Mooring failure prediction method, device and equipment based on floating wind turbine

Abstract

The invention provides a mooring failure prediction method, device and equipment based on a floating type wind turbine, and belongs to the technical field of floating type wind turbines.The method comprises the steps that platform data of the floating type wind turbine and environment data of the environment where the floating type wind turbine is located are obtained; analyzing the platform data, and constructing a platform numerical simulation model of the floating wind turbine; splitting the environmental data, and extracting key environmental parameters of each environmental parameter type; on the basis of the key environment parameters, motion change information and mooring force change information of the floating wind turbine are predicted through a platform numerical simulation model and a motion mooring prediction model; according to the method, the mooring failure evaluation strategy is used for carrying out anomaly evaluation on the motion change information and the mooring force change information, a mooring failure evaluation result of the floating type wind turbine is generated, and the accuracy of motion response prediction of the floating type wind turbine is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of floating wind turbine technology, and in particular to a method, apparatus and equipment for predicting mooring failure based on floating wind turbines. Background Technology

[0002] For currently large-scale offshore stationary wind power systems, operation and maintenance costs often account for more than one-third of the total investment. For floating wind turbines, in addition to routine maintenance of the turbines, close attention must be paid to the platform's operational status to ensure safety. One of the most significant factors contributing to the overall failure of floating platforms is mooring failure. Accidents such as broken anchor chains can cause the platform to drift, changing its operational area and motion patterns. Understanding and predicting the extreme characteristics of platform mooring forces allows for anticipating failure risks before extreme weather events, enabling better preparedness. Therefore, predicting the extreme values ​​of floating wind turbine motion and mooring forces is crucial for improving operational efficiency and ensuring structural safety.

[0003] Currently, short-term motion response prediction for floating wind turbines mainly adopts a data-driven, univariate input prediction architecture based on historical motion data, and is mostly a single-step prediction mode. Such models are highly dependent on the platform's historical motion data, and their applicability is limited when facing scenarios that require long-term motion characteristic evolution prediction, such as maintenance window forecasting, resulting in low accuracy in predicting the motion response of floating wind turbines. Summary of the Invention

[0004] This invention provides a method, apparatus, and equipment for predicting mooring failure based on a floating wind turbine, in order to address the shortcomings of poor accuracy in mooring failure prediction in existing technologies.

[0005] In a first aspect, the present invention provides a mooring failure prediction method based on a floating wind turbine, comprising:

[0006] Acquire platform data of the floating wind turbine and environmental data of the environment in which the floating wind turbine is located;

[0007] The platform data was analyzed to construct a platform numerical simulation model of the floating wind turbine;

[0008] The environmental data is split, and key environmental parameters of each environmental parameter type are extracted;

[0009] Based on the key environmental parameters, the motion change information and mooring force change information of the floating wind turbine are predicted through the platform numerical simulation model and the motion mooring prediction model.

[0010] Using a mooring failure assessment strategy, anomaly assessments are performed on the motion change information and the mooring force change information to generate the mooring failure assessment results for the floating wind turbine.

[0011] According to the present invention, a method for predicting mooring failure based on a floating wind turbine, wherein the step of parsing the platform data and constructing a platform numerical simulation model of the floating wind turbine includes:

[0012] The platform data is broken down into structural data and structural performance parameters of each platform structure;

[0013] Based on the structural data and structural performance parameters, a numerical simulation model of the substructure of each platform structure is constructed.

[0014] The numerical simulation models of all the substructures are spliced ​​together to obtain the platform numerical simulation model of the floating wind turbine.

[0015] According to the present invention, a mooring failure prediction method based on a floating wind turbine is provided, wherein the environmental data is split and key environmental parameters of each environmental parameter type are extracted, including:

[0016] The environmental data is split into environmental wind data and wave current data. Based on the environmental wind data and the wave current data, the environmental wind simulation data and wave current simulation data of the environment in which the floating wind turbine is located are simulated by a time-domain calculation model.

[0017] Using a preset data extraction strategy, target environmental wind simulation data and target wave current simulation data are extracted from the environmental wind simulation data and the wave current simulation data, respectively.

[0018] Based on the target environment wind simulation data and the target wave current simulation data, key environmental parameters for each environmental parameter type are extracted.

[0019] According to the present invention, a mooring failure prediction method based on a floating wind turbine is provided, wherein the motion mooring prediction model includes a motion prediction model;

[0020] Based on the aforementioned key environmental parameters, the motion change information of the floating wind turbine is predicted using the platform's numerical simulation model and motion mooring prediction model, including:

[0021] Among the key environmental parameters, key environmental parameters of each first environmental parameter type required by the motion prediction model are selected.

[0022] Based on the key environmental parameters of each of the first environmental parameter types, a first simulation dataset for the motion prediction model is constructed.

[0023] Based on the first simulation dataset, the motion change information of the floating wind turbine is predicted through the motion prediction model and the platform numerical simulation model.

