelligent control and execution subsystem achieves attitude stability, load optimization, and safety protection of the basic platform under extreme environments through multi-dimensional dynamic control strategies and emergency protection mechanisms. It has stronger adaptability and flexibility. The tilt angle-extension mapping relationship combined with PID closed-loop control can accurately control the platform tilt angle and ensure the stability of the platform attitude. The genetic algorithm optimizes the deck equipment layout with the goal of minimizing the maximum structural strain, effectively reducing the risk of local load concentration. The emergency mode activated when the comprehensive environmental intensity index is >0.8 rapidly improves the platform's ability to resist extreme environments through multi-measure coordinated response. The magnetorheological damper has a real-time adjustable damping force based on horizontal loads and a response time ≤50ms, which can efficiently absorb impact loads and solve the problems of fixed damping force and limited impact resistance of traditional dampers, thus comprehensively ensuring the structural safety and stable operation of the basic platform.

[0088] The specific working process of the safety early warning decision subsystem includes:

[0089] A risk assessment indicator system was constructed, including environmental risk, structural risk and operational risk. The weights were determined and the comprehensive risk index was calculated using the analytic hierarchy process.

[0090] The process of determining weights using the analytic hierarchy process includes:

[0091] Construct a judgment matrix, calculate the largest eigenvalue and corresponding eigenvector of the judgment matrix, normalize the eigenvector to obtain the weight vector;

[0092] Perform a consistency check, calculate the consistency index, and find the average random consistency index. When the ratio of the consistency index to the average random consistency index is less than 0.1, the matrix is ​​considered to meet the consistency requirements and the weights are valid.

[0093] The warning levels are divided based on a comprehensive risk index, including Level 1, Level 2, Level 3, and Level 4.

[0094] A response decision database is generated for different early warning levels.

[0095] In the above embodiments, the safety early warning decision subsystem achieves accurate quantification and efficient response to risks of the basic platform under extreme environments through a risk assessment system and a hierarchical decision-making mechanism. It overcomes the shortcomings of traditional early warning systems, such as subjective risk assessment and poor decision-making targeting. When determining weights using the analytic hierarchy process, the system ensures the scientific and reasonable allocation of weights through the construction of a judgment matrix and consistency checks. The comprehensive risk index integrates the three major risk dimensions of environment, structure, and operation to achieve comprehensive risk quantification. The four-level early warning classification clearly defines the degree of risk and corresponds to different response decision libraries, making the early warning response more targeted and avoiding the problems of over-response or under-response. The decision generation combines fuzzy inference rules to optimize priorities, ensuring that key response measures are executed first.

[0096] The specific functions of the shore-based collaborative management subsystem include:

[0097] The system receives real-time monitoring data, control command execution results, and early warning information uploaded by the basic platform via satellite communication and a maritime 5G dual-mode network.

[0098] Build a 3D visualization platform to render the 3D model of the basic platform in real time, overlay and display environmental parameters, load data and warning levels, and generate daily and weekly data reports.

[0099] When a Level 4 warning is received, the system automatically associates information on shore-based maintenance vessels, helicopters, and spare parts depots, plans the optimal rescue route based on the Dijkstra algorithm, generates a dispatch plan, and pushes it to the relevant maintenance units, while also recording the warning processing process.

[0100] In the above embodiments, the shore-based collaborative management subsystem achieves remote real-time control and emergency coordination of the deep-sea wind power infrastructure platform through dual-mode communication, three-dimensional visualization, and intelligent scheduling. This breaks through the bottlenecks of traditional shore-based management data silos and scheduling lags. Satellite communication and the offshore 5G dual-mode network ensure uninterrupted communication in extreme environments. The three-dimensional visualization platform restores the basic platform structure 1:1, overlays multi-dimensional data in real time, and intuitively presents the platform status. Daily and weekly reports facilitate operation and maintenance analysis. When a level 4 warning is received, the system automatically associates operation and maintenance resources and plans the optimal rescue route, realizing rapid scheduling and closed-loop management of emergency resources. This solves the problems of slow operation and maintenance response and difficult resource coordination in deep-sea areas, and improves the overall operation and maintenance efficiency and emergency response capabilities of the system.

[0101] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. An integrated system for deep-sea wind power foundation platforms and loads under extreme environmental conditions, characterized in that, It includes the basic platform body, multi-dimensional monitoring subsystem, environment-load coupling analysis subsystem, intelligent control and execution subsystem, safety early warning decision-making subsystem, and shore-based collaborative management subsystem; The multi-dimensional monitoring subsystem is configured to simultaneously collect extreme marine environment data, basic platform structural load data, and operational status data. The environment-load coupling analysis subsystem is configured to perform fusion processing and coupled modeling of data collected by the multi-dimensional monitoring subsystem. The intelligent control and execution subsystem is configured to dynamically adjust the structural attitude and load-bearing configuration of the base platform based on the modeling results of the environment-load coupling analysis subsystem. The safety early warning decision-making subsystem is configured to generate risk early warning information and push it to the shore-based collaborative management subsystem. The multi-dimensional monitoring subsystem communicates unidirectionally with the environment-load coupling analysis subsystem. The environment-load coupling analysis subsystem communicates bidirectionally with the intelligent control and execution subsystem and the safety early warning decision subsystem. The safety early warning decision subsystem communicates bidirectionally with the shore-based collaborative management subsystem. The intelligent control and execution subsystem is connected to the basic platform body for execution.

2. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 1, characterized in that, The specific data acquisition process of the multi-dimensional monitoring subsystem includes: Meteorological stations, wave radars, current meters, and water quality sensors are deployed on the top and around the basic platform to collect data on wind speed, wind direction, significant wave height, wave period, ocean current speed, ocean current direction, seawater temperature, salinity, and sediment content. Time-series environmental data are generated based on a 1-minute sampling interval. Fiber optic strain sensors, piezoelectric pressure sensors, and displacement gauges are installed at the weld joints of the foundation platform pile legs, the connection nodes of the tower platform, and the load-bearing beams of the deck to collect structural strain, contact pressure, and node displacement, and simultaneously generate a multi-dimensional load data matrix. Using high-definition cameras, vibration sensors, and tilt sensors, the system collects data on the positional offset of equipment on the platform deck, structural vibration acceleration, and overall tilt angle of the platform to generate an operational status dataset. The preprocessed time-series environmental data sequence, multidimensional load data matrix, and operating status dataset are integrated to obtain a standardized monitoring dataset, which includes environmental data, load data, and status data.

3. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 1, characterized in that, The specific processing steps of the environment-load coupling analysis subsystem include: Based on the improved DS evidence theory, the standardized monitoring dataset is fused, the basic probability assignment function of environmental data, load data and state data is defined, and the fused data is obtained through evidence synthesis. Construct an environment-load dual-drive coupled model, using environmental parameters from the fused data as the input vector and load parameters as the output vector, and train the model using a deep belief network. The input layer corresponds to environmental parameters; the number of neurons in each hidden layer is determined by particle swarm optimization algorithm, and is 16, 8, and 4 respectively; the output layer corresponds to load parameters. The model accuracy was verified by the root mean square error (RMSE), and the final coupled model was determined when the RMSE was less than 5%. Based on the final coupled model, the system takes into account the environmental forecast data for the next 24 hours, predicts the load forecast data for the corresponding time, and generates a load change trend curve.

4. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 1, characterized in that, The specific control process of the intelligent control execution subsystem includes: Based on the overall platform tilt angle and ocean current direction, the hydraulic expansion joint at the pile legs is controlled to establish a tilt angle-expansion amount mapping relationship: ； in, Let i be the extension / retraction amount of the i-th pile leg. For the overall tilt angle of the platform, According to the direction of the ocean current, This is the tilt adjustment coefficient. Let be the azimuth angle of the i-th pile leg. The reference expansion / contraction amount is collected in real time via a displacement feedback sensor. PID algorithm closed-loop control is adopted to make Maintain within the range of [-0.5°, 0.5°]; Based on load prediction data, a genetic algorithm is used to optimize the layout of deck equipment. The objective function is to minimize the maximum structural strain, and the constraints are equipment weight and equipment spacing. The optimal layout scheme is output, and the deck robotic arm is controlled to adjust the position of the equipment. Calculate the comprehensive environmental intensity index: ； in, , and As weight, For wind speed, Design limits for wind speed. For the effective wave height, The effective wave height is the design limit value. For ocean current speed, Design limit values ​​for ocean current speed; when the comprehensive environmental intensity index When necessary, activate the emergency mode, shut down non-essential equipment on the platform, activate the anti-lateral displacement damper, tighten the pile leg locking device, and reduce the platform's windward area.

5. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 1, characterized in that, The specific working process of the security early warning decision subsystem includes: A risk assessment indicator system was constructed, including environmental risk, structural risk and operational risk. The weights were determined and the comprehensive risk index was calculated using the analytic hierarchy process. Early warning levels are classified based on a comprehensive risk index, including Level 1, Level 2, Level 3, and Level 4. A response decision database is generated for different early warning levels.

6. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 1, characterized in that, The specific functions of the shore-based collaborative management subsystem include: The system receives real-time monitoring data, control command execution results, and early warning information uploaded by the basic platform via satellite communication and a maritime 5G dual-mode network. Build a 3D visualization platform to render the 3D model of the basic platform in real time, overlay and display environmental parameters, load data and warning levels, and generate daily and weekly data reports. When a Level 4 warning is received, the system automatically associates information on shore-based maintenance vessels, helicopters, and spare parts depots, plans the optimal rescue route based on the Dijkstra algorithm, generates a dispatch plan, and pushes it to the relevant maintenance units, while also recording the warning processing process.

7. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 2, characterized in that, The fiber optic strain sensor adopts a distributed deployment method, with one monitoring point deployed every 1 meter along the pile leg axis. Each monitoring point integrates three strain sensing units in different directions to collect axial strain, circumferential strain and shear strain respectively. Distributed acquisition and positioning of strain data are realized through optical time domain reflection technology.

8. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 3, characterized in that, The training process of a deep belief network includes: Randomly initialize the weights and biases of each layer, and set the learning rate and number of iterations; An unsupervised learning approach is used to train each layer of the restricted Boltzmann machine to minimize the reconstruction error of the input data. A supervised learning approach is adopted, with the error between the predicted and actual values ​​of load parameters as the objective function. The weights and biases of the entire network are adjusted through the backpropagation algorithm until the number of iterations or the error is less than a preset threshold is reached.

9. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 4, characterized in that, The anti-lateral displacement damper is a magnetorheological damper, which adjusts the damping force in real time based on the horizontal load in the load prediction data; the damping force is continuously adjustable by adjusting the magnetic field strength inside the damper through a current controller.

10. The deep-sea wind power foundation platform and integrated load system under extreme environmental conditions as described in claim 5, characterized in that, The process of determining weights using the analytic hierarchy process includes: Construct a judgment matrix, calculate the largest eigenvalue and corresponding eigenvector of the judgment matrix, normalize the eigenvector to obtain the weight vector; Perform a consistency check, calculate the consistency index, and find the average random consistency index. When the ratio of the consistency index to the average random consistency index is less than 0.1, the judgment matrix meets the consistency requirement and the weights are effective.

Images