Wind turbine generator running state analysis method and system based on rotating speed monitoring

Abstract

The invention discloses a wind turbine generator running state analysis method and system based on rotating speed monitoring, and relates to the technical field of wind power generation, and the method comprises the steps: collecting wind turbine generator data and historical data of the wind turbine generator data, and carrying out the preprocessing of the collected wind turbine generator data; performing rotation speed monitoring verification based on fault tree analysis, training a machine learning model in combination with historical data, and performing fault diagnosis by using a rotation speed monitoring verification result; and automatically generating a repair scheme based on the fault diagnosis result, executing the repair scheme, and recovering the normal working state of the wind turbine generator. According to the wind turbine generator running state analysis method based on rotating speed monitoring, data modeling is carried out by using the optimal characteristics of the wind turbine generator, the correlation among the characteristics is captured through the multi-dimensional Gaussian model, the fault probability of each component is calculated, the fault diagnosis precision is improved, and the fault diagnosis efficiency is improved. And the limitation of only depending on a single data source in the prior art is overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of wind power generation, and in particular to a method and system for analyzing the operating status of a wind turbine generator set based on rotation speed monitoring. Background Art

[0002] As the core equipment of wind power plants, the operating status of wind turbines directly affects the operating efficiency and safety of wind power plants. With the rapid development of the wind power industry, the number and scale of wind turbines are constantly expanding, and their failure problems are gradually becoming an important factor limiting the efficiency of wind power development. In order to ensure the reliability and safety of wind turbines in operation, the research and application of wind turbine fault diagnosis technology has become particularly important.

[0003] Existing wind turbine fault diagnosis methods mainly include methods based on vibration signal analysis, monitoring of physical parameters such as oil temperature and oil pressure, and pattern recognition based on artificial intelligence and machine learning. Although these methods can provide certain fault diagnosis functions, they still have some shortcomings, which are mainly reflected in the following aspects: Many existing fault diagnosis technologies mainly rely on real-time monitoring and analysis of data during the operation of wind turbines. However, due to the complex working environment of wind turbines, the variety of fault types, and the fact that the occurrence of faults is often a gradual process, traditional diagnostic methods are difficult to fully consider the mutual influence between the various components of wind turbines, resulting in insufficient accuracy in fault diagnosis.

[0004] Traditional fault diagnosis methods mostly ignore the impact of environmental factors (such as wind speed, temperature, humidity, etc.) and wind turbine operating conditions (such as load, power factor, etc.) on wind turbine failures. The environment and operating conditions have a significant impact on wind turbine failures, and ignoring these factors may lead to deviations in diagnostic results.

[0005] Most of the existing wind turbine fault diagnosis technologies lack systematic and comprehensive analysis, especially in multi-component systems. It is difficult to fully evaluate the fault path of wind turbines by simply relying on a single monitoring indicator. There are complex relationships between different fault paths, and these relationships need to be evaluated through more systematic analysis methods in order to identify potential high-risk paths in advance and carry out targeted prevention and repair.

[0006] At present, the application of machine learning methods in wind turbine fault diagnosis is gradually increasing. However, due to the unreasonable data quality and feature selection, the accuracy and generalization ability of the model are often poor, making it difficult to cope with complex wind turbine fault types.

[0007] In response to the above problems, the existing technology urgently needs a multi-dimensional fault diagnosis method that can better combine the operating status of wind turbines, environmental impacts, and the relationships between various components to improve the accuracy of fault diagnosis, timely identify high-risk fault paths, and automatically generate targeted repair plans based on the diagnosis results to reduce the downtime and maintenance costs of wind turbines. Summary of the invention

[0008] In view of the above-mentioned problems, the present invention is proposed.

[0009] Therefore, the technical problem solved by the present invention is that the existing wind turbine fault diagnosis method relies on a single data source and ignores the correlation between data. The occurrence of wind turbine faults is usually related to multiple factors. Traditional fault diagnosis methods are difficult to deal with the mutual influence between various characteristics of wind turbines, resulting in low diagnostic accuracy.

