A cloud platform system for offshore wind power construction

Abstract

The present invention relates to the technical field of offshore wind power construction, and discloses a cloud platform system for offshore wind power construction, which includes a multi-dimensional data fusion module for collecting and fusing ship, pile hammer and environmental data; a construction path optimization module for modeling the ship's navigation trajectory and optimizing the path; a resource dynamic scheduling module for decomposing tasks and scheduling resources; a safety situation awareness module for extracting equipment abnormality features and generating early warnings; and a feedback closed-loop optimization module for correcting task scheduling results. In addition, it also includes an analysis software integration module for integrating multiple analysis software and realizing data intercommunication; and a geographic information service module for realizing multi-dimensional spatial data visualization. This cloud platform system realizes the intelligent management of offshore wind power construction, improves construction efficiency and safety, reduces construction costs and risks, and promotes the development of the offshore wind power industry.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore wind power construction, and in particular to a cloud platform system for offshore wind power construction. Background Art

[0002] Offshore wind power, a clean, renewable energy source, has experienced rapid growth worldwide in recent years. As offshore wind power projects continue to expand in scale and construction environments become increasingly complex, traditional construction management methods have gradually exposed numerous problems and are no longer able to meet the needs of the industry's rapid development.

[0003] In terms of data management, offshore wind turbine construction involves numerous different types of data, such as vessel data, pile hammer data, and environmental monitoring data. This data comes from a wide range of sources and in various formats, making it heterogeneous. Previously, a lack of effective data collection and integration methods resulted in fragmented data, hindering the formation of comprehensive and accurate construction information. This made it difficult for construction personnel to comprehensively understand the construction status, resulting in a lack of sufficient data support for construction decision-making, increasing construction risks. For example, in complex sea conditions, the inability to timely integrate and analyze vessel navigation data with environmental monitoring data could result in the vessel being unable to adjust its navigation strategy in the event of severe weather, compromising construction safety.

[0004] Construction route planning is also a major challenge. The marine environment of offshore wind farms is complex and subject to numerous restrictions, including restricted areas such as shoals, reefs, and other operating areas, as well as constantly changing weather conditions and ocean currents. Traditional path planning methods are often designed based on static environments and are unable to adapt to the dynamically changing construction environment in real time. When faced with sudden weather changes or adjustments to construction tasks, it is difficult to quickly and effectively optimize the construction route. This not only prolongs the construction period but can also cause safety accidents such as equipment collisions, leading to economic losses and project delays.

[0005] Resource scheduling is also a critical issue. Offshore wind power construction projects involve numerous tasks, placing significant and complex demands on various resources, including vessels, equipment, and manpower. Different construction tasks have varying resource requirements and priorities, but previous resource scheduling methods lacked a scientific and rational planning model, preventing efficient and dynamic scheduling based on task priorities and real-time resource status. This often leads to uneven resource allocation, with some construction tasks sitting idle and wasted, while tasks requiring urgent resources are left unsettled, severely impacting construction progress and overall efficiency.

[0006] Safety management is paramount in offshore wind turbine construction. Due to the harsh offshore construction environment and the complex ocean conditions under which equipment operates, failures and anomalies are prone to occur. However, traditional safety monitoring methods are limited and outdated, making it difficult to monitor and analyze equipment status in real time and comprehensively. This inability to promptly identify potential safety hazards increases the risk of accidents. Once an accident occurs, it not only results in casualties and property damage, but also negatively impacts the reputation and future development of the entire project.

[0007] Furthermore, during the construction process, various analytical software, such as pile insertion and extraction calculation and analysis software and pile driveability analysis software, play an important role in assisting construction decision-making. However, these software programs are often independent of each other, preventing data from being shared or interoperable, making user operation inconvenient. Furthermore, the lack of effective geographic information services makes it difficult for construction personnel to intuitively understand the relationship between wind farm geographic information, real-time ship locations, and geological exploration data, hindering the overall planning and management of construction. Summary of the Invention

[0008] The purpose of the present invention is to provide a cloud platform system for offshore wind power construction to solve the problems raised in the above background technology.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a cloud platform system for offshore wind power construction, the system comprising:

[0010] The multidimensional data fusion module is used to collect and process ship data, pile hammer data, and environmental monitoring data from offshore wind power construction in real time through heterogeneous data acquisition interfaces to obtain the original construction data set; it uses a data cleaning algorithm to remove noise from the original construction data set to obtain cleaned construction data; and it uses a spatiotemporal correlation fusion algorithm to perform multidimensional data fusion processing on the cleaned construction data to obtain fused construction data;

[0011] The construction path optimization module is used to model the ship's navigation trajectory in the integrated construction data using a dynamic path planning algorithm to obtain an initial construction path. It also uses a constrained optimization algorithm combined with the wind farm's sea area topology to perform conflict detection and correction on the initial construction path to obtain an optimized construction path.

