A method for diagnosing and predicting the operating state fault of a wind generator and a gear box

Abstract

The present application relates to the technical field of fan fault diagnosis and prediction, and discloses a kind of operating state fault diagnosis and prediction method of fan generator and gear box, comprising: the establishment of multi-field coupling model, including: fan generator and fan gear box equivalent model;Establish data processing model to fan operation data pre-processing and analysis, obtain the characteristic data of generator and gear box;The accuracy of simulation result is verified to multi-field coupling model using characteristic data;Real-time operating state data of wind power generator and gear box is obtained, and the simulation data and characteristic data corresponding to the real-time operating state data of wind power generator and gear box are mixed and then input into fault diagnosis and prediction model respectively, to obtain real-time operating state fault diagnosis result and future preset time fault prediction result.The present application monitors the operating state of fan gear box and generator in real time, diagnoses and predicts the occurrence of fault, improves the operation and maintenance efficiency of fan, and effectively reduces the operation and maintenance cost.

Description

Technical Field

[0001] This invention relates to the field of wind turbine fault diagnosis and prediction technology, specifically to a method for diagnosing and predicting operational status faults in wind turbine generators and gearboxes. Background Technology

[0002] Wind power, as an emerging clean energy source, is one of the most technologically mature and commercially viable renewable energy sources. However, frequent failures lead to significant operation and maintenance costs, and among the many components of a wind turbine, such as… Figure 1 and Figure 2 The gearboxes and generators shown account for a significant proportion in terms of both failure rate and downtime. The operation and maintenance process generally includes steps such as fault occurrence, alarm, personnel dispatch, fault location, solution development, and component replacement. This process involves extremely long maintenance cycles and high labor costs, resulting in low efficiency and further increasing the economic losses caused by wind turbine failures. Therefore, efficient wind turbine operation status monitoring, fault diagnosis, and prediction systems are indispensable for the further development of wind power.

[0003] Existing diagnostic techniques either simplify the gearbox and generator into parametric / algebraic models, focusing only on parameter transmission and losing the description of the actual physical mechanism of wind turbine operation; or use finite element models to describe the multiphysics of the gearbox and generator, but such models have an excessive amount of computation, place extremely high demands on the equipment's computing power, and have a slow calculation speed, making it far from being able to achieve real-time monitoring and timely fault diagnosis and prediction of the operating status. Summary of the Invention

[0004] In view of this, in order to solve the shortcomings of low monitoring efficiency in the existing wind power operation process, which leads to low operation efficiency and high cost, this invention provides a method, device, equipment and medium for fault diagnosis and prediction of the operating status of wind turbine generators and gearboxes. It can monitor the operating status of wind turbine gearboxes and generators in real time, diagnose and predict the occurrence of faults. With the help of a web publishing platform, it can realize real-time online monitoring of their status, which greatly improves the operation and maintenance efficiency during wind turbine operation and effectively reduces operation and maintenance costs.

[0005] In a first aspect, the present invention provides a method for diagnosing and predicting operational status faults in wind turbine generators and gearboxes, comprising:

[0006] A multi-field coupling model is established based on the physical laws, energy transfer, and mutual coupling relationships between multiple physical fields involved in the actual operation of the equipment. The multi-field coupling model includes: an equivalent model of the wind turbine generator and an equivalent model of the wind turbine gearbox.

[0007] A data processing model is established to preprocess and analyze the wind turbine operation data stored in the SCADA system and CMS system to obtain the characteristic data of the wind turbine generator and gearbox.

[0008] The accuracy of the simulation results of the multi-field coupling model is verified by using the feature data provided by the data processing model;

[0009] The real-time operating status data of the wind turbine and gearbox are acquired, and the corresponding simulation data is obtained by using the multi-field coupling model after accuracy verification, and the corresponding feature data is extracted by the data processing model. The simulation data and feature data corresponding to the real-time operating status data of the wind turbine and gearbox are mixed to obtain the mixed data.

[0010] The mixed data is input into the fault diagnosis model and the fault prediction model respectively to obtain the corresponding real-time operating status fault diagnosis results and the fault prediction results within a preset time period.

[0011] The method for fault diagnosis and prediction of the operating status of wind turbine generators and gearboxes provided in this embodiment can monitor the operating status of wind turbine gearboxes and generators in real time, perform fault diagnosis and prediction, and, in conjunction with a web publishing platform, realize real-time online monitoring of their status, which greatly improves the operation and maintenance efficiency during wind turbine operation and effectively reduces operation and maintenance costs.

[0012] In one alternative implementation, an equivalent model of the wind turbine gearbox is established based on the theoretical foundation of the temperature field and dynamic model.

[0013] The equivalent model of the wind turbine generator constructed in this embodiment involves multiple complex physical processes based on circuits, magnetic circuits, thermal circuits, and mechanical forces. It can predict the electromagnetic characteristics, magnetic field distribution, thermal state, and mechanical stress of the wind turbine generator under various operating conditions, providing a scientific basis for the design, manufacturing, and maintenance of the generator, improving the performance and efficiency of the generator, while reducing energy consumption and failure rate.

[0014] In one alternative implementation, an equivalent model of the wind turbine generator is established based on the theoretical foundations of circuits, magnetic circuits, thermal circuits, and mechanical forces.