[0024] According to the present invention, a mooring failure prediction method based on a floating wind turbine is provided, wherein the method predicts the motion change information of the floating wind turbine based on the first simulation dataset, through the motion prediction model and the platform numerical simulation model, including:

[0025] Based on the first simulation dataset, the motion data distribution information of each motion data type of the floating wind turbine is predicted by the motion prediction model and the platform numerical simulation model.

[0026] The distribution information of each motion data is subjected to distribution fitting processing to obtain the target motion data distribution information corresponding to each motion data type, and the target motion data distribution information is used as the motion change information of the floating wind turbine.

[0027] According to the present invention, a mooring failure prediction method based on a floating wind turbine is provided, wherein the moving mooring prediction model includes a mooring tension prediction model;

[0028] Based on the aforementioned key environmental parameters, the prediction of mooring force changes of the floating wind turbine using the platform's numerical simulation model and motion mooring prediction model includes:

[0029] Among the key environmental parameters of each of the environmental parameter types, the key environmental parameters of each second environmental parameter type required by the mooring tension prediction model are selected.

[0030] Based on the key environmental parameters of each of the second environmental parameter types, a second simulation dataset for the mooring tension prediction model is constructed.

[0031] Based on the second simulation dataset, the mooring force variation information of the floating wind turbine is predicted using the mooring tension prediction model and the platform numerical simulation model.

[0032] According to the present invention, a mooring failure prediction method based on a floating wind turbine, based on the second simulation dataset, predicts the mooring force variation information of the floating wind turbine through the mooring tension prediction model and the platform numerical simulation model, including:

[0033] Based on the second simulation dataset, the tension distribution information of each anchor chain of the floating wind turbine is predicted by the mooring tension prediction model and the platform numerical simulation model, and the model error data of the mooring tension prediction model is determined.

[0034] Based on the tension prediction distribution information of each anchor chain and the model error data, the target tension change distribution information of each anchor chain is identified, and the target tension change distribution information is used as the mooring force change information of the floating wind turbine.

[0035] According to the present invention, a mooring failure prediction method based on a floating wind turbine is provided, wherein the method utilizes a mooring failure assessment strategy to perform anomaly assessment on the motion change information and the mooring force change information, and generates a mooring failure assessment result for the floating wind turbine, including:

[0036] Based on the target motion data distribution information corresponding to each of the motion data types, the motion anomaly assessment results of the floating wind turbine are identified through a motion anomaly assessment strategy.

[0037] Based on the target tension change distribution information of each anchor chain, the mooring force distribution information of each anchor chain is identified, and the mooring force anomaly assessment result of the floating wind turbine is identified through the mooring force anomaly assessment strategy.

[0038] Based on the motion anomaly assessment results and the mooring force anomaly assessment results, the mooring failure assessment results of the floating wind turbine are generated.

[0039] Secondly, the present invention provides a mooring failure prediction device based on a floating wind turbine, comprising:

[0040] The acquisition module is used to acquire platform data of the floating wind turbine and environmental data of the environment in which the floating wind turbine is located;

[0041] A construction module is used to parse the platform data and construct a platform numerical simulation model of the floating wind turbine;

[0042] The extraction module is used to split the environmental data and extract key environmental parameters for each type of environmental parameter.

[0043] The prediction module is used to predict the motion change information and mooring force change information of the floating wind turbine based on the key environmental parameters, through the platform numerical simulation model and the motion mooring prediction model.

[0044] The evaluation module is used to perform anomaly evaluation on the motion change information and the mooring force change information using a mooring failure evaluation strategy, and generate the mooring failure evaluation result of the floating wind turbine.

[0045] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the mooring failure prediction method based on a floating wind turbine as described above.

[0046] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the mooring failure prediction method based on a floating wind turbine as described above.

[0047] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the mooring failure prediction method based on a floating wind turbine as described above.

[0048] This invention provides a method, apparatus, and equipment for predicting mooring failure based on a floating wind turbine, comprising: acquiring platform data of the floating wind turbine and environmental data of the environment in which the floating wind turbine is located; analyzing the platform data to construct a platform numerical simulation model of the floating wind turbine; decomposing the environmental data and extracting key environmental parameters of each environmental parameter type; based on each key environmental parameter, predicting the motion change information and mooring force change information of the floating wind turbine through the platform numerical simulation model and the motion mooring prediction model; and using a mooring failure assessment strategy to perform anomaly assessment on the motion change information and mooring force change information, generating a mooring failure assessment result for the floating wind turbine. By constructing a platform numerical simulation model, the performance and function of floating wind turbines can be effectively simulated, thereby improving the comprehensiveness and accuracy of the analysis of floating wind turbine performance data. The motion mooring prediction model is combined with the platform data simulation model for simulation prediction, ensuring that the prediction results meet the prediction accuracy requirements of engineering projects. Moreover, the model only relies on environmental parameter inputs and does not require real-time motion data from the platform, which can reduce the dependence on the monitoring system and provide a basis for application scenarios such as decision-making on safe operation windows of floating wind power. This improves the applicability of predictions when facing scenarios that require long-term motion characteristic evolution prediction, such as operation and maintenance window forecasting, thereby comprehensively improving the accuracy of motion response prediction for floating wind turbines. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0050] Figure 1This is a flowchart illustrating the mooring failure prediction method based on a floating wind turbine provided in this embodiment.