[0012] The resource dynamic scheduling module is used to decompose the optimized construction path based on the construction task priority model and generate a set of construction subtasks. It also uses the dynamic priority scheduling algorithm to match and schedule resources for the construction subtasks and obtain the task scheduling results.

[0013] The security situation awareness module is used to extract abnormal features from the device status data in the task scheduling results through real-time stream data processing technology to obtain a device abnormal feature set; it also calculates the security risk probability of the device abnormal feature set and generates a security warning instruction;

[0014] The feedback closed-loop optimization module is used to parse and process the feedback data of safety warning instructions through an incremental learning algorithm to obtain feedback optimization parameters; and use an adaptive adjustment algorithm combined with feedback optimization parameters to perform real-time correction processing on the task scheduling results to generate a closed-loop optimization scheduling strategy.

[0015] Preferably, the multidimensional data fusion module includes:

[0016] Through the heterogeneous data acquisition interface, ship data, pile hammer data and environmental monitoring data are converted into standardized formats to obtain structured construction data;

[0017] Interpolate and filter missing values ​​and outliers in structured construction data to obtain clean construction data;

[0018] The spatiotemporal correlation fusion algorithm is used to perform correlation matching on the spatiotemporal labels of the cleaning construction data to generate a spatiotemporal correlation matrix.

[0019] The multi-dimensional data is weightedly fused based on the spatiotemporal correlation matrix to obtain fused construction data.

[0020] Preferably, the function formula of the spatiotemporal correlation fusion algorithm is as follows:

[0021]

[0022] Where F is the fusion construction data, For the Time series data, is the time weight coefficient, is the time decay factor, is the time interval, For the Spatial data, is the spatial weight coefficient, is the spatial distance correlation function, is the number of time series data, is the number of spatial data.

[0023] Preferably, the construction path optimization module includes:

[0024] The ship's navigation trajectory is segmented and modeled using a dynamic path planning algorithm to generate an initial path node set.

[0025] The constraint optimization algorithm is used in combination with the coordinates of the prohibited sea area to perform collision detection on the initial path node set and eliminate conflicting nodes. The function formula of the constraint optimization algorithm is as follows:

[0026]

[0027] Where, To optimize the construction path, is the cost function of the k-th constraint condition, is the constraint weight coefficient, is the path smoothing coefficient, is the initial path node set, P is the variable of the path node set, representing the path candidate to be optimized, is the number of constraints;

[0028] Path smoothing is performed based on the remaining path node set to generate an optimized construction path.

[0029] Preferably, the resource dynamic scheduling module includes:

[0030] The urgency and resource consumption of the construction subtask set are evaluated and processed through the task priority model to generate a task priority sequence;

[0031] The dynamic priority scheduling algorithm is combined with the real-time position data of the ship to perform resource allocation processing on the task priority sequence to obtain the task scheduling result; the function formula of the dynamic priority scheduling algorithm is as follows:

[0032]

[0033] Where, For the Dynamic priority of tasks, For the task urgency, The deadline for the task, is the real-time available resources, and is the adjustment parameter, and t is the current system time.

[0034] Preferably, the security situation awareness module includes:

[0035] The device status data is sampled and processed by sliding windows through real-time stream data processing technology to obtain device status time series data;

[0036] Use convolutional neural networks to extract abnormal features from device status time series data and generate a device abnormal feature set;

[0037] Based on the Markov chain model, the risk probability evolution of the equipment abnormality feature set is calculated to generate safety warning instructions.

[0038] Preferably, the feedback closed-loop optimization module includes:

[0039] The incremental learning algorithm is used to extract the features of the equipment abnormality type and occurrence frequency in the safety warning instructions and generate feedback feature vectors;

[0040] Based on the gradient descent method, the parameter update direction of the feedback feature vector is calculated and processed to obtain the feedback optimization parameters;

[0041] The feedback optimization parameters are compared and verified with the preset weight threshold interval to screen the effective optimization parameter set.

[0042] Preferably, the feedback closed-loop optimization module further includes:

[0043] An adaptive adjustment algorithm is used to calculate multi-dimensional correction coefficients for task priority weights, resource allocation weights, and ship path weights in the task scheduling results. The correction coefficients are dynamically adjusted in combination with the feedback optimization parameter set to generate an updated scheduling weight matrix.

[0044] The scheduling weight matrix is ​​matched with the real-time construction task data through the weighted fusion method to generate a closed-loop optimization scheduling strategy.