[0010] The connection between the fault diagnosis results and the repair solutions in the prior art is not close enough. Fault diagnosis and repair solutions are usually performed separately, and there is a lack of comprehensive assessment of fault risks, resulting in low efficiency of the fault handling process and inability to achieve real-time and automated repair.

[0011] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for analyzing the operating status of a wind turbine generator set based on speed monitoring, comprising: Collect wind turbine data and historical data of wind turbine data, and pre-process the collected wind turbine data; Speed ​​monitoring verification is performed based on fault tree analysis, and the machine learning model is trained with historical data, and fault diagnosis is performed using the speed monitoring verification results; Automatically generate a repair plan based on the fault diagnosis results, execute the repair plan, and restore the normal working state of the wind turbine; Before speed monitoring verification based on fault tree analysis, the preprocessed data is used to extract wind turbine characteristics; The wind turbine characteristics include load index, power factor, mechanical stress coefficient and environmental impact coefficient; The load index is expressed as, , , , , in, represents the load index, Indicates the actual power output, Indicates the maximum power, represents the power factor, represents the air density, is the atmospheric pressure, is the gas constant, It is the temperature It represents the wind swept area of ​​the wind turbine. is the blade radius of the wind turbine, Indicates wind speed; The power factor is expressed as, , , , in, Indicates the power factor, Indicates the actual power, Indicates voltage, represents the current, represents the phase angle between current and voltage, Indicates apparent power; The mechanical stress coefficient is expressed as, , in, represents the mechanical stress coefficient, Indicates the current speed of the wind turbine. Indicates the maximum speed of the wind turbine. Indicates the vibration speed of the wind turbine. Indicates the maximum allowable vibration speed of the wind turbine; The environmental impact coefficient is expressed as, , in, represents the environmental impact coefficient, Indicates the current wind speed. Indicates the rated wind speed of the wind turbine. Indicates the current ambient temperature.

[0012] As a preferred solution of the wind turbine operating status analysis method based on speed monitoring described in the present invention, wherein: the wind turbine data includes wind turbine operating status data, environmental data and electrical system data; The preprocessing includes denoising, filling missing values ​​and standardizing the data.

[0013] As a preferred solution of the method for analyzing the operating status of a wind turbine based on speed monitoring described in the present invention, the speed monitoring verification includes modeling the characteristics of the wind turbine using a multidimensional Gaussian model, and obtaining the failure probability of each component through the probability density function of the multidimensional Gaussian distribution combined with the covariance matrix.

[0014] As a preferred solution of the wind turbine operating status analysis method based on speed monitoring described in the present invention, wherein: the fault probability is calculated by a multidimensional Gaussian distribution model for each feature combination. Given a feature vector , the probability density function is expressed as, , , in, represents the mean vector, represents the failure probability, Represents the dimension of the feature, Represents the covariance matrix The determinant of Represents the covariance matrix The inverse matrix of represents the deviation of each feature from its mean, represents the transpose operation, represents the exponential function, represents the mean value of the load index, represents the mean value of the power factor, represents the mean value of the mechanical stress coefficient, represents the mean value of environmental impact coefficient; Based on the logic gates in the fault tree model and the failure probability, the risk level of each fault path is evaluated and the potential fault paths are determined.

[0015] As a preferred solution of the wind turbine operating status analysis method based on speed monitoring described in the present invention, wherein: the determination of the potential fault path includes combining the fault probabilities of various components through the logic gates in the fault tree analysis model to obtain the overall fault probability of the wind turbine; When the AND gate is used, the probability of overall failure of the wind turbine is the product of the failure probabilities of each subcomponent; when the OR gate is used, the probability of overall failure of the wind turbine is the weighted sum of the failure probabilities of each subcomponent; when the NOT gate is used, the probability of no failure of the component is calculated; Through the combination of logic gates, the failure probability of each fault path is evaluated and the risk level is classified. If the failure probability is greater than the preset threshold, it is judged as a high-risk path; if the failure probability is less than or equal to the preset threshold, it is judged as a low-risk path.