[0015] The equivalent model of the gearbox established in this embodiment of the invention takes into account the physical laws of temperature field and dynamics, which can help predict the performance and life of the gearbox, thereby optimizing the design and maintenance strategy of the gearbox and improving its operating efficiency and reliability.

[0016] In one optional implementation, the data processing model is established to preprocess and analyze the wind turbine operating data stored in the SCADA and CMS systems to obtain characteristic data of the wind turbine generator and gearbox, including:

[0017] The system performs tagging, calibration and classification of operating conditions, replacement of abnormal data, selection of feature data and data standardization for the wind turbine operating data stored in the SCADA system and CMS system.

[0018] The standardized data is analyzed to form characteristic data of electrical parameters, control signals, system status, and temperature distribution of components from the data in the SCADA system, and characteristic data of component system performance, friction and vibration, mechanical defects, and fatigue damage from the data in the CMS system.

[0019] The feature data obtained through the data processing model implemented in this invention not only helps to deeply understand the operating status of wind turbines, but also provides strong support for predictive maintenance. By analyzing the operating data of wind turbines, potential problems can be identified in a timely manner, and preventive maintenance can be carried out before the problems become serious. This approach can significantly reduce downtime caused by wind turbine failures, thereby improving the overall operating efficiency and power generation of wind turbines. It can also be used to optimize wind turbine design and operating strategies, and to provide feedback to wind turbine designers to achieve better gearbox and generator designs.

[0020] In one optional implementation, the process of constructing the fault diagnosis model includes:

[0021] Simulation data and feature data corresponding to the historical operating status data of wind turbine generators and gearboxes are obtained and mixed to obtain hybrid data;

[0022] Clustering the mixed data yields operational failure data;

[0023] The operational failure data is further clustered to obtain the failure type.

[0024] In one alternative implementation, the mixed data and operational failure data are clustered separately based on the Pearson correlation coefficient.

[0025] In one optional implementation, the fault diagnosis model obtains fault diagnosis results based on time series similarity metrics, and the process includes:

[0026] Based on time series Motif features, and taking the normal operating condition data of the equipment as a benchmark, the Pearson correlation coefficient between the current operating condition data and the normal operating condition data is obtained.

[0027] The presence and type of fault are determined by the Pearson correlation coefficient between the current operating status data and the normal operating condition data.

[0028] In this embodiment of the invention, the Pearson correlation coefficient is used to reflect the degree of linear correlation between two variables. The distance between the operational data is determined by the Pearson correlation coefficient, thereby identifying the operational data relative to normal and abnormal operation.

[0029] In this embodiment of the invention, the process of determining whether a fault exists and the type of fault based on the Pearson correlation coefficient includes:

[0030] If the Pearson correlation coefficient between the current operating status data and the normal operating status data is greater than the preset threshold, it is judged to be operating normally;

[0031] If the Pearson correlation coefficient between the current operating status data and the normal operating status data is less than a preset threshold, the operating fault type is determined according to the preset threshold range corresponding to its fault type.

[0032] In one alternative implementation, the fault prediction model is trained based on a temporal convolutional neural network.

[0033] The embodiments of this invention use a temporal convolutional neural network (TCN) because it has the following characteristics: parallelism, which allows it to process sentences in parallel, unlike RNNs which process them sequentially; flexible receptive field, whose size is determined by the number of layers, kernel size, and dilation coefficients, and can be flexibly customized according to different tasks and characteristics; stable gradients, as RNNs often suffer from vanishing and exploding gradients, mainly due to the sharing of parameters across different time periods, unlike traditional convolutional neural networks; and lower memory usage, as RNNs require storing information at each step, which consumes a lot of memory, while TCNs share kernels within a single layer, resulting in lower memory consumption.

[0034] In one optional implementation, the process of training a temporal convolutional neural network to obtain the fault prediction model includes:

[0035] The simulation data and feature data corresponding to the operating status data of the wind turbine and gearbox in the preset historical period are obtained, and the mixed data after mixing is divided into training set and validation set according to the preset ratio.

[0036] The mixed data of a preset number of historical days in the training set is input into the temporal convolutional neural network to obtain the operating status of the wind turbine and gearbox on a future day.

[0037] The predicted operating status of the wind turbine and gearbox is added to the historical data. The next step is to update the mixed data of the preset number of historical days and input it into the temporal convolutional neural network to obtain the substation operating status on a future day. This process is repeated to obtain the predicted data of the operating status of the wind turbine and gearbox for multiple future days.

[0038] The trained model is validated using a validation set, and the training parameters are adjusted appropriately based on the validation results to optimize the model. The model with the best validation performance is selected as the fault prediction model.

[0039] This invention trains a temporal convolutional neural network to learn the causal relationship between the occurrence of an event and changes in the data prior to the event in historical data, and uses this relationship to predict the occurrence of faults.

[0040] Secondly, the present invention provides a device for diagnosing and predicting operational status faults in wind turbine generators and gearboxes, the device comprising:

[0041] The multi-field coupling model construction module is used to establish a multi-field coupling model based on the physical laws, energy transfer and mutual coupling relationships between multiple physical fields involved in the actual operation of the equipment. The multi-field coupling model includes: an equivalent model of the wind turbine generator and an equivalent model of the wind turbine gearbox.