[0045] Preferably, the system further comprises:

[0046] An analysis software integration module is used to deploy pile insertion and extraction calculation and analysis software, pile driveability analysis software, floating vessel operability prediction software, pile foundation penetration analysis software, and wind farm resource scheduling software on a cloud platform through cloud integration or offline call, and to achieve cross-software data interoperability through a unified data protocol;

[0047] In the analysis software integration module:

[0048] The cloud integration method is to deploy the software on a cloud platform virtual machine. Users can directly operate the software through single sign-on. The calculation results are automatically uploaded to the document database after approval by the technical center.

[0049] The offline call method is that the user submits a data form to the cloud platform, and the technical center operates the local computing node to generate a report and transmits it back to the cloud platform;

[0050] Cross-software data interoperability is achieved through a unified data identifier, which includes the wind farm code OW+4 digits, the machine position code TK+5 digits and the ship's unique ID.

[0051] Preferably, the system further comprises:

[0052] The geographic information service module integrates wind farm geographic information, real-time ship AIS data and geological exploration data based on WebGIS technology, and uses tiled map services and dynamic layer overlay technology to achieve multi-dimensional spatial data visualization.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] The multidimensional data fusion module plays a key role in data processing and integration. Through heterogeneous data acquisition interfaces, the module collects ship data, pile hammer data, and environmental monitoring data in real time and converts them into standardized formats, resolving data heterogeneity. Data cleaning algorithms remove noise to ensure data accuracy and reliability. Using spatiotemporal correlation fusion algorithms, multidimensional data is integrated to generate comprehensive and accurate fused construction data. This enables construction personnel to obtain more complete and valuable construction information, providing a solid data foundation for subsequent construction decisions. For example, by integrating ship navigation data with environmental monitoring data, construction personnel can predict the impact of severe weather on construction in advance, adjust construction plans in a timely manner, avoid safety incidents, and improve construction efficiency.

[0055] The construction path optimization module significantly enhances the scientific nature and flexibility of construction path planning. A dynamic path planning algorithm is used to model the ship's navigation trajectory, generating an initial construction path. Combined with the wind farm's sea area topology, a constrained optimization algorithm is used for conflict detection and correction. This approach fully considers factors such as prohibited areas and weather conditions in the sea area, effectively avoiding ship collisions and ensuring construction safety. At the same time, the path is smoothed to facilitate smoother navigation, reduce energy consumption, and lower construction costs. In actual construction, when encountering sudden weather changes or new construction tasks, the module can quickly adjust the path to ensure smooth construction progress, significantly shortening the construction period.

[0056] The dynamic resource scheduling module enables efficient and rational resource allocation. Based on the construction task priority model, construction tasks are broken down, the urgency and resource consumption of each subtask are assessed, and a task priority sequence is generated. In combination with real-time vessel position data, a dynamic priority scheduling algorithm is used to match and schedule resources, ensuring that resources are allocated promptly and accurately to the construction tasks that need them most. This not only improves resource utilization and avoids idleness and waste of resources, but also rationally arranges resources based on the urgency of the task, ensuring the smooth completion of critical tasks and effectively safeguarding the construction progress. For example, when multiple construction tasks are being carried out simultaneously, the module can prioritize resource allocation for urgent and important tasks, avoiding construction delays due to insufficient resources.

[0057] The safety situation awareness module provides a strong guarantee for construction safety. By extracting abnormal features from equipment status data using real-time streaming data processing technology and applying convolutional neural networks and Markov chain models for analysis and calculation, it can promptly and accurately identify potential safety hazards and generate safety warnings. Construction personnel can take timely action based on these warnings to maintain and repair equipment, effectively preventing accidents and ensuring the safety of construction workers and the normal operation of equipment. This significantly reduces the economic losses and construction delays caused by safety accidents, and improves the safety and reliability of the entire project.

[0058] The feedback closed-loop optimization module enables continuous optimization of the construction process. It analyzes safety warning instructions through an incremental learning algorithm, obtains feedback optimization parameters, and uses an adaptive adjustment algorithm to make real-time corrections to task scheduling results, generating a closed-loop optimization scheduling strategy. This enables the construction process to be continuously adjusted and optimized based on actual conditions, improving adaptability and efficiency. As construction progresses, the system continuously learns and improves, gradually enhancing construction management and further improving construction quality and efficiency.