[0016] As a preferred solution of the method for analyzing the operating status of a wind turbine generator set based on speed monitoring described in the present invention, wherein: the fault diagnosis includes inputting the characteristic data of the high-risk path into the machine learning model for training, the training uses historical fault data and combines the fault probability information of each high-risk path, and the historical fault data is used to learn the fault mode of the wind turbine generator set; during the training process, by evaluating the correlation between each feature and the fault probability, feature selection is performed to remove redundant features; the model is optimized by cross-validation and hyperparameter tuning methods; The machine learning model is trained using a supervised learning method. The model learns the failure modes and path characteristics in historical data. When the model training is completed, the characteristic data of the current high-risk path is input, each path is diagnosed, and the fault type, faulty component and location of the faulty component of each path are output.

[0017] As a preferred solution of the wind turbine operating status analysis method based on speed monitoring described in the present invention, wherein: the execution of the repair plan includes generating a repair plan according to the fault type, faulty component and location of the faulty component output by the machine learning model, and the repair plan includes repair steps, required tools, personnel arrangements and time scheduling, and the repair order and urgency in the repair plan are optimized based on the priority of the faulty component.

[0018] Another object of the present invention is to provide a speed monitoring test system, which can solve the problem that the existing wind turbine operating status analysis method based on speed monitoring usually relies on a large amount of prior knowledge and is difficult to adapt to the dynamically changing wind turbine operating environment. At the same time, the traditional method has certain limitations when processing multi-dimensional data, and it is difficult to accurately evaluate the relationship and fault path between components in a complex system.

[0019] To solve the above technical problems, the present invention provides the following technical solutions: a speed monitoring and testing system, comprising: a data acquisition module, a fault diagnosis module and a fault repair module; the data acquisition module is used to collect wind turbine data and historical data of wind turbine data, and pre-process the collected wind turbine data; the fault diagnosis module is used to perform speed monitoring verification based on fault tree analysis, train a machine learning model in combination with historical data, and perform fault diagnosis using the speed monitoring verification results; the fault repair module is used to automatically generate a repair plan based on the fault diagnosis results, execute the repair plan, and restore the normal working state of the wind turbine.

[0020] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the wind turbine operating status analysis method based on rotation speed monitoring as described above are implemented.

[0021] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the wind turbine operating status analysis method based on rotation speed monitoring as described above.

[0022] The beneficial effects of the present invention are as follows: the wind turbine operating status analysis method based on speed monitoring provided by the present invention utilizes the preferred features of the wind turbine to perform data modeling, captures the correlation between the features through a multidimensional Gaussian model, calculates the failure probability of each component, improves the accuracy of fault diagnosis, and overcomes the limitation of the prior art that only relies on a single data source; through fault tree analysis combined with the failure probability calculated by the multidimensional Gaussian model, it is possible to automatically identify high-risk paths, perform fault diagnosis through a machine learning model, and automatically generate repair plans. The repair plans obtained not only have higher accuracy, but also can optimize the repair sequence and urgency, thereby improving the operation and maintenance efficiency of the wind turbine; combined with the logic gates in the fault tree analysis, it is possible to accurately evaluate the risk level of each component or fault path, obtain high-risk fault paths through reasonable probability calculations, and perform effective sorting. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0024] Figure 1 An overall flow chart of a method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring is provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0026] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0027] Example 1, reference Figure 1, which is an embodiment of the present invention, provides a method for analyzing the operating status of a wind turbine generator set based on speed monitoring, comprising: Collect wind turbine data and historical data of wind turbine data, and pre-process the collected wind turbine data; Speed ​​monitoring verification is performed based on fault tree analysis, and the machine learning model is trained with historical data, and fault diagnosis is performed using the speed monitoring verification results; Automatically generate a repair plan based on the fault diagnosis results, execute the repair plan, and restore the normal working state of the wind turbine.