[0042] The data processing model building module is used to establish a data processing model to preprocess and analyze the wind turbine operation data stored in the SCADA system and CMS system, and obtain the characteristic data of the wind turbine generator and gearbox.

[0043] The multi-field coupling model verification module is used to verify the accuracy of the simulation results of the multi-field coupling model using the feature data provided by the data processing model;

[0044] The real-time operation data acquisition and processing module is used to acquire real-time operation status data of wind turbine generators and gearboxes, and use the corresponding simulation data obtained by the multi-field coupling model after accuracy verification, and extract the corresponding feature data through the data processing model. The simulation data and feature data corresponding to the real-time operation status data of wind turbine generators and gearboxes are mixed to obtain mixed data.

[0045] The fault diagnosis and prediction module is used to input the mixed data into the fault diagnosis model and the fault prediction model respectively to obtain the corresponding real-time operating status fault diagnosis results and the fault prediction results within a preset time period.

[0046] In one optional implementation, the multi-field coupling model construction module includes:

[0047] The gearbox equivalent model building unit is used to establish the equivalent model of the wind turbine gearbox based on the theoretical basis of temperature field and dynamic model.

[0048] The wind turbine generator equivalent model construction unit is used to establish an equivalent model of the wind turbine generator based on the theoretical foundations of circuit, magnetic circuit, thermal circuit, and mechanical force.

[0049] In one optional implementation, the data processing model building module includes:

[0050] The preprocessing unit is used to set tags, calibrate and classify operating conditions, replace abnormal data, select feature data, and standardize data for the wind turbine operating data stored in the SCADA system and CMS system.

[0051] The analysis unit is used to analyze standardized data, transforming data from the SCADA system into characteristic data of electrical parameters, control signals, system status, and temperature distribution of components, and transforming data from the CMS system into characteristic data of component system performance, friction and vibration, mechanical defects, and fatigue damage.

[0052] Thirdly, the present invention provides a computer device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the above-described method for diagnosing and predicting the operating status faults of a wind turbine generator and gearbox according to the first aspect or any corresponding embodiment thereof.

[0053] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for diagnosing and predicting the operating status faults of a wind turbine generator and gearbox as described in the first aspect or any corresponding embodiment thereof. Attached Figure Description

[0054] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0055] Figure 1 A diagram illustrating the maintenance cost ratios of various components of wind power equipment;

[0056] Figure 2 A schematic diagram showing the downtime ratio of individual components in wind power equipment;

[0057] Figure 3 This is a flowchart illustrating the method for diagnosing and predicting the operating status faults of wind turbine generators and gearboxes according to an embodiment of the present invention.

[0058] Figure 4 This is a schematic diagram of multi-field coupling of a wind turbine generator provided in an embodiment of the present invention;

[0059] Figure 5This is a schematic diagram of multi-field coupling in a wind turbine gearbox provided in an embodiment of the present invention;

[0060] Figure 6 This is a schematic diagram of the three-stage gearbox structure of the wind turbine generator provided in an embodiment of the present invention;

[0061] Figure 7 This is a schematic diagram illustrating the interrelationships between the various models provided in the embodiments of the present invention;

[0062] Figure 8 This is a schematic diagram illustrating the results of clustering mixed data provided in an embodiment of the present invention;

[0063] Figure 9 This is a schematic diagram showing the results of clustering fault types according to an embodiment of the present invention;

[0064] Figure 10 This is a schematic diagram illustrating the verification of the fault prediction model provided in the embodiments of the present invention;

[0065] Figure 11 This is a structural block diagram of the wind turbine generator and gearbox operation status fault diagnosis and prediction device provided in an embodiment of the present invention;

[0066] Figure 12 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0068] According to an embodiment of the present invention, a method for diagnosing and predicting the operating status faults of wind turbine generators and gearboxes is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0069] This embodiment provides a method for diagnosing and predicting operational status faults in wind turbine generators and gearboxes, which can be used in computer equipment terminals. Figure 3 This is a flowchart of a method for diagnosing and predicting operational status faults in wind turbine generators and gearboxes according to an embodiment of the present invention, such as... Figure 3 As shown, the process includes the following steps:

[0070] Step S101: Establish a multi-field coupling model based on the physical laws, energy transfer and mutual coupling relationship between multiple physical fields involved in the actual operation of the equipment. The multi-field coupling model includes: the wind turbine generator equivalent model and the wind turbine gearbox equivalent model.

[0071] Because most existing simulation models use mathematical models to describe wind turbines, they fail to consider the complexity of actual operation or the interference between multiple fields, resulting in simulation results that deviate significantly from reality. This invention integrates multiphysics information, making the simulation results closer to the actual operation of the wind turbine. By establishing a multiphysics coupling model, this invention helps to better understand and optimize the equipment's operation. By simulating the actual operation of the equipment, the model can predict and analyze the equipment's behavior and performance under various operating conditions.

[0072] A schematic diagram of multi-field coupling of the generator in an embodiment of the present invention is shown below. Figure 4 As shown, the constructed equivalent model of the wind turbine generator involves multiple complex physical processes based on circuits, magnetic circuits, thermal circuits, and mechanical forces. It can predict the electromagnetic characteristics, magnetic field distribution, thermal state, and mechanical stress of the wind turbine generator under various operating conditions, providing a scientific basis for the design, manufacturing, and maintenance of the generator, improving its performance and efficiency, while reducing energy consumption and failure rate. Specifically, the equivalent model includes:

[0073] The mechanical force equivalent model includes the mechanical power P1 input and the electrical power P2 output of the wind turbine system, and the corresponding calculation formulas are as follows:

[0074] P1 = Tn / 9550

[0075] P2 = UI / 1000

[0076] Where T is the input torque, n is the rotational speed, 9550 is the torque coefficient, U is the stator phase voltage, I is the stator winding current, and 1000 is the unit conversion factor.