[0051] Figure 2 This is a three-dimensional structural schematic diagram of the floating wind power platform provided in this embodiment;

[0052] Figure 3 This is a schematic diagram of the mooring failure prediction device based on a floating wind turbine provided in this embodiment;

[0053] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this embodiment. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0055] Figure 1 This is a flowchart illustrating the mooring failure prediction method based on a floating wind turbine provided in this embodiment.

[0056] like Figure 1 As shown, this embodiment provides a mooring failure prediction method based on a floating wind turbine, which can be applied to the application environment of a mooring failure prediction system based on a floating wind turbine. This system can be applied to a terminal, a server, or a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, mid-range computers, etc. The method mainly includes the following steps:

[0057] 101. Obtain platform data of the floating wind turbine and environmental data of the environment in which the floating wind turbine is located.

[0058] Specifically, in this embodiment, in response to the information upload operation by the staff, the terminal obtains the basic platform parameters of the floating wind turbine platform, as well as the platform structure data, such as... Figure 2 The diagram shown is a three-dimensional structural schematic of a floating wind turbine. The basic parameters of this platform include, but are not limited to, the parameters listed in Table 1.

[0059] Table 1 Basic Platform Parameters

[0060]

[0061]

[0062] Then, the terminal collects real-time wind and wave data of the platform's location to obtain environmental data of the floating wind turbine's environment. The wind data includes, but is not limited to, wind speed data and wind speed variation data, while the wave data includes wave direction data, flow velocity data, and angle data.

[0063] 102. Analyze the platform data and construct a platform numerical simulation model of the floating wind turbine.

[0064] The platform numerical simulation model is used to simulate the performance and operation of the floating wind turbine. After obtaining the platform data, the platform data is broken down into structural data and structural performance parameters of each platform structure; based on the structural data and structural performance parameters, substructure numerical simulation models of each platform structure are constructed; and all substructure numerical simulation models are spliced ​​together to obtain the platform numerical simulation model of the floating wind turbine.

[0065] Specifically, frequency domain calculations were first performed, establishing the platform's columns and damping chamber as plate models and using wave diffraction / radiation theory for calculations. The trusses between the columns were not involved in the frequency domain calculations or coefficient transfer; instead, they were directly established as finite elements in the time domain calculation model. To fully simulate the viscous effect of the damping chamber, a virtual annular Morrison rod model was established at its bottom, and the drag force coefficients used were calibrated through pool experiments. Since the research object is a floating wind turbine platform under operating conditions, ballast water and mooring elements need to be defined to ensure that the mass distribution conforms to actual conditions. The frequency domain calculation results include the platform's mass matrix, restoring force matrix, kinematic response RAO, force response RAO, mean wave drift force coefficient, and frequency-dependent additional mass and additional damping matrices.

[0066] To account for the stress and motion characteristics of the platform's three columns under wave action, the frequency-domain hydrodynamic coefficients of the floating platform are transferred to the time-domain computational model. The trusses between the platform columns are modeled separately as finite elements for time-domain computation, which better accounts for the nonlinear effects of Morrison's towing force. The towing force coefficient C used is... d and inertial force coefficient C m After being calibrated through tank testing, the mooring system was also modeled in detail as a finite element model with vertical loads added to balance the vertical anchor chain pretension.

[0067] In this embodiment, the platform integrates wind power generation and aquaculture functions and is equipped with a double-layer fishing net. Under the influence of waves and ocean currents, it generates complex hydrodynamic loads, which have a significant impact on the platform's motion attitude, necessitating coupled dynamic analysis in the time domain. The specific configuration of the double-layer fishing net in this embodiment is as follows: the outer net has a diameter of 5.5 mm, a mesh size of 150 × 150 mm, and is made of T70; the inner net has a diameter of 3.0 mm, a mesh size of 37.5 × 37.5 mm, and is made of Dyneema SK78. In the simulation, a nonlinear finite element method coupled with potential flow theory is used to simulate the deformation of the fishing net itself and its impact on the platform. The nodes at both ends of the fishing net unit are rigidly connected to the platform. The following hydrodynamic parameters are used to characterize the interaction between the fishing net and the fluid: inner net density 0.1536, drag coefficient 0.8; outer net density 0.073, drag coefficient 0.9. Simulation results show that the fishing net structure has a significant impact on the platform's equilibrium position under convective loads.