[0028] The wind turbine data includes wind turbine operation status data, environmental data and electrical system data; The wind turbine operating status data includes wind turbine speed, power output and blade angle; The environmental data include wind speed, wind direction and temperature; The electrical system data includes current, voltage and frequency data; The preprocessing includes denoising, filling missing values ​​and standardizing the data; De-noising removes noise data through low-pass filters, moving averages and other methods to ensure that the input data is smoother and more reliable.

[0029] Missing value filling fills the missing values ​​through interpolation methods (linear interpolation, Lagrange interpolation, etc.) to ensure that each data point is complete.

[0030] Standardization processing standardizes the data, eliminates dimensional differences, and ensures that data of different dimensions have the same weight in the model.

[0031] Before speed monitoring verification based on fault tree analysis, the preprocessed data is used to extract wind turbine characteristics; The wind turbine characteristics include load index, power factor, mechanical stress coefficient and environmental impact coefficient; The load index is expressed as, , , , , in, represents the load index, Indicates the actual power output, Indicates the maximum power, represents the power factor, represents the air density, is the atmospheric pressure, is the gas constant, It is the temperature It represents the wind swept area of ​​the wind turbine. is the blade radius of the wind turbine, Indicates wind speed; The power factor is expressed as, , , , in, Indicates the power factor, Indicates the actual power, Indicates voltage, represents the current, represents the phase angle between current and voltage, Indicates apparent power; The mechanical stress coefficient is expressed as, , in, represents the mechanical stress coefficient, Indicates the current speed of the wind turbine. Indicates the maximum speed of the wind turbine. Indicates the vibration speed of the wind turbine. Indicates the maximum allowable vibration speed of the wind turbine; The environmental impact coefficient is expressed as, , in, represents the environmental impact coefficient, Indicates the current wind speed. Indicates the rated wind speed of the wind turbine. Indicates the current ambient temperature.

[0032] The speed monitoring verification includes modeling the characteristics of the wind turbine using a multi-dimensional Gaussian model, and obtaining the failure probability of each component through the probability density function of the multi-dimensional Gaussian distribution combined with the covariance matrix; The failure probability is calculated through the multi-dimensional Gaussian distribution model for each feature combination. Given the feature vector , the probability density function is expressed as, , , , in, represents the mean vector, represents the failure probability, Indicates the dimension of the feature, which is 4. Represents the covariance matrix The determinant of Represents the covariance matrix The inverse matrix of represents the deviation of each feature from its mean, represents the transpose operation, represents the exponential function, represents the mean value of the load index, represents the mean value of the power factor, represents the mean value of the mechanical stress coefficient, represents the mean value of the environmental impact coefficient, represents the variance of the load index, represents the variance of the power factor, represents the variance of the mechanical stress coefficient, represents the variance of the environmental impact coefficient, is the covariance between different features, reflecting the correlation between them; It should be noted that load index, power factor, mechanical stress coefficient and environmental impact coefficient are four key characteristics that describe the operating status of wind turbines. They can reflect the working health status of wind turbines from different aspects. These four characteristics reveal various aspects of the operating status of wind turbines from different dimensions and can comprehensively and accurately describe the potential risks of wind turbine failures. Therefore, they are preferred as key characteristics for wind turbine fault diagnosis.

[0033] Moreover, these four characteristics are not only independent, but also have complex relationships with each other. For example, an increase in the load index may lead to an increase in mechanical stress, a decrease in the power factor may aggravate the loss of mechanical components, and changes in ambient temperature and wind speed will affect the load and mechanical stress. The interaction between these characteristics shows that considering the failure probability of each characteristic alone is not enough to fully describe the overall failure risk of wind turbines.