[0077] The inputs to the magnetic circuit equivalent model are the permeability and current magnitude, and the output is the flux linkage at each observation node. The corresponding calculation formulas are as follows:

[0078]

[0079] Where Ψ represents the magnetic flux linkage, N represents the number of coil turns, B is the magnetic field strength, S is the equivalent cross-sectional area through which the magnetic field lines pass, and f e Let l be the magnitude of the electromagnetic force, l be the length of the conductor subjected to the electromagnetic force, and i be the induced current.

[0080] The calculations involved in the circuit equivalent model are as follows:

[0081]

[0082] Where e is the electromotive force, B is the magnetic field strength, v is the linear velocity of the rotor, l is the length of the conductor subjected to electromagnetic force, U is the stator phase voltage, I is the stator winding current, and R is the equivalent resistance of the motor.

[0083] The calculations involved in the thermal equivalent model are as follows:

[0084]

[0085] Among them, P Cu For copper loss, P Fe For iron loss, I rms R is the effective value of the current. p It is the equivalent resistance of the winding, K h ,K c ,K e B is a constant coefficient. m Where is the maximum magnetic field strength, and f is the voltage frequency. In practice, the resistance and permeability of magnetic materials are affected by temperature; the resistance and permeability of a specific material at a specific temperature need to be obtained from a table.

[0086] A schematic diagram of the multi-field coupling of the gearbox in an embodiment of the present invention is shown below. Figure 5 As shown, the establishment of the equivalent model of the gearbox, considering the physical laws of temperature field and dynamics, can help predict the performance and lifespan of the gearbox, thereby optimizing the design and maintenance strategies of the gearbox and improving its operating efficiency and reliability. Specifically, the equivalent model includes:

[0087] This invention provides a kinematic modeling of the gearbox by transforming the loads acting on the structure into concentrated forces acting on concentrated mass blocks using an equivalent transfer method. These concentrated mass blocks are connected by equivalent elastic and damping elements, thus forming a discrete dynamic model consisting of concentrated masses, springs, and damping elements, with concentrated forces acting on the relevant mass blocks. This simplifies complex engineering problems due to the often intricate geometry and boundary conditions of mechanical structures into relatively simpler problems. Analyzing the simplified dynamic model and constructing kinematic differential equations simplifies the solution. In a specific embodiment, the structure of a three-stage gearbox for a wind turbine is as follows: Figure 6 As shown, Figure 6 In the diagram, 'c' represents the planet carrier, 'p' represents the planet gears, 's' represents the sun gear, '1' and '2' are the medium-speed gears of the parallel shaft system, and '3' and '4' are the high-speed gears of the parallel shaft system. In the first-stage planetary gear train, the internal gear ring is fixed to the inner wall of the gearbox and is assumed to be a rigid body. All other gears are fixed to specific positions in the gearbox via bearings.

[0088] This invention models the temperature field of a gearbox, employing a circuit calculation model to represent the gearbox's temperature field. This allows for the use of circuit calculation methods to calculate the temperature field, significantly reducing computational burden while ensuring model accuracy. In the temperature field calculation, temperature is equated to voltage, heat to current, and thermal resistance to resistance, thus simplifying the temperature field model and greatly reducing computational scale. This method subdivides the research object into several unit nodes, assuming each node is a unit with lumped parameters. Specifically, it assumes uniform internal temperature within each node, and that heat transfer between nodes, whether through conduction, convection, or radiation, is achieved through thermal resistance connections, forming a thermal network. Node heat flux includes self-generated heat, external heating loads, and heat transfer flux between nodes. The thermal network model introduces the concepts of thermal resistance and heat capacity, and, using Kirchhoff's laws, establishes the following set of node thermal balance equations based on the law of conservation of heat.

[0089]

[0090] Where, q n Let Vn be the heat generation rate of the heat source per unit volume at node n; Vn be the volume of the heat flux cylinder; f(T) be a function of temperature T; and Rn be the heat transfer rate. j-n C is the thermal resistance between node j and node n; n Δt is the heat capacity in J / (kg·K); Δt is the time interval from time i to time i+1.

[0091] The temperature, heat flux, and their variations at each node can be solved using the nodal heat balance equations. Thermal resistance can be used to simulate the steady-state temperature field of a fan gearbox, while adding heat capacity allows for the establishment of a transient temperature field model. The temperature, heat flux, and their variations at each node can be solved using the nodal heat balance equations.

[0092] The structure-based simulation model provided in this invention can explain phenomena in data models, helping to understand various complex phenomena and problems in equipment operation and providing coping strategies, thereby further improving equipment operating efficiency and reliability. Through the above equivalent model, the equipment operation process can be better understood and optimized, improving equipment performance and lifespan, saving energy and maintenance costs, and further promoting the sustainable development of equipment.