[0068] Because the fishing net structure contains a large number of mesh elements, performing a fully refined finite element simulation would require an extremely small computation time step (e.g., 0.0001 seconds) to ensure the convergence of the numerical calculation, resulting in extremely high computational costs. Considering that the focus of this invention is on the impact of the fishing net on the overall platform motion rather than the local deformation of the net itself, this invention proposes a simplified equivalent modeling method: simulating each complete piece of netting as a single net element. To verify the feasibility and accuracy of this simplified model, we performed a convergence analysis on the mesh density. Specifically, under a limiting condition (defined as a significant wave height H...),... s = 4.77m, peak period T p =10.65s, flow velocity U c =2.53m / s, wind speed U 10 At a speed of 30 m / s, the calculation results of the 1-unit net model were compared with those of a more refined model (the inner net is divided into 10 units, and the outer net into 30 units). The results show that the average absolute differences between the two models in the directions of sway, roll, heave, pitch, pitch, and bow are 0.016 m, 0.013 m, 0.013 m, 0.031°, 0.029°, and 0.038°, respectively. The differences between the simplified and refined models are very small across all six degrees of freedom of the platform, fully demonstrating that the "1-unit fishing net model" possesses sufficient engineering accuracy in the simulation analysis of this invention.

[0069] The floating wind power platform in this embodiment has two wind load components: one is the platform wind load, which is considered using a second-order wind transfer coefficient; the other is the wind load on the turbine and tower, which is applied directly to the connection between the tower bottom and the platform using the load time program provided by the turbine manufacturer. The 4MW turbine used in the platform has a cut-in wind speed of 3 m / s and a cut-out wind speed of 25 m / s. When the wind speed is less than 3 m / s, the turbine stops, and the load on the turbine can be ignored on the platform's movement. When the wind speed is between 3 m / s and 25 m / s, the turbine generates electricity normally. When the wind speed increases to above 25 m / s, the turbine idles to reduce the wind load and protect the system from damage. To expand the limited manufacturer data, data is expanded based on the square relationship between wind pressure and wind speed (Equation 1).

[0070]

[0071] Where ρ is air density and v is wind speed. Assuming the wind turbine always faces the wind, load data for different wind directions are generated through rotational transformation. The expanded database covers 2760 load combinations with wind speeds from 0 m / s to 46 m / s (1 m / s intervals) and wind directions from 0 to 354° (6° intervals). Each combination contains a 10-minute wind turbine load sequence with two different random seeds. Furthermore, the flow load on the platform is simulated using a second-order transfer coefficient calibrated through a water tank experiment.

[0072] Based on the above scheme, by modeling different structural parts separately, a platform numerical simulation model of the floating wind turbine is constructed, thereby improving the accuracy of the performance simulation of the floating wind turbine.

[0073] 103. Segment the environmental data and extract the key environmental parameters for each type of environmental parameter.

[0074] Environmental data is broken down into environmental wind data and wave / current data. Based on these data, a time-domain computation model is used to simulate the environmental wind and wave / current conditions of the floating wind turbine. A pre-defined data extraction strategy is employed to extract target environmental wind simulation data from the wind simulation data and target wave / current simulation data from the wave / current simulation data. Based on these target wind and wave / current simulation data, key environmental parameters for each environmental parameter type are extracted. These environmental parameter types include, but are not limited to, significant wave height, peak period, wave direction sine, wave direction cosine, current velocity, current direction sine, current direction cosine, wind speed, wind direction sine, and wind direction cosine.

[0075] Specifically, the method for extracting target simulation data is to set the terminal time-domain simulation duration to 10 minutes or longer to adapt to the temporal characteristics of the wind turbine load. Given that the platform is in a static equilibrium position under steady flow load and initial wind load at the start of the simulation, which does not match the actual dynamic response state, the calculation results of the first 5 minutes of transient response stage are discarded based on the platform's free decay characteristics. Only the steady-state response time-series data is selected, and the extreme values ​​of its time-series absolute values ​​are extracted as target features. Assuming that the results conform to the Gumbel distribution, for each environmental feature point, two different random seeds are used for independent simulation calculation, and the position parameter μ of its Gumbel distribution is fitted using the moment estimation method, as shown in Equation (2).

[0076]

[0077] In the formula, represents the average of the two calculations; s represents the standard deviation. Using a distribution fitting method can reduce the impact of randomness in subsequent deep learning.

[0078] Because floating wind turbines primarily bear the combined effects of wind, waves, and currents, this embodiment extracts the following key parameters to characterize this multidimensional environmental parameter space: wind parameters are represented by the average wind speed U at a height of 10m above sea level. 10 And the mean wind direction is represented, and the wave parameter is expressed as the meaningful wave height H. s Peak period T p And the average wave direction characterization, ocean current parameters in terms of surface current velocity U c In addition, the average flow direction is characterized. In the time-domain computational model, the Davenport wind speed spectrum and the JONSWAP wave spectrum are used to simulate and characterize the wind and wave environment, while the ocean current is simplified to a steady current.