[0034] The multidimensional Gaussian model can handle the correlation between multiple variables and comprehensively consider the mutual influence between these features. By constructing a multidimensional Gaussian distribution, we can not only calculate the independent failure probability of each feature, but also reveal the relationship between the features through the covariance matrix, and then accurately estimate the overall failure probability of the wind turbine. In this way, the multidimensional Gaussian model is very suitable for analyzing these features because it can handle the complex dependencies between features and improve the accuracy of fault diagnosis.

[0035] Combining these four features with a multidimensional Gaussian model can significantly improve the accuracy and robustness of fault diagnosis. First, this combination can reveal the complex interactions between features and avoid relying solely on a single feature or a simple weighted sum. For example, although the load index is closely related to the risk of failure, if the power factor of the electrical system and the influence of environmental factors are ignored, diagnostic bias may occur. Through the multidimensional Gaussian model, the method of the present invention can fully consider the interaction of these features and derive a more accurate probability of failure. Secondly, it can automatically discover the implicit correlation between the features without the need to manually explicitly set the relationship between the features. The covariance matrix can capture these potential dependencies, thereby more accurately calculating the probability of failure, avoiding the interference of redundant features, and improving the efficiency of the model. At the same time, the model can be optimized in a high-dimensional space to reduce the impact of irrelevant features on the results.

[0036] Based on the logic gates in the fault tree model and the failure probability, the risk level of each fault path is evaluated and the potential fault paths are determined.

[0037] A potential fault path refers to a fault chain formed by connecting the faults of various components through logical relationships, which may cause the wind turbine system to fail.

[0038] Determining the potential fault path includes combining the fault probabilities of the various components through logic gates in the fault tree analysis model to obtain the overall fault probability of the wind turbine; When the AND gate is used, the probability of overall failure of the wind turbine is the product of the failure probabilities of each subcomponent; when the OR gate is used, the probability of overall failure of the wind turbine is the weighted sum of the failure probabilities of each subcomponent; when the NOT gate is used, the probability of no failure of the component is calculated; Through the combination of logic gates, the failure probability of each fault path is evaluated and the risk level is classified. If the failure probability is greater than the preset threshold, it is judged as a high-risk path; if the failure probability is less than or equal to the preset threshold, it is judged as a low-risk path.

[0039] A path refers to a fault chain in a wind turbine consisting of multiple components and the fault propagation relationships between them (such as components connected by logic gates). Each path consists of a series of components and their failure probabilities, which may be combined through logic gates such as AND and OR to form a more complex path.

[0040] High-risk paths refer to wind turbine failure paths obtained by combining different logic gates (AND gates, OR gates, etc.) in the fault tree model, and their failure probability is greater than the preset threshold, which means that these paths may cause major failure or shutdown of the wind turbine.

[0041] The fault tree model of a wind turbine determines the fault propagation relationship between components through engineering analysis. The failure mode of each component and its relationship with other components are defined according to the wind turbine system structure. These relationships include: which components are critical components, and they must fail at the same time to cause the overall failure of the wind turbine, using the AND gate. Which components are redundant or backup, and any failure of any component will cause the overall failure of the wind turbine, using the OR gate. Which components are redundant in normal operation, and the system can continue to operate as long as there are redundant components, using the NOT gate.

[0042] The AND gate is represented as, , in, represents the total failure probability of the fault path, Represents the components that make up the fault path The probability of failure.

[0043] The OR gate is represented as, , in, represents the total failure probability of the fault path, Represents the components that make up the fault path The probability of failure.

[0044] The NOT gate is represented as, , in, Representation Components The probability of no failure, Representation Components The probability of failure.

[0045] The fault diagnosis includes inputting the characteristic data of the high-risk path into the machine learning model for training, the training uses historical fault data and combines the fault probability information of each high-risk path, and the historical fault data is used to learn the fault mode of the wind turbine; during the training process, by evaluating the correlation between each feature and the fault probability, feature selection is performed to remove redundant features; the model is optimized by cross-validation and hyperparameter tuning methods; The machine learning model is trained using a supervised learning method. The model learns the failure modes and path characteristics in historical data. When the model training is completed, the characteristic data of the current high-risk path is input, each path is diagnosed, and the fault type, faulty component and location of the faulty component of each path are output.