[0093] Step S102: Establish a data processing model to preprocess and analyze the wind turbine operation data stored in the SCADA system and CMS system to obtain the characteristic data of the wind turbine generator and gearbox.

[0094] In practical applications, wind turbine operating data is typically stored in a Supervisory Control and Data Acquisition (SCADA) system and a Condition Monitoring and Management (CMS) system. This invention cleans, re-summarizes, and structures the data stored in these two systems to obtain characteristic data for the wind turbine generator and gearbox, forming a data-level description of the wind turbine gearbox and generator. Data from the SCADA system can be used to describe the electrical parameters, control signals, system status, and temperature distribution of components. Data from the CMS system can be used to describe the system performance, friction and vibration, mechanical defects, and fatigue damage of components. This data not only helps in understanding the wind turbine's operating status but also provides strong support for predictive maintenance. By analyzing the wind turbine's operating data, potential problems can be identified in a timely manner, and preventative maintenance can be performed before problems become serious, significantly reducing downtime due to wind turbine failures and thus improving the overall operating efficiency and power generation of the wind turbine.

[0095] Furthermore, data-level component descriptions can be used to optimize wind turbine design and operating strategies. For example, by analyzing electrical parameters, control signals, system states, and temperature distribution, optimal operating parameters can be identified to achieve the highest energy efficiency. Analysis of system performance, frictional vibration, mechanical defects, and fatigue damage can provide feedback to wind turbine designers for better gearbox and generator designs.

[0096] Step S103: Verify the accuracy of the simulation results of the multi-field coupling model using the feature data provided by the data processing model.

[0097] This invention, through verification of the accuracy of the multi-field coupling model, improves the model's accuracy and reliability, thereby enhancing the wind turbine's performance prediction and fault diagnosis capabilities. This is of great significance for the long-term stable operation of wind turbines, especially under extreme climatic conditions.

[0098] Step S104: Obtain real-time operating status data of wind turbine and gearbox, and use the multi-field coupling model after accuracy verification to obtain the corresponding simulation data and extract the corresponding feature data through the data processing model.

[0099] In practical applications, there is a phenomenon of insufficient or missing implementation status data. Therefore, in this embodiment of the invention, the simulation data obtained by the multi-field coupling model after accuracy verification and the corresponding feature data extracted by the data processing model are mixed to obtain mixed data, which can yield rich and comprehensive operation status data.

[0100] Step S105: Input the mixed data into the fault diagnosis model and the fault prediction model respectively to obtain the corresponding real-time operating status fault diagnosis results and the fault prediction results within a preset time period.

[0101] This invention employs hybrid data instead of conventional single simulation data or single actual data. Simulation data can supplement fault data with events at zero cost, while actual data calibrates and verifies the simulation model. The simulation model can also help to inversely analyze and verify the mechanism of fault occurrence. Diagnostic and prediction results based on hybrid data are superior.

[0102] The relationships between the various models provided in the embodiments of the present invention are as follows: Figure 7 As shown, the fault diagnosis and prediction model uses a hybrid data model based on multi-field coupling and data processing as input. In the fault diagnosis part, based on time-series Motif features and using normal operating condition data of the equipment as a benchmark, if the similarity distance between the current event and the normal operating condition exceeds a specified threshold, a fault is determined to have occurred. The fault prediction part is based on a temporal convolutional neural network, which learns the causal relationship between the occurrence of an event and the changes in data before its occurrence in historical data, and uses this as a basis to predict the occurrence of faults. Both the fault diagnosis and prediction calculations of this invention can be completed in a very short time, meeting real-time requirements.

[0103] In one specific embodiment, the process of constructing a fault diagnosis model includes:

[0104] A1. Obtain simulation data and feature data corresponding to the historical operating status data of wind turbine generator and gearbox, and mix them to obtain mixed data;

[0105] A2, clustering the mixed data to obtain operational failure data;

[0106] A3, further cluster the operational fault data to obtain the fault type.

[0107] Specifically, such as Figure 8 As shown, clustering the mixed data yielded a clear division, measured by the Pearson correlation coefficient, whose absolute value always falls between 0 and 1. A value closer to 1 indicates a stronger correlation, while a value closer to 0 indicates a weaker correlation. The clustering results show that Class_0 is significantly different from the other three classes and is very close to 0, indicating that the equipment operating state represented by these data is far from normal operating conditions, thus classifying it as a fault. Based on this, further... Figure 7 Class_0 in the dataset was further clustered to distinguish its fault types, and the results are as follows: Figure 9As shown, the fault category Class_0 was clustered with 3 subcategories and further divided according to the severity of the fault, labeled as Fault_0, Fault_1, and Fault_2. Subsequent verification of these three fault categories yielded results consistent with previous conclusions: the closer a fault is to 0, the further it is from the normal operating state of the equipment.