[0059] The analysis software integration module and geographic information service module further enhance the convenience and visualization of construction management. The analysis software integration module deploys multiple analysis software programs on the cloud platform through cloud integration or offline call-based access. It also enables cross-software data interoperability through a unified data protocol, facilitating user operations and data sharing. The geographic information service module, based on WebGIS technology, integrates wind farm geographic information, real-time ship AIS data, and geological exploration data. It utilizes tiled map services and dynamic layer overlay technology to visualize multi-dimensional spatial data. This enables construction personnel to intuitively understand the construction environment and progress, facilitating overall construction planning and management, and improving construction efficiency and decision-making accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a working principle diagram of the cloud platform system for offshore wind power construction according to the present invention;

[0061] Figure 2 This is the workflow diagram of the multidimensional data fusion module;

[0062] Figure 3 This is the workflow diagram of the construction path optimization module;

[0063] Figure 4 This is the workflow diagram of the security situation awareness module. DETAILED DESCRIPTION

[0064] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0065] See also Figure 1-4 The present invention provides a technical solution: a cloud platform system for offshore wind power construction, which includes a multi-dimensional data fusion module, a construction path optimization module, a resource dynamic scheduling module, a safety situation awareness module, and a feedback closed-loop optimization module. In the offshore wind power construction scenario, the multi-dimensional data fusion module collects ship data, pile hammer data, and environmental monitoring data in real time through a heterogeneous data acquisition interface. For example, during the construction of a certain offshore wind farm, the navigation speed and position information of the construction ship, the hitting frequency and force data of the pile hammer, and environmental monitoring data such as wind speed and wave height are collected to obtain the original construction data set. Then, a data cleaning algorithm is used to eliminate data noise to obtain cleaned construction data. Then, a spatiotemporal correlation fusion algorithm is used to perform multi-dimensional data fusion to generate fused construction data.

[0066] The construction path optimization module uses a dynamic path planning algorithm to model the path based on the ship's navigation trajectory from the integrated construction data, generating an initial construction path. Incorporating the wind farm's offshore topology, a constrained optimization algorithm is used to detect and correct conflicts within the initial construction path. For example, conflicts with existing submarine cables within the wind farm are avoided, resulting in an optimized construction path.

[0067] The dynamic resource scheduling module, based on a construction task priority model, decomposes the optimized construction path into a set of construction subtasks. For example, a specific phase of offshore wind turbine construction involves multiple subtasks, such as piling tasks at different machine positions. Using a dynamic priority scheduling algorithm, resources are matched and scheduled across these subtasks, resulting in a task scheduling result.

[0068] The security situation awareness module uses real-time streaming data processing technology to extract abnormal features from equipment status data in task scheduling results. For example, it monitors the operating temperature, pressure, and other status data of construction equipment in real time to extract abnormal feature sets. It then calculates the safety risk probability based on these abnormal feature sets and generates safety warning instructions.

[0069] The feedback closed-loop optimization module uses an incremental learning algorithm to analyze the feedback data from safety warning instructions to obtain feedback optimization parameters. Combined with the feedback optimization parameters, an adaptive adjustment algorithm is used to make real-time corrections to the task scheduling results, generating a closed-loop optimization scheduling strategy.

[0070] The present invention will be further described below in conjunction with Examples 1 to 7:

[0071] Example 1:

[0072] This embodiment corresponds to a specific implementation of a multidimensional data fusion module. In actual offshore wind power construction, heterogeneous data acquisition interfaces connect to various data sources, such as sensors on ships, monitoring systems for pile hammer equipment, and environmental monitoring stations. For example, a construction vessel's sensors collect real-time data on the vessel's heading, speed, draft, and other data in various formats. The heterogeneous data acquisition interface first converts the collected ship data, pile hammer data, and environmental monitoring data into a standardized format, transforming it into unified, structured construction data for easier subsequent processing.

[0073] After obtaining structured construction data, some missing values ​​and outliers may exist. For example, at a certain moment, due to a sensor failure, the ship's speed data may show an abnormally large value. In this case, the missing and outlier values ​​in the structured construction data are interpolated and filtered. Methods such as mean interpolation and median interpolation are used to fill in the missing values, and statistical methods are used to identify and filter outliers, thereby obtaining clean construction data.

[0074] Next, the spatiotemporal labels of the cleaning construction data are correlated and matched using the spatiotemporal correlation fusion algorithm. The function formula of the spatiotemporal correlation fusion algorithm is:

[0075] Where F is the fusion construction data, For the Time series data, such as the frequency of a pile hammer hitting at a certain moment; The time weight coefficient is set according to the importance of the data in the time dimension, and its value range is usually arrive between; It is the time decay factor, which is used to measure the influence of time on data fusion. The larger its value is, the faster the influence of time on data decays. is the time interval, that is, the time difference between two adjacent time series data; For the Spatial data, such as the geographic location of a construction vessel at a certain moment; is the spatial weight coefficient, which reflects the importance of spatial data in fusion; is a spatial distance correlation function, which is used to measure the degree of correlation between data at different spatial locations. Indicates the distance between different spatial data.