[0046] The execution of the repair plan includes generating a repair plan based on the fault type, faulty component and location of the faulty component output by the machine learning model. The repair plan includes repair steps, required tools, personnel arrangements and time scheduling. The repair order and urgency in the repair plan are optimized based on the priority of the faulty component.

[0047] Embodiment 2, an embodiment of the present invention provides a rotation speed monitoring test system, comprising: Data acquisition module, fault diagnosis module and fault repair module; The data acquisition module is used to collect wind turbine data and historical data of wind turbine data, and pre-process the collected wind turbine data; The fault diagnosis module is used to perform speed monitoring verification based on fault tree analysis, train a machine learning model in combination with historical data, and perform fault diagnosis using the speed monitoring verification results; The fault repair module is used to automatically generate a repair plan based on the fault diagnosis result, execute the repair plan, and restore the normal working state of the wind turbine generator set.

[0048] Embodiment 3, an embodiment of the present invention, is different from the first two embodiments in that: If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0049] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0050] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0051] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0052] Embodiment 4 is an embodiment of the present invention, which provides a method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring. In order to verify the beneficial effects of the present invention, a scientific demonstration is carried out through simulation experiments.

[0053] The experiment uses 10 wind turbines. The operating status and fault data of each wind turbine are collected in real time by the wind turbine monitoring system, including wind speed, rotation speed, voltage, current, ambient temperature, vibration speed, load and other data. Each wind turbine has different types of potential faults, and each fault path is marked by empirical rules and historical data. The experiment is divided into two parts. The first part uses the traditional wind turbine fault diagnosis method, and the second part uses the multifunctional wind turbine operating status analysis method based on rotation speed monitoring of the present invention.

[0054] In the experiment of the traditional method, the collected data of each wind turbine is first preprocessed by standardization and analyzed according to the traditional fault tree model. In the fault tree model, the risk assessment of the fault path is based only on the rules set by experience, and a simple threshold judgment is used to identify the high-risk path. The failure probability of each component of each wind turbine is combined through logic gates to obtain the overall failure probability. For each path, if the failure probability exceeds the preset threshold, it is judged as a high-risk path; if it is lower than the threshold, it is a low-risk path. Finally, based on the high-risk path, the corresponding repair plan is generated. The generation of the repair plan depends on the preset rules, usually based on the fault type and the severity of the component, and adopts a general repair step.

[0055] In the experiment of the method of the present invention, the collected wind turbine data is firstly subjected to denoising, missing value filling and standardization. Afterwards, the four features of the present invention are used to extract features of the wind turbine. Based on these features, a multidimensional Gaussian model is combined to perform modeling, the failure probability of each component is calculated, and the relationship between the features is described by the covariance matrix. The failure probability of each feature combination is obtained through the probability density function of the multidimensional Gaussian distribution, and then the failure probability of the wind turbine as a whole is calculated. Then, the high-risk path is evaluated by the logic gate in the fault tree analysis model. Different from the traditional method, the method of the present invention uses multidimensional features combined with the probability obtained by fault tree analysis to accurately identify potential high-risk paths. Finally, the data based on the high-risk path is input into the machine learning model for training to optimize the repair plan. The repair plan is automatically generated, taking into account the location of the faulty component, the repair steps, the required tools and personnel arrangements, and intelligently adjusting the repair order and urgency according to historical data. The experimental results are shown in Table 1.