[0108] Based on the above clustering process, the process of obtaining fault diagnosis results based on time series similarity measurement in this embodiment of the invention includes: obtaining the Pearson correlation coefficient between the current operating status data and the normal operating status data based on time series Motif features and using the normal operating status data of the equipment as a benchmark; determining whether there is a fault and the fault type based on the Pearson correlation coefficient between the current operating status data and the normal operating status data; if the Pearson correlation coefficient between the current operating status data and the normal operating status data is greater than a preset threshold, it is determined that the operation is normal; if the Pearson correlation coefficient between the current operating status data and the normal operating status data is less than a preset threshold, the operating fault type is determined according to the preset threshold interval corresponding to its fault type. For example: the Pearson correlation coefficient P between the current operating status data and the normal operating status data, the fault and fault type are determined based on the value of P, and the specific correspondence is as follows:

[0109] When 0 ≤ P < 0.23, it represents a fault, and the fault type is Fault_0;

[0110] When 0.23≤P<0.27, it represents a fault, and the fault type is Fault_1;

[0111] When 0.27≤P<0.5, it represents a fault, and the fault type is Fault_2;

[0112] When 0.5≤P≤1, it indicates no fault.

[0113] It should be noted that the various thresholds mentioned above are for illustrative purposes only and are not intended to be limiting.

[0114] The fault prediction stage in this embodiment of the invention employs a temporal convolutional neural network. With its unique causal convolution, dilated convolution, and residual connection methods, it can learn the causal relationships between events and data in historical data over a wide time range, and predict the occurrence of events based on this.

[0115] The process of training a temporal convolutional neural network to obtain the fault prediction model in this embodiment of the invention includes:

[0116] B1: Obtain simulation data and feature data corresponding to the operating status data of the wind turbine generator and gearbox in the preset historical period, and divide the mixed data into training set and validation set according to the preset ratio; for example, training set: validation set = 8:2;

[0117] B2: Input the mixed data of a preset number of historical days in the training set into the temporal convolutional neural network to obtain the operating status of the wind turbine and gearbox on a future day;

[0118] B3: Add the predicted operating status of the wind turbine and gearbox to the historical data, update the next step of the mixed data with a preset number of historical days, input it into the temporal convolutional neural network, and obtain the substation operating status on a future day; repeat this process to obtain the predicted data of the operating status of the wind turbine and gearbox for multiple future days.

[0119] B4: Validate the trained model using the validation set and optimize the model by adjusting the training parameters appropriately based on the validation results. Select the model with the best validation performance as the fault prediction model.

[0120] This invention uses a real-world validation set to validate the trained prediction model. The validation results are as follows: Figure 10 As shown, when the features of the real data show a trend of moving towards the fault area, this prediction model can capture this trend well and predict its subsequent movement.

[0121] In one specific embodiment, the mixed data corresponding to the historical operation data from the 1st to the 10th of a certain month can be input into a temporal convolutional neural network to obtain the predicted operation status on the 11th. The data obtained on the 11th is then added to the historical data. The mixed data corresponding to the operation data from the 2nd to the 11th of a certain month is then input into the temporal convolutional neural network to predict the operation data on the 12th. This process is repeated, for example, to obtain the predicted indicator data for seven days from the 11th to the 16th of a certain month. This is just an example and is not a limitation.

[0122] This embodiment also provides a fault diagnosis and prediction device for the operating status of a wind turbine generator and gearbox. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0123] This embodiment provides a device for diagnosing and predicting operational status faults in wind turbine generators and gearboxes, such as... Figure 11 As shown, it includes:

[0124] The multi-field coupling model construction module 1101 is used to establish a multi-field coupling model based on the physical laws, energy transfer and mutual coupling relationship between multiple physical fields involved in the actual operation of the equipment. The multi-field coupling model includes: an equivalent model of the wind turbine generator and an equivalent model of the wind turbine gearbox.

[0125] The data processing model construction module 1102 is used to establish a data processing model to preprocess and analyze the wind turbine operation data stored in the SCADA system and CMS system to obtain the characteristic data of the wind turbine generator and gearbox.

[0126] The multi-field coupling model verification module 1103 is used to verify the accuracy of the simulation results of the multi-field coupling model using the feature data provided by the data processing model;

[0127] The real-time operation data acquisition and processing module 1104 is used to acquire real-time operation status data of wind turbine generator and gearbox, and use the corresponding simulation data obtained by the multi-field coupling model after accuracy verification, and extract the corresponding feature data through the data processing model, and mix the simulation data and feature data corresponding to the real-time operation status data of wind turbine generator and gearbox to obtain mixed data.

[0128] The fault diagnosis and prediction module 1105 is used to input the mixed data into the fault diagnosis model and the fault prediction model respectively to obtain the corresponding real-time operating status fault diagnosis results and the fault prediction results within a preset time period.

[0129] In one optional implementation, the multi-field coupling model construction module includes:

[0130] The gearbox equivalent model building unit is used to establish the equivalent model of the wind turbine gearbox based on the theoretical basis of temperature field and dynamic model.

[0131] The wind turbine generator equivalent model construction unit is used to establish an equivalent model of the wind turbine generator based on the theoretical foundations of circuit, magnetic circuit, thermal circuit, and mechanical force.

[0132] In one optional implementation, the data processing model building module includes:

[0133] The preprocessing unit is used to set tags, calibrate and classify operating conditions, replace abnormal data, select feature data, and standardize data for the wind turbine operating data stored in the SCADA system and CMS system.

[0134] The analysis unit is used to analyze standardized data, transforming data from the SCADA system into characteristic data of electrical parameters, control signals, system status, and temperature distribution of components, and transforming data from the CMS system into characteristic data of component system performance, friction and vibration, mechanical defects, and fatigue damage.