[0076] According to this formula, a spatiotemporal correlation matrix is ​​first calculated. Based on this matrix, multidimensional data is weighted and fused to produce fused construction data. For example, within a specific time period, spatial data on ship positions and time series data on pile hammer impact frequencies at different times are weighted and fused according to the above formula. Ultimately, fused construction data that comprehensively reflects the construction situation is generated, providing accurate data support for subsequent modules such as construction path optimization.

[0077] Example 2:

[0078] This embodiment is directed to a specific implementation method of the construction path optimization module. In offshore wind power construction, construction vessels need to navigate in a complex sea environment and reach various construction positions to perform operations. The dynamic path planning algorithm performs segmented modeling on the ship's navigation trajectory. Taking a certain construction stage as an example, it is assumed that the ship needs to navigate from the starting point to multiple different machine positions for piling operations. The dynamic path planning algorithm divides the ship's navigation trajectory into multiple small segments based on the ship's current position, the target machine position, and the real-time sea conditions (such as water speed, wind direction, etc.), and generates an initial path node set. Each node contains relevant information about the ship at that location, such as position coordinates, navigation direction, etc.

[0079] The constraint optimization algorithm is used in combination with the coordinates of the prohibited sea area to perform collision detection on the initial path node set. The function formula of the constraint optimization algorithm is as follows:

[0080] Where, To optimize the construction path; is the cost function of the kth constraint condition, for example, for the constraint of the prohibited area in the sea area, when the path passes through the prohibited area, The value of will be large to avoid path conflicts; is the constraint weight coefficient, which is used to adjust the importance of different constraint conditions; is the path smoothing coefficient, which is used to balance the degree of path optimization and path smoothness. The larger the value, the smoother the path, but it may deviate from the optimal path; is the initial path node set; P is the variable of the path node set, which represents the path candidate to be optimized.

[0081] Based on this formula, the initial path node set is collided with, and nodes that conflict with restricted areas in the sea area are removed. For example, if an initial path node falls within a restricted area for already laid submarine cables, the algorithm removes that node from the initial path node set. Path smoothing is then performed on the remaining path node set, using methods such as spline interpolation to generate an optimized construction path. This optimized construction path not only meets the construction task requirements but also avoids conflicts with restricted areas, while ensuring a smooth path, facilitating the safe and efficient navigation of construction vessels.

[0082] Example 3:

[0083] This embodiment describes in detail the specific implementation method of the resource dynamic scheduling module. During the offshore wind power construction process, there are multiple construction tasks, such as piling at different machine positions, tower installation, and other tasks, which have different urgency and resource requirements. The urgency and resource consumption of the construction subtask set are evaluated and processed through the task priority model. For example, for the piling task that is about to miss the optimal construction window, its urgency is higher; and for some resource-intensive tasks, such as tower installation tasks that require a large amount of concrete and steel, the resource consumption is relatively large. Based on these factors, a task priority sequence is generated.

[0084] The dynamic priority scheduling algorithm is combined with the real-time position data of the ship to allocate resources to the task priority sequence. The function formula of the dynamic priority scheduling algorithm is as follows:

[0085] Where, For the Dynamic priority of tasks; The value range can be set according to the importance and time urgency of the task, for example, arrive The larger the value, the more urgent it is; The deadline for the task; The real-time available resources, such as the amount of concrete, steel, and other resources currently remaining at the construction site; and It is an adjustment parameter used to adjust the impact of task urgency and real-time available resources on dynamic priority, which is usually set through a large number of experiments and practical experience; t is the current system time.

[0086] Assume that at a certain moment, there are two construction subtasks, and the urgency of task A is , deadline There are still t days left from the current system time Hours, real-time available resources Can satisfy Task A ; The urgency of Task B , deadline There are still t days left from the current system time Hours, real-time available resources Can satisfy Task B . Adjustment parameters , According to the formula, the dynamic priority of task A is , the dynamic priority of task B Therefore, resources are allocated to Task A first, such as deploying construction ships and construction equipment, so as to obtain reasonable task scheduling results and improve construction efficiency.

[0087] Example 4:

[0088] This example illustrates a specific implementation of the security situation awareness module. In offshore wind power construction, the operating status of construction equipment, such as piling equipment and ship propulsion systems, is crucial. Real-time stream data processing technology is used to perform sliding window sampling on equipment status data. For example, to sample the oil temperature data of piling equipment, a sliding window size of 10 sampling points is set. Oil temperature data is collected at regular intervals (e.g., 1 minute). After each new data acquisition, the oldest data is removed from the window to generate time-series data on the equipment status.