[0056] Table 1 Comparison of experimental results , First, in terms of fault diagnosis accuracy, the present invention uses a multidimensional Gaussian model to model multiple characteristics of wind turbines (load index, power factor, mechanical stress coefficient, and environmental impact coefficient). Traditional methods often make fault judgments based on simplified empirical rules or single features, ignoring the complex correlations between the characteristics, and react slowly to changes in the operating environment of wind turbines. In contrast, the present invention considers the covariance relationship between the various characteristics through a multidimensional Gaussian model, which can comprehensively integrate the operating status and environmental impact of the wind turbine, thereby improving the accuracy of fault diagnosis. Specifically, the multidimensional Gaussian distribution calculates the fault probability of the feature combination through a probability density function. This refined modeling method enables the fault type and location of the wind turbine to be more accurately identified, avoiding missed diagnosis or misdiagnosis in traditional methods.

[0057] Secondly, in terms of risk assessment accuracy, the present invention combines a multidimensional Gaussian model with a fault tree analysis, and performs risk assessment after comprehensively calculating the failure probability of each component. Traditional methods generally use simple logic gates (such as AND gates, OR gates, etc.) to combine failure probabilities, but this method cannot accurately consider the correlation between components and the complexity of the failure chain. In contrast, the present invention combines a fault tree model with a multidimensional Gaussian distribution to more accurately assess the risk of each fault path. Specifically, the failure probability calculated based on the covariance matrix between each feature can more realistically reflect the chain reaction of the failure, thereby more accurately determining which components or paths are high-risk and which are low-risk. Through this precise risk assessment, the present invention can effectively identify potential high-risk paths, reducing the problem of the inability to comprehensively assess the risk of fault paths in traditional methods.

[0058] Finally, in terms of optimization of repair solutions, the present invention automatically learns historical fault data through a machine learning model, and combines it with the characteristic data of high-risk paths for training, thereby generating an intelligent repair solution. The generation of repair solutions by traditional methods usually relies on manual experience or simple rules, which is prone to inefficiency and inaccuracy. The present invention uses a machine learning model to optimize the repair sequence and urgency according to the characteristics of the wind turbine fault path and historical data, making the repair solution more accurate and efficient. Specifically, the machine learning model can not only determine the type of fault and the location of the component, but also dynamically adjust the details of the repair solution according to the current operating status of the wind turbine and historical fault data, thereby improving the accuracy and efficiency of the repair and reducing the repair time and cost.

[0059] In summary, the present invention, by adopting multidimensional Gaussian models and machine learning optimization technology, can fully consider the complex interactions between the various components of wind turbines and intelligently generate optimized repair plans. These technical features greatly improve the accuracy of wind turbine fault diagnosis, risk assessment, and repair plan generation, solving the problems of missed diagnosis, misdiagnosis, and inefficient repair plans in traditional methods, thereby providing a more reliable guarantee for the reliability and safety of wind turbines.

[0060] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring, characterized in that: include: Collect wind turbine data and historical data of wind turbine data, and pre-process the collected wind turbine data; Speed ​​monitoring verification is performed based on fault tree analysis, and the machine learning model is trained with historical data, and fault diagnosis is performed using the speed monitoring verification results; Automatically generate a repair plan based on the fault diagnosis results, execute the repair plan, and restore the normal working state of the wind turbine; Before speed monitoring verification based on fault tree analysis, the preprocessed data is used to extract wind turbine characteristics; The wind turbine characteristics include load index, power factor, mechanical stress coefficient and environmental impact coefficient; The load index is expressed as, ， ， ， ， in, represents the load index, Indicates the actual power output, Indicates the maximum power, represents the power factor, represents the air density, is the atmospheric pressure, is the gas constant, It is the temperature It represents the wind swept area of ​​the wind turbine. is the blade radius of the wind turbine, Indicates wind speed; The power factor is expressed as, ， ， ， in, Indicates the power factor, Indicates the actual power, Indicates voltage, represents the current, represents the phase angle between current and voltage, Indicates apparent power; The mechanical stress coefficient is expressed as, ， in, represents the mechanical stress coefficient, Indicates the current speed of the wind turbine. Indicates the maximum speed of the wind turbine. Indicates the vibration speed of the wind turbine. Indicates the maximum allowable vibration speed of the wind turbine; The environmental impact coefficient is expressed as, ， in, represents the environmental impact coefficient, Indicates the current wind speed. Indicates the rated wind speed of the wind turbine. Indicates the current ambient temperature.