[0079] The environmental space defined by the above seven dimensions of parameters covers all potential environmental combinations in the platform's service area, but the actual probability of occurrence varies at different points in the space. To optimize computational efficiency, the boundaries of the environmental space need to be compressed as much as possible. This embodiment sets reasonable parameter boundaries for the environmental space based on statistical data of the platform's service area. Through analysis of the significant wave height H... s Wind speed U 10 With the spectral peak period T p The correlation between variables, as well as the correlation between wind direction and wave direction, effectively reduces the dimensionality of independent variables in the environmental space, significantly narrowing its computational domain. Then, to ensure the generalization performance and stability of the deep learning model throughout the entire environmental space, a random sampling strategy within the environmental parameter space is adopted.

[0080] This embodiment focuses on the extreme value prediction of the six-degree-of-freedom motion response of a floating wind turbine platform and the tension of its mooring system under specific environmental conditions. A time-domain numerical computation model can map environmental characteristic parameters to the platform's motion response and mooring load characteristics. By accumulating numerical computation results, a deep learning model is trained to replace the time-consuming time-domain numerical simulation process. To ensure the generalization performance and stability of the deep learning model across the entire environmental space, a random sampling strategy within the environmental parameter space is adopted. Compared to the commonly used method of manually selecting environmental combinations, the random sampling strategy has significant scalability advantages, easily expanding the training dataset without redesigning environmental combinations. Furthermore, when the number of sampling points is sufficient, a finite sample can effectively approximate the statistical characteristics of the complete environmental space.

[0081] 104. Based on key environmental parameters, the motion change information and mooring force change information of the floating wind turbine are predicted through the platform numerical simulation model and the motion mooring prediction model.

[0082] The motion mooring prediction model includes a motion prediction model and a mooring tension prediction model. Motion change information includes the distribution information of target motion data corresponding to each motion data type, which includes, but is not limited to, sway, roll, pitch, and other data types. Mooring force change includes the distribution information of tension changes in each anchor chain of the floating wind turbine.

[0083] Specifically, among the key environmental parameters, key environmental parameters of each first environmental parameter type required by the motion prediction model are selected; based on the key environmental parameters of each first environmental parameter type, a first simulation dataset for the motion prediction model is constructed; based on the first simulation dataset, the motion change information of the floating wind turbine is predicted through the motion prediction model and the platform numerical simulation model. Among these, each first environmental parameter type includes, but is not limited to, significant wave height H. s Peak period T p Wave direction sine value sin(θ), wave direction cosine value cos(θ), flow velocity U c Flow direction sine value sin(α), flow direction cosine value cos(α), wind speed U 10 , wind direction sine value sin(ψ), wind direction cosine value cos(ψ).

[0084] Then, among the key environmental parameters of each environmental parameter type, key environmental parameters of each second environmental parameter type required by the mooring tension prediction model are selected. Based on the key environmental parameters of each second environmental parameter type, a second simulation dataset for the mooring tension prediction model is constructed. Based on the second simulation dataset, the mooring force change information of the floating wind turbine is predicted through the mooring tension prediction model and the platform numerical simulation model. The second environmental parameter type is the anchor chain parameter type of each anchor chain.

[0085] The simulation datasets are constructed by simulating the platform values ​​corresponding to the key environmental parameters of the above environmental parameter types through the simulation process of the time-domain numerical simulation model.

[0086] The motion prediction model and mooring tension prediction model constructed in this embodiment predict the motion change information and mooring force change information of the floating wind turbine, respectively, which improves the accuracy of the prediction and enhances the applicability of the prediction when facing scenarios that require long-term motion characteristic evolution prediction, such as maintenance window forecasting.

[0087] 105. Using a mooring failure assessment strategy, perform anomaly assessment on motion change information and mooring force change information to generate mooring failure assessment results for floating wind turbines.

[0088] Based on the target motion data distribution information corresponding to each type of motion data, an anomaly assessment strategy is used to identify the motion anomaly assessment results of the floating wind turbine. Based on the target tension change distribution information of each anchor chain, the mooring force distribution information of each anchor chain is identified, and a mooring force anomaly assessment strategy is used to identify the mooring force anomaly assessment results of the floating wind turbine. Based on the motion anomaly assessment results and the mooring force anomaly assessment results, a mooring failure assessment result for the floating wind turbine is generated. The mooring failure assessment result is used to characterize the motion data anomaly assessment results and mooring tension anomaly information of the floating wind turbine.