[0089] Convolutional neural networks are used to extract abnormal features from equipment status time series data. Convolutional neural networks have powerful feature extraction capabilities. Through multiple convolutional and pooling layers, they can automatically learn features from equipment status data. The equipment status time series data is input into the convolutional neural network, and after a series of convolution and pooling operations, a set of equipment abnormality features is generated. For example, if the oil temperature of a pile driving device continues to rise and exceeds the normal range for a period of time, the convolutional neural network can extract this abnormal feature and include it in the equipment abnormality feature set.

[0090] Based on the Markov chain model, the risk probability evolution of the equipment anomaly feature set is calculated to generate safety warning instructions. The Markov chain model assumes that the system's state at a future moment is only related to the current state, and not to the past state. Based on the equipment anomaly feature set and the transition probability matrix of the Markov chain model, the probability of safety risks such as equipment failure is calculated. For example, after the abnormal feature of an abnormal increase in the oil temperature of the piling equipment is extracted, the Markov chain model calculates the probability of the equipment failing in the future based on previous equipment failure data and the current oil temperature anomaly. When this probability exceeds a preset threshold, a safety warning instruction is generated, such as sending an alarm message to construction personnel, notifying them to inspect and maintain the equipment to prevent the impact of equipment failure on construction.

[0091] Example 5:

[0092] This embodiment details the portion of the feedback loop optimization module that uses an incremental learning algorithm to analyze and process feedback data from safety warning instructions. When a safety warning is issued, such as an alert indicating excessive oil temperature in piling equipment, the incremental learning algorithm extracts features from the warning, including the type of equipment anomaly (e.g., excessive oil temperature) and its frequency. Assuming that five excessive oil temperature warnings are issued for piling equipment over a period of time, the incremental learning algorithm extracts this information and generates a feedback feature vector containing information such as the type of equipment anomaly and its frequency.

[0093] The feedback eigenvector is processed based on the gradient descent method to calculate the parameter update direction and obtain feedback optimization parameters. The gradient descent method is a commonly used optimization algorithm that minimizes the objective function by calculating the gradient of the objective function and continuously adjusting the parameters. In this embodiment, the feedback eigenvector is used as input, and the gradient descent method is used to calculate the parameter update direction that can optimize the task scheduling results, thereby obtaining feedback optimization parameters. For example, if the pile driving equipment oil temperature abnormality is found to be excessively high frequently, the gradient descent method will calculate the direction for adjusting parameters related to the pile driving equipment (such as equipment usage time, maintenance plan, etc.) in the task scheduling results based on this feedback eigenvector, and obtain the corresponding feedback optimization parameters.

[0094] Compare and verify the feedback optimization parameters with the preset weight threshold interval to screen the effective optimization parameter set. The preset weight threshold interval is set based on factors such as construction experience and equipment performance. For example, for the feedback optimization parameter that adjusts the use time of the piling equipment, the preset weight threshold interval is If the calculated feedback optimization parameter is within this interval, the parameter is considered valid and included in the valid optimization parameter set; if it is not within this interval, it is adjusted or discarded to ensure that the subsequent correction of the task scheduling result is reasonable and effective.

[0095] Example 6:

[0096] This embodiment continues to introduce the part of the feedback closed-loop optimization module that uses the adaptive adjustment algorithm to correct the task scheduling results. After obtaining the effective optimization parameter set, the adaptive adjustment algorithm is used to calculate the multi-dimensional correction coefficients of the task priority weights, resource allocation weights and ship path weights in the task scheduling results. For example, based on the feedback optimization parameters, if it is found that the urgency of a construction task needs to be re-evaluated, the adaptive adjustment algorithm will calculate the correction coefficient of the task priority weight according to the relevant rules. Assuming that the original priority weight of task A is 0.6, after calculation by the adaptive adjustment algorithm, the correction coefficient is 1.2, then the adjusted priority weight of task A is .

[0097] Combined with the feedback optimization parameter set, the correction coefficients are dynamically adjusted to generate an updated scheduling weight matrix. The correction coefficients of each task's priority weight, resource allocation weight, and ship path weight are integrated to form a scheduling weight matrix. For example, in a certain construction phase, there are 3 construction tasks, each of which corresponds to a different priority weight, resource allocation weight, and ship path weight. By dynamically adjusting the correction coefficients of these weights, a scheduling weight matrix is ​​generated. The scheduling weight matrix.