2. The method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring according to claim 1, characterized in that: The wind turbine data includes wind turbine operation status data, environmental data and electrical system data; The preprocessing includes denoising, filling missing values ​​and standardizing the data.

3. The method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring according to claim 2, characterized in that: The speed monitoring verification includes using a multi-dimensional Gaussian model to model the characteristics of the wind turbine generator set, and obtaining the failure probability of each component through the probability density function of the multi-dimensional Gaussian distribution combined with the covariance matrix.

4. The method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring according to claim 3, characterized in that: The failure probability is calculated through the multi-dimensional Gaussian distribution model for each feature combination. Given the feature vector , the probability density function is expressed as, ， ， in, represents the mean vector, represents the probability of failure, Represents the dimension of the feature, Represents the covariance matrix The determinant of Represents the covariance matrix The inverse matrix of represents the deviation of each feature from its mean, represents the transpose operation, represents the exponential function, represents the mean value of the load index, represents the mean value of the power factor, represents the mean value of the mechanical stress coefficient, represents the mean value of environmental impact coefficient; Based on the logic gates in the fault tree model and the failure probability, the risk level of each fault path is evaluated and the potential fault paths are determined.

5. The method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring according to claim 4, characterized in that: Determining the potential fault path includes combining the fault probabilities of the various components through logic gates in the fault tree analysis model to obtain the overall fault probability of the wind turbine; When the AND gate is used, the probability of overall failure of the wind turbine is the product of the failure probabilities of each subcomponent; when the OR gate is used, the probability of overall failure of the wind turbine is the weighted sum of the failure probabilities of each subcomponent; when the NOT gate is used, the probability of no failure of the component is calculated; Through the combination of logic gates, the failure probability of each fault path is evaluated and classified into risk levels. If the failure probability is greater than the preset threshold, it is judged as a high-risk path; If the failure probability is less than or equal to the preset threshold, it is judged as a low-risk path.

6. The method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring according to claim 5, characterized in that: The fault diagnosis includes inputting the characteristic data of the high-risk path into the machine learning model for training, the training uses historical fault data and combines the fault probability information of each high-risk path, and the historical fault data is used to learn the fault mode of the wind turbine; during the training process, by evaluating the correlation between each feature and the fault probability, feature selection is performed to remove redundant features; the model is optimized by cross-validation and hyperparameter tuning methods; The machine learning model is trained using a supervised learning method. The model learns the failure modes and path characteristics in historical data. When the model training is completed, the characteristic data of the current high-risk path is input, each path is diagnosed, and the fault type, faulty component and location of the faulty component of each path are output.

7. The method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring according to claim 6, characterized in that: The execution of the repair plan includes generating a repair plan based on the fault type, faulty component and location of the faulty component output by the machine learning model. The repair plan includes repair steps, required tools, personnel arrangements and time scheduling. The repair order and urgency in the repair plan are optimized based on the priority of the faulty component.

8. A system using the wind turbine operating status analysis method based on speed monitoring as described in any one of claims 1 to 7, characterized in that: include: Data acquisition module, fault diagnosis module and fault repair module; The data acquisition module is used to collect wind turbine data and historical data of wind turbine data, and pre-process the collected wind turbine data; The fault diagnosis module is used to perform speed monitoring verification based on fault tree analysis, train a machine learning model in combination with historical data, and perform fault diagnosis using the speed monitoring verification results; The fault repair module is used to automatically generate a repair plan based on the fault diagnosis result, execute the repair plan, and restore the normal working state of the wind turbine generator set.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for analyzing the operating status of a wind turbine set based on rotational speed monitoring according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for analyzing the operating status of a wind turbine generator system based on rotation speed monitoring according to any one of claims 1 to 7 are implemented.

Images