[0089] Specifically, firstly, motion anomaly assessment information corresponding to the motion data distribution feature ranges of different motion data types is preset. Then, a linear feature extraction network is used to extract the motion data distribution features corresponding to each target motion data distribution information. Next, range adaptation is used to identify the motion anomaly assessment results corresponding to each motion data type. Then, the correspondence between each mooring force distribution feature range and the mooring force anomaly assessment information is preset. Finally, range adaptation is used to identify the sub-mooring force anomaly assessment results corresponding to each anchor chain, and all sub-mooring force anomaly assessment results are used as the final mooring force anomaly assessment result. Finally, the motion anomaly assessment results and the mooring force anomaly assessment results are filled into a preset mooring failure assessment report template to obtain the mooring failure assessment result for the floating wind turbine.

[0090] This embodiment constructs a platform numerical simulation model to effectively simulate the performance and function of floating wind turbines, thereby significantly improving the comprehensiveness and accuracy of performance data analysis. Then, this solution uses a constructed motion mooring prediction model, combined with the platform data simulation model, to perform simulation predictions, accurately outputting extreme values ​​for motion such as heave and pitch, ensuring that the prediction results meet engineering accuracy requirements. Finally, the above model relies only on environmental parameter inputs and does not require real-time platform motion data, reducing dependence on monitoring systems and providing a basis for application scenarios such as safe operation window decisions for floating wind power. This improves the predictive applicability in scenarios requiring long-term motion characteristic evolution prediction, such as maintenance window forecasting, thus comprehensively enhancing the accuracy of motion response prediction for floating wind turbines.

[0091] Furthermore, based on the above embodiments, this embodiment predicts the motion change information of the floating wind turbine based on the first simulation dataset, using a motion prediction model and a platform numerical simulation model. This includes: predicting the motion data distribution information of each motion data type of the floating wind turbine based on the first simulation dataset, using a motion prediction model and a platform numerical simulation model; performing distribution fitting processing on each motion data distribution information to obtain the target motion data distribution information corresponding to each motion data type; and using each target motion data distribution information as the motion change information of the floating wind turbine.

[0092] Specifically, the motion prediction model is input with 10-dimensional wind, wave, and current information, including meaningful wave height H. s Peak period T p Wave direction sine value sin(θ), wave direction cosine value cos(θ), flow velocity U c Flow direction sine value sin(α), flow direction cosine value cos(α), wind speed U 10 The wind direction sine (sin(ψ)) and cosine (cos(ψ)) values ​​are used to transform the one-dimensional angle information into x and y-direction information, compared to directly using the wind, wave, and current direction angle values ​​as feature inputs. This better reflects the periodicity of the angle. Verification shows that this method can reduce the model loss by about 50% without changing the structure. The model output is the maximum amplitude information of the six-degree-of-freedom motion of the 6-dimensional platform. The 10-dimensional features input to the model and the 6-dimensional labels output have different dimensions and orders of magnitude. Direct training would cause weight bias. Therefore, the Z-score standardization method is used to transform the input features and output labels into standard distributions with a mean of 0 and a standard deviation of 1, respectively.

[0093]

[0094] Where μ is the mean of sample X; σ is the standard deviation of sample X. In this embodiment, the model architecture is a 4-layer fully connected network with the following topology: input layer (10 nodes), hidden layers 1-3 (256 nodes), hidden layer 4 (128 nodes), and output layer (6 nodes). To suppress model overfitting, a dropout regularization mechanism is introduced after each hidden layer, with a dropout rate of 0.015 for each dropout layer. ReLU function activation is used after each dropout layer to introduce nonlinearity. The model optimization algorithm uses the Adam optimizer with an initial learning rate of 0.0005. The loss function uses the Mean Absolute Error (MAE) criterion instead of the Mean Squared Error (MSE) criterion, as defined in equations (4) and (5), respectively.

[0095]

[0096] In the formula, y i Represents the actual value; This represents the predicted value. Since the training data is based on 10-minute simulations with two different seeds calculated by a time-domain numerical model, the extreme value characteristics of the output are random. Therefore, MAE is used to reduce the influence of extreme values.

[0097] Table 2 lists the MAE and MSE of the motion prediction model on the test set. The MAE represents the mean absolute error of the model's predictions and can be directly used to measure the model's accuracy. The model's average error in predicting horizontal motion is about 15 cm, and the average error in predicting vertical motion is about 3 cm, which meets the needs of practical use.

[0098] Table 2: Accuracy Prediction Table of Motion Prediction Model

[0099] MAE MSE Swaying (m) 0.1394 0.0418 Swaying (m) 0.1536 0.0540 Drooping (m) 0.0300 0.0019 Sway (°) 0.1290 0.0316 Sway (°) 0.1305 0.0326 Bow rocking (°) 0.3447 0.2070

[0100] Furthermore, in this embodiment, based on the second simulation dataset, the mooring tension prediction model and the platform numerical simulation model are used to predict the mooring force change information of the floating wind turbine. This includes: based on the second simulation dataset, using the mooring tension prediction model and the platform numerical simulation model, predicting the tension distribution information of each anchor chain of the floating wind turbine, and determining the model error data of the mooring tension prediction model; based on the tension distribution information of each anchor chain and the model error data, identifying the target tension change distribution information of each anchor chain, and using the target tension change distribution information as the mooring force change information of the floating wind turbine.