[0098] A weighted fusion method is used to match the scheduling weight matrix with real-time construction task data to generate a closed-loop optimization scheduling strategy. This method applies weighted calculations to real-time construction task data based on the weights in the scheduling weight matrix. For example, the specific amount of resources allocated for Task A is calculated based on the resource allocation weights for Task A in the scheduling weight matrix and real-time resource data from the construction site. This method allows for real-time corrections to task scheduling results, generating a closed-loop optimization scheduling strategy that continuously optimizes the construction process and improves efficiency and safety.

[0099] Example 7:

[0100] This example describes the specific implementation of the analysis software integration module and geographic information service module in a cloud platform system. The analysis software integration module is used to deploy pile insertion and extraction calculation and analysis software, pile driveability analysis software, floating vessel operability prediction software, pile foundation penetration analysis software, and wind farm resource scheduling software on the cloud platform via cloud integration or offline call access.

[0101] Cloud integration involves deploying the software on a virtual machine on the cloud platform, allowing users to directly operate the software through single sign-on. For example, a construction technician at the construction site logs into the cloud platform using a tablet computer, accesses the plug-and-pull pile calculation and analysis software interface, and enters relevant construction parameters such as pile type, size, and geological conditions. The software then runs the calculation on the cloud platform virtual machine. After approval by the technical center, the results are automatically uploaded to the document database. The technical center can then review the results to determine whether they meet construction requirements and specifications.

[0102] The offline call method involves users submitting data forms to the cloud platform. The technical center then operates local computing nodes to generate reports and transmits them back to the cloud platform. For example, before a construction worker begins piling, they fill out a data form with data such as the piling location and hammer parameters and submit it to the cloud platform. Upon receiving the data form, the technical center runs the pile driveability analysis software on a local computing node, generates an analysis report based on the submitted data, and then transmits the report back to the cloud platform for the construction worker's reference.

[0103] Cross-software data interoperability is achieved through unified data identifiers, consisting of the wind farm code OW followed by a four-digit number, the machine site code TK followed by a five-digit number, and the vessel's unique ID. For example, in a wind farm with the wind farm code OW0001, construction at the machine site code TK00001 involves operations on vessel 001. These unified data identifiers enable accurate identification and retrieval of relevant data when exchanging data between different software, enabling cross-software data interoperability and improving the efficiency and accuracy of construction data utilization.

[0104] The geographic information service module integrates wind farm geographic information, real-time ship AIS data, and geological exploration data based on WebGIS technology. For example, it integrates geographic information such as the wind farm's geographical location and seabed topography, AIS data such as the real-time location and navigation trajectory of construction vessels, and data such as stratigraphic structure and soil characteristics obtained through geological exploration. Multi-dimensional spatial data visualization is achieved using tiled map services and dynamic layer overlay technology. Tile-based map services divide map data into multiple small tiles, increasing map loading speed. Dynamic layer overlay technology can overlay different data layers (such as ship location layers and geological layers) on the map according to construction needs, allowing construction personnel to intuitively view the geographic information, ship locations, and geological conditions of the construction area, providing strong support for construction decision-making.

[0105] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0106] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A cloud platform system for offshore wind power construction, characterized in that: include: The multidimensional data fusion module is used to collect and process ship data, pile hammer data, and environmental monitoring data from offshore wind power construction in real time through heterogeneous data acquisition interfaces to obtain the original construction data set; it uses a data cleaning algorithm to remove noise from the original construction data set to obtain cleaned construction data; and it uses a spatiotemporal correlation fusion algorithm to perform multidimensional data fusion processing on the cleaned construction data to obtain fused construction data; The construction path optimization module is used to model the ship's navigation trajectory in the integrated construction data using a dynamic path planning algorithm to obtain an initial construction path. It also uses a constrained optimization algorithm combined with the wind farm's sea area topology to perform conflict detection and correction on the initial construction path to obtain an optimized construction path. The resource dynamic scheduling module is used to decompose the optimized construction path based on the construction task priority model and generate a set of construction subtasks. It also uses the dynamic priority scheduling algorithm to match and schedule resources for the construction subtasks and obtain the task scheduling results. The security situation awareness module is used to extract abnormal features from the device status data in the task scheduling results through real-time stream data processing technology to obtain a device abnormal feature set; it also calculates the security risk probability of the device abnormal feature set and generates a security warning instruction; The feedback closed-loop optimization module is used to analyze and process the feedback data of safety warning instructions through an incremental learning algorithm to obtain feedback optimization parameters. It also uses an adaptive adjustment algorithm combined with the feedback optimization parameters to perform real-time correction processing on the task scheduling results to generate a closed-loop optimization scheduling strategy. The construction path optimization module includes: The ship's navigation trajectory is segmented and modeled using a dynamic path planning algorithm to generate an initial path node set. The constraint optimization algorithm is used in combination with the coordinates of the prohibited sea area to perform collision detection on the initial path node set and eliminate conflicting nodes. The function formula of the constraint optimization algorithm is as follows: ， Where, To optimize the construction path, For the The constraint cost function, is the constraint weight coefficient, is the path smoothing coefficient, is the initial path node set, is a variable of the path node set, representing the path candidate to be optimized, is the number of constraints; Path smoothing is performed based on the remaining path node set to generate an optimized construction path.