[0135] In one optional implementation, the process of constructing the fault diagnosis model includes:

[0136] Simulation data and feature data corresponding to the historical operating status data of wind turbine generators and gearboxes are obtained and mixed to obtain hybrid data;

[0137] Clustering the mixed data yields operational failure data;

[0138] The operational failure data is further clustered to obtain the failure type.

[0139] In one alternative implementation, the mixed data and operational failure data are clustered separately based on the Pearson correlation coefficient.

[0140] In one optional implementation, the fault diagnosis model obtains fault diagnosis results based on time series similarity metrics, and the process includes:

[0141] Based on time series Motif features, and taking the normal operating condition data of the equipment as a benchmark, the Pearson correlation coefficient between the current operating condition data and the normal operating condition data is obtained.

[0142] The presence and type of fault are determined by the Pearson correlation coefficient between the current operating status data and the normal operating condition data.

[0143] In this embodiment of the invention, the process of determining whether a fault exists and the type of fault based on the Pearson correlation coefficient includes:

[0144] If the Pearson correlation coefficient between the current operating status data and the normal operating status data is greater than the preset threshold, it is judged to be operating normally;

[0145] If the Pearson correlation coefficient between the current operating status data and the normal operating status data is less than a preset threshold, the operating fault type is determined according to the preset threshold range corresponding to its fault type.

[0146] In one optional implementation, the fault prediction model is obtained by training a temporal convolutional neural network. The process of training the temporal convolutional neural network to obtain the fault prediction model includes:

[0147] The simulation data and feature data corresponding to the operating status data of the wind turbine and gearbox in the preset historical period are obtained, and the mixed data after mixing is divided into training set and validation set according to the preset ratio.

[0148] The mixed data of a preset number of historical days in the training set is input into the temporal convolutional neural network to obtain the operating status of the wind turbine and gearbox on a future day.

[0149] The predicted operating status of the wind turbine and gearbox is added to the historical data. The next step is to update the mixed data of the preset number of historical days and input it into the temporal convolutional neural network to obtain the substation operating status on a future day. This process is repeated to obtain the predicted data of the operating status of the wind turbine and gearbox for multiple future days.

[0150] The trained model is validated using a validation set, and the training parameters are adjusted appropriately based on the validation results to optimize the model. The model with the best validation performance is selected as the fault prediction model.

[0151] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0152] This invention also provides a computer device having the above-described features. Figure 11 The device shown is for diagnosing and predicting the operating status of wind turbine generators and gearboxes.

[0153] Please see Figure 12 , Figure 12 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 12 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 12 Take a processor 10 as an example.

[0154] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0155] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.

[0156] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0157] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0158] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0159] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0160] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for diagnosing and predicting operational status faults in wind turbine generators and gearboxes, characterized in that, include: A multi-field coupling model is established based on the physical laws, energy transfer, and mutual coupling relationships between multiple physical fields involved in the actual operation of the equipment. The multi-field coupling model includes: an equivalent model of the wind turbine generator and an equivalent model of the wind turbine gearbox. A data processing model is established to preprocess and analyze the wind turbine operation data stored in the SCADA system and CMS system to obtain the characteristic data of the wind turbine generator and gearbox. The accuracy of the simulation results of the multi-field coupling model is verified by using the feature data provided by the data processing model; The real-time operating status data of the wind turbine and gearbox are acquired, and the corresponding simulation data is obtained by using the multi-field coupling model after accuracy verification, and the corresponding feature data is extracted by the data processing model. The simulation data and feature data corresponding to the real-time operating status data of the wind turbine and gearbox are mixed to obtain the mixed data. The mixed data is input into the fault diagnosis model and the fault prediction model respectively to obtain the corresponding real-time operating status fault diagnosis results and the fault prediction results within a preset time period. The fault diagnosis model obtains fault diagnosis results based on time series similarity metrics, and the process includes: Based on time series Motif features, and taking the normal operating condition data of the equipment as a benchmark, the Pearson correlation coefficient between the current operating condition data and the normal operating condition data is obtained. Determine whether there is a fault and the type of fault based on the Pearson correlation coefficient between the current operating status data and the normal operating condition data. The fault prediction model is obtained by training a temporal convolutional neural network. The process of training the temporal convolutional neural network to obtain the fault prediction model includes: The simulation data and feature data corresponding to the operating status data of the wind turbine and gearbox in the preset historical period are obtained, and the mixed data after mixing is divided into training set and validation set according to the preset ratio. The mixed data of a preset number of historical days in the training set are input into the temporal convolutional neural network to obtain the operating status of the wind turbine and gearbox on a future day. The predicted operating status of the wind turbine and gearbox is added to the historical data. The next step is to update the mixed data of the preset number of historical days and input it into the temporal convolutional neural network to obtain the substation operating status on a future day. This process is repeated to obtain the predicted data of the operating status of the wind turbine and gearbox for multiple future days. The trained model is validated using a validation set, and the training parameters are adjusted appropriately based on the validation results to optimize the model. The model with the best validation performance is selected as the fault prediction model.

2. The method according to claim 1, characterized in that, An equivalent model of the wind turbine gearbox is established based on the theoretical foundation of temperature field and dynamic model.

3. The method according to claim 1, characterized in that, An equivalent model of a wind turbine generator is established based on the theoretical foundations of circuits, magnetic circuits, thermal circuits, and mechanical forces.