[0101] Specifically, the mooring prediction model is similar to the aforementioned motion prediction model, but it is simplified because the dimensions and orders of magnitude of the maximum 9-dimensional mooring force output are consistent. In this embodiment, the mooring force prediction model adopts a 3-layer fully connected network with the following topology: input layer (10 nodes), hidden layers 1-3 (256 nodes), and output layer (9 nodes). The dropout rate of each dropout layer is set to 0.010, and other parameters are set consistent with the motion prediction model. The anchor chain tension prediction results are shown in Table 3.

[0102] Table 3: Prediction Accuracy of Mooring Tension Prediction Model

[0103]

[0104]

[0105] Table 3 details the mean absolute error (MAE) and mean square error (MSE) of the mooring tension prediction model on the test set. The results show that the maximum mean absolute error occurs on anchor chain 2, with a value of 75.29 kN, equivalent to 11% of the cable pretension. This error level meets the accuracy requirements for engineering applications.

[0106] Based on the above scheme, the 4-layer fully connected network motion prediction model used in this embodiment has a mean absolute error of only 0.133 on the test set, a training error of 0.084, and an error ratio of 1.59, showing no noteworthy overfitting. The coefficient of determination R between the predicted and actual values ​​on the validation set is [value missing]. 2 The accuracy rate (R0.97) is above 0.97, demonstrating that the model can effectively predict nonlinear relationships in platform motion data and output motion extremes such as heave and pitch with high precision. Secondly, the mooring tension prediction model shows controllable prediction errors for the maximum tension of the nine mooring cables. Except for anchor chains 1 and 2, whose prediction results are affected by extreme values, the R0.97 of the other anchor chains is consistent. 2 All exceeded 0.97. The anchor chain with the largest error, No. 2, had a MAE of 75.29 kN, which was only 11% of the pretension. The predicted results met the engineering accuracy requirements.

[0107] Based on the same general inventive concept, this invention also protects a mooring failure prediction device based on a floating wind turbine. The mooring failure prediction device based on a floating wind turbine described below and the mooring failure prediction method based on a floating wind turbine described above can be referred to in correspondence.

[0108] Figure 3 This is a schematic diagram of the mooring failure prediction device based on a floating wind turbine provided in this embodiment.

[0109] like Figure 3 As shown, this embodiment provides a mooring failure prediction device based on a floating wind turbine, comprising:

[0110] The acquisition module 301 is used to acquire platform data of the floating wind turbine and environmental data of the environment in which the floating wind turbine is located;

[0111] Module 302 is used to parse platform data and build a platform numerical simulation model of the floating wind turbine;

[0112] Extraction module 303 is used to split environmental data and extract key environmental parameters of each environmental parameter type;

[0113] The prediction module 304 is used to predict the motion change information and mooring force change information of the floating wind turbine based on various key environmental parameters, through the platform numerical simulation model and the motion mooring prediction model.

[0114] The evaluation module 305 is used to perform anomaly assessment on motion change information and mooring force change information using a mooring failure assessment strategy, and generate mooring failure assessment results for the floating wind turbine.

[0115] Figure 4 This is a schematic diagram of the structure of the electronic device provided in this embodiment.

[0116] like Figure 4 As shown, the electronic device may include a processor 401, a communications interface 402, a memory 403, and a communication bus 404. The processor 401, communications interface 402, and memory 403 communicate with each other via the communication bus 404. The processor 401 can call logical instructions from the memory 403 to execute a mooring failure prediction method based on a floating wind turbine.

[0117] Furthermore, the logical instructions in the aforementioned memory 403 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0118] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the mooring failure prediction method based on the floating wind turbine provided by the above methods.

[0119] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the mooring failure prediction method based on a floating wind turbine provided by the methods described above.

[0120] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0121] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting mooring failure based on a floating wind turbine, characterized in that, include: Acquire platform data of the floating wind turbine and environmental data of the environment in which the floating wind turbine is located; The platform data was analyzed to construct a platform numerical simulation model of the floating wind turbine; The environmental data is split, and key environmental parameters of each environmental parameter type are extracted; Based on the key environmental parameters, the motion change information and mooring force change information of the floating wind turbine are predicted through the platform numerical simulation model and the motion mooring prediction model. Using a mooring failure assessment strategy, anomaly assessments are performed on the motion change information and the mooring force change information to generate the mooring failure assessment results for the floating wind turbine.

2. The mooring failure prediction method based on a floating wind turbine according to claim 1, characterized in that, The process of parsing the platform data and constructing a platform numerical simulation model of the floating wind turbine includes: The platform data is broken down into structural data and structural performance parameters of each platform structure; Based on the structural data and structural performance parameters, a numerical simulation model of the substructure of each platform structure is constructed. The numerical simulation models of all the substructures are spliced ​​together to obtain the platform numerical simulation model of the floating wind turbine.