2. The cloud platform system for offshore wind power construction according to claim 1, characterized in that: The multidimensional data fusion module includes: Through the heterogeneous data acquisition interface, ship data, pile hammer data and environmental monitoring data are converted into standardized formats to obtain structured construction data; Interpolate and filter missing values ​​and outliers in structured construction data to obtain clean construction data; The spatiotemporal correlation fusion algorithm is used to perform correlation matching on the spatiotemporal labels of the cleaning construction data to generate a spatiotemporal correlation matrix. The multi-dimensional data is weightedly fused based on the spatiotemporal correlation matrix to obtain fused construction data.

3. The cloud platform system for offshore wind power construction according to claim 2, characterized in that: The function formula of the spatiotemporal correlation fusion algorithm is as follows: ， Where, To integrate construction data, For the Time series data, is the time weight coefficient, is the time decay factor, is the time interval, For the Spatial data, is the spatial weight coefficient, is the spatial distance correlation function, is the number of time series data, is the number of spatial data.

4. The cloud platform system for offshore wind power construction according to claim 1, characterized in that: The resource dynamic scheduling module includes: The urgency and resource consumption of the construction subtask set are evaluated and processed through the task priority model to generate a task priority sequence; The dynamic priority scheduling algorithm is combined with the real-time position data of the ship to perform resource allocation processing on the task priority sequence to obtain the task scheduling result; the function formula of the dynamic priority scheduling algorithm is as follows: ， Where, For the Dynamic priority of tasks, For the task urgency, The deadline for the task, is the real-time available resources, and To adjust the parameters, The current system time.

5. The cloud platform system for offshore wind power construction according to claim 1, characterized in that: The security situation awareness module includes: The device status data is sampled and processed by sliding windows through real-time stream data processing technology to obtain device status time series data; Use convolutional neural networks to extract abnormal features from device status time series data and generate a device abnormal feature set; Based on the Markov chain model, the risk probability evolution of the equipment abnormality feature set is calculated to generate safety warning instructions.

6. The cloud platform system for offshore wind power construction according to claim 1, characterized in that: The feedback closed-loop optimization module includes: The incremental learning algorithm is used to extract the features of the equipment abnormality type and occurrence frequency in the safety warning instructions and generate feedback feature vectors; Based on the gradient descent method, the parameter update direction of the feedback feature vector is calculated and processed to obtain the feedback optimization parameters; The feedback optimization parameters are compared and verified with the preset weight threshold interval to screen the effective optimization parameter set.

7. The cloud platform system for offshore wind power construction according to claim 6, characterized in that: The feedback closed-loop optimization module also includes: An adaptive adjustment algorithm is used to calculate multi-dimensional correction coefficients for task priority weights, resource allocation weights, and ship path weights in the task scheduling results. The correction coefficients are dynamically adjusted in combination with the feedback optimization parameter set to generate an updated scheduling weight matrix. The scheduling weight matrix is ​​matched with the real-time construction task data through the weighted fusion method to generate a closed-loop optimization scheduling strategy.

8. The cloud platform system for offshore wind power construction according to claim 1, characterized in that: Also includes: An analysis software integration module is used to deploy pile insertion and extraction calculation and analysis software, pile driveability analysis software, floating vessel operability prediction software, pile foundation penetration analysis software, and wind farm resource scheduling software on a cloud platform through cloud integration or offline call, and to achieve cross-software data interoperability through a unified data protocol; In the analysis software integration module: The cloud integration method is to deploy the software on a cloud platform virtual machine. Users can directly operate the software through single sign-on. The calculation results are automatically uploaded to the document database after approval by the technical center. The offline call method is that the user submits a data form to the cloud platform, and the technical center operates the local computing node to generate a report and transmits it back to the cloud platform; Cross-software data interoperability is achieved through a unified data identifier, which includes the wind farm code OW+4 digits, the machine position code TK+5 digits and the ship's unique ID.

9. The cloud platform system for offshore wind power construction according to claim 1, characterized in that: Also includes: The geographic information service module integrates wind farm geographic information, real-time ship AIS data and geological exploration data based on WebGIS technology, and uses tiled map services and dynamic layer overlay technology to achieve multi-dimensional spatial data visualization.

Images