4. The method according to claim 1, characterized in that, The established data processing model preprocesses and analyzes the wind turbine operating data stored in the SCADA and CMS systems to obtain characteristic data of the wind turbine generator and gearbox, including: The system performs tagging, calibration and classification of operating conditions, replacement of abnormal data, selection of feature data and data standardization for the wind turbine operating data stored in the SCADA system and CMS system. The standardized data is analyzed to form characteristic data of electrical parameters, control signals, system status, and temperature distribution of components from the data in the SCADA system, and characteristic data of component system performance, friction and vibration, mechanical defects, and fatigue damage from the data in the CMS system.

5. The method according to claim 1, characterized in that, The process of constructing the fault diagnosis model includes: Simulation data and feature data corresponding to the historical operating status data of wind turbine generators and gearboxes are obtained and mixed to obtain hybrid data; Clustering the mixed data yields operational failure data; The operational failure data is further clustered to obtain the failure type.

6. The method according to claim 5, characterized in that, Clustering was performed on the mixed data and operational failure data based on the Pearson correlation coefficient.

7. The method according to claim 1, characterized in that, The process of determining whether there is a fault and the type of fault based on the Pearson correlation coefficient includes: If the Pearson correlation coefficient between the current operating status data and the normal operating status data is greater than the preset threshold, it is judged to be operating normally; If the Pearson correlation coefficient between the current operating status data and the normal operating status data is less than a preset threshold, the operating fault type is determined according to the preset threshold range corresponding to its fault type.

8. A device for diagnosing and predicting operational status faults in a wind turbine generator and gearbox, characterized in that, The device includes: The multi-field coupling model construction module is used to establish a multi-field coupling model based on the physical laws, energy transfer and mutual coupling relationships between multiple physical fields involved in the actual operation of the equipment. The multi-field coupling model includes: an equivalent model of the wind turbine generator and an equivalent model of the wind turbine gearbox. The data processing model building module is used to establish a data processing model to preprocess and analyze the wind turbine operation data stored in the SCADA system and CMS system, and obtain the characteristic data of the wind turbine generator and gearbox. The multi-field coupling model verification module is used to verify the accuracy of the simulation results of the multi-field coupling model using the feature data provided by the data processing model; The real-time operation data acquisition and processing module is used to acquire real-time operation status data of wind turbine generators and gearboxes, and use the corresponding simulation data obtained by the multi-field coupling model after accuracy verification, and extract the corresponding feature data through the data processing model. The simulation data and feature data corresponding to the real-time operation status data of wind turbine generators and gearboxes are mixed to obtain mixed data. The fault diagnosis and prediction module is used to input the mixed data into the fault diagnosis model and the fault prediction model respectively to obtain the corresponding real-time operating status fault diagnosis results and the fault prediction results within a preset time period. The fault diagnosis model obtains fault diagnosis results based on time series similarity metrics, and the process includes: Based on time series Motif features, and taking the normal operating condition data of the equipment as a benchmark, the Pearson correlation coefficient between the current operating condition data and the normal operating condition data is obtained. Determine whether there is a fault and the type of fault based on the Pearson correlation coefficient between the current operating status data and the normal operating condition data. The fault prediction model is obtained by training a temporal convolutional neural network. The process of training the temporal convolutional neural network to obtain the fault prediction model includes: The simulation data and feature data corresponding to the operating status data of the wind turbine and gearbox in the preset historical period are obtained, and the mixed data after mixing is divided into training set and validation set according to the preset ratio. The mixed data of a preset number of historical days in the training set are input into the temporal convolutional neural network to obtain the operating status of the wind turbine and gearbox on a future day. The predicted operating status of the wind turbine and gearbox is added to the historical data. The next step is to update the mixed data of the preset number of historical days and input it into the temporal convolutional neural network to obtain the substation operating status on a future day. This process is repeated to obtain the predicted data of the operating status of the wind turbine and gearbox for multiple future days. The trained model is validated using a validation set, and the training parameters are adjusted appropriately based on the validation results to optimize the model. The model with the best validation performance is selected as the fault prediction model.

9. The apparatus according to claim 8, characterized in that, The multi-field coupling model construction module includes: The gearbox equivalent model building unit is used to establish the equivalent model of the wind turbine gearbox based on the theoretical basis of temperature field and dynamic model. The wind turbine generator equivalent model construction unit is used to establish an equivalent model of the wind turbine generator based on the theoretical foundations of circuit, magnetic circuit, thermal circuit, and mechanical force.

10. The apparatus according to claim 7, characterized in that, The data processing model construction module includes: The preprocessing unit is used to set tags, calibrate and classify operating conditions, replace abnormal data, select feature data, and standardize data for the wind turbine operating data stored in the SCADA system and CMS system. The analysis unit is used to analyze standardized data, transforming data from the SCADA system into characteristic data of electrical parameters, control signals, system status, and temperature distribution of components, and transforming data from the CMS system into characteristic data of component system performance, friction and vibration, mechanical defects, and fatigue damage.

11. A computer device, characterized in that, include: A memory and a processor are interconnected, the memory stores computer instructions, and the processor executes the computer instructions to perform the method for diagnosing and predicting the operating status of a wind turbine generator and gearbox as described in any one of claims 1 to 7.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the method for diagnosing and predicting the operating status faults of the wind turbine generator and gearbox as described in any one of claims 1 to 7.

Images