3. The mooring failure prediction method based on a floating wind turbine according to claim 1, characterized in that, The process of splitting the environmental data and extracting key environmental parameters for each environmental parameter type includes: The environmental data is split into environmental wind data and wave current data. Based on the environmental wind data and the wave current data, the environmental wind simulation data and wave current simulation data of the environment in which the floating wind turbine is located are simulated by a time-domain calculation model. Using a preset data extraction strategy, target environmental wind simulation data and target wave current simulation data are extracted from the environmental wind simulation data and the wave current simulation data, respectively. Based on the target environment wind simulation data and the target wave current simulation data, key environmental parameters for each environmental parameter type are extracted.

4. The mooring failure prediction method based on a floating wind turbine according to claim 1, characterized in that, The motion mooring prediction model includes a motion prediction model; Based on the aforementioned key environmental parameters, the motion change information of the floating wind turbine is predicted using the platform's numerical simulation model and motion mooring prediction model, including: Among the key environmental parameters, key environmental parameters of each first environmental parameter type required by the motion prediction model are selected. Based on the key environmental parameters of each of the first environmental parameter types, a first simulation dataset for the motion prediction model is constructed. Based on the first simulation dataset, the motion change information of the floating wind turbine is predicted through the motion prediction model and the platform numerical simulation model.

5. The mooring failure prediction method based on a floating wind turbine according to claim 4, characterized in that, The step of predicting the motion change information of the floating wind turbine based on the first simulation dataset, through the motion prediction model and the platform numerical simulation model, includes: Based on the first simulation dataset, the motion data distribution information of each motion data type of the floating wind turbine is predicted by the motion prediction model and the platform numerical simulation model. The distribution information of each motion data is subjected to distribution fitting processing to obtain the target motion data distribution information corresponding to each motion data type, and the target motion data distribution information is used as the motion change information of the floating wind turbine.

6. The mooring failure prediction method based on a floating wind turbine according to claim 5, characterized in that, The motion mooring prediction model includes a mooring tension prediction model; Based on the aforementioned key environmental parameters, the prediction of mooring force changes of the floating wind turbine using the platform's numerical simulation model and motion mooring prediction model includes: Among the key environmental parameters of each of the environmental parameter types, the key environmental parameters of each second environmental parameter type required by the mooring tension prediction model are selected. Based on the key environmental parameters of each of the second environmental parameter types, a second simulation dataset for the mooring tension prediction model is constructed. Based on the second simulation dataset, the mooring force variation information of the floating wind turbine is predicted using the mooring tension prediction model and the platform numerical simulation model.

7. The mooring failure prediction method based on a floating wind turbine according to claim 6, characterized in that, Based on the second simulation dataset, the mooring force variation information of the floating wind turbine is predicted using the mooring tension prediction model and the platform numerical simulation model, including: Based on the second simulation dataset, the tension distribution information of each anchor chain of the floating wind turbine is predicted by the mooring tension prediction model and the platform numerical simulation model, and the model error data of the mooring tension prediction model is determined. Based on the tension prediction distribution information of each anchor chain and the model error data, the target tension change distribution information of each anchor chain is identified, and the target tension change distribution information is used as the mooring force change information of the floating wind turbine.

8. The mooring failure prediction method based on a floating wind turbine according to claim 7, characterized in that, The method of utilizing a mooring failure assessment strategy to perform anomaly assessment on the motion change information and the mooring force change information, and generating a mooring failure assessment result for the floating wind turbine, includes: Based on the target motion data distribution information corresponding to each of the motion data types, the motion anomaly assessment results of the floating wind turbine are identified through a motion anomaly assessment strategy. Based on the target tension change distribution information of each anchor chain, the mooring force distribution information of each anchor chain is identified, and the mooring force anomaly assessment result of the floating wind turbine is identified through the mooring force anomaly assessment strategy. Based on the motion anomaly assessment results and the mooring force anomaly assessment results, the mooring failure assessment results of the floating wind turbine are generated.

9. A mooring failure prediction device based on a floating wind turbine, characterized in that, include: The acquisition module is used to acquire platform data of the floating wind turbine and environmental data of the environment in which the floating wind turbine is located; A construction module is used to parse the platform data and construct a platform numerical simulation model of the floating wind turbine; The extraction module is used to split the environmental data and extract key environmental parameters for each type of environmental parameter. The prediction module is used to predict the motion change information and mooring force change information of the floating wind turbine based on the key environmental parameters, through the platform numerical simulation model and the motion mooring prediction model. The evaluation module is used to perform anomaly evaluation on the motion change information and the mooring force change information using a mooring failure evaluation strategy, and generate the mooring failure evaluation result of the floating wind turbine.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the mooring failure prediction method based on a floating wind turbine as described in any one of claims 1 to 8.

Images