A safety protection and monitoring method and system for a wind turbine transmission chain

Abstract

The application relates to the technical field of computer wireless communication, and discloses a safety protection and monitoring method and system for a wind turbine transmission chain, which comprises the following steps: collecting wind turbine operation data; establishing a dynamic model of a wind turbine transmission system; extracting fault features through a signal processing method, and constructing an early warning model of system stability deterioration trend; and formulating a maintenance strategy based on early warning results to carry out equipment predictive maintenance. The application significantly improves the operation reliability of the wind turbine, reduces the failure rate and maintenance cost, and has important practical application value and broad popularization prospect. Based on the analysis of the predicted values at multiple time points, the maintenance strategy can be dynamically adjusted, preventive maintenance is implemented before the system state deteriorates, the unplanned downtime is effectively reduced, and the overall operation efficiency and service life of the wind turbine are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine maintenance, and in particular to a safety protection and monitoring method and system for a transmission chain of a wind turbine. Background Art

[0002] As a clean, renewable energy source, wind power has seen widespread application and rapid development. As the core equipment for wind power generation, the operational stability and reliability of wind turbines are directly related to power generation efficiency and safety. Among the many components of a wind turbine, the drive train is a key component in converting wind energy into electricity. It includes components such as the main shaft, gearbox, coupling, and bearings. These components operate under complex operating conditions for long periods of time and are susceptible to multiple factors, including fatigue, wear, and vibration, which can lead to failures and even downtime. Therefore, the safety protection and monitoring of wind turbine drive trains are particularly important.

[0003] Existing wind turbine drivetrain monitoring technologies primarily focus on condition monitoring and fault diagnosis. Common methods include vibration analysis, oil analysis, and temperature monitoring. While these technologies can detect abnormal equipment conditions and provide fault warnings to a certain extent, they have certain limitations. First, traditional monitoring methods are mostly passive, identifying faults only after they occur and lack effective predictive capabilities. Second, due to the complex operating environment of wind turbines, traditional methods are susceptible to noise and external interference, resulting in inaccurate monitoring results. Furthermore, most existing technologies rely on a single monitoring parameter, such as vibration or temperature, failing to fully consider the combined effects of multiple factors and making it difficult to fully reflect the health of the drivetrain. Therefore, existing technologies still have significant shortcomings in improving the accuracy, comprehensiveness, and real-time nature of fault prediction. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is that the existing technology still has great deficiencies in improving the accuracy, comprehensiveness and real-time performance of fault prediction.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for protecting and monitoring the transmission chain of a wind turbine generator set, comprising:

[0007] Collect wind turbine operation data;

[0008] Establish a dynamic model of the wind turbine transmission system;

[0009] Extract fault characteristics through signal processing methods and build an early warning model for system stability degradation trends;

[0010] Formulate maintenance strategies based on early warning results to perform predictive maintenance on equipment.

[0011] As a preferred solution of the wind turbine transmission chain safety protection and monitoring method of the present invention, the wind turbine operation data includes selecting sensors and placing them at key positions of the transmission chain, including the gearbox, bearings, and main shaft, to collect data; selecting a suitable sampling frequency to ensure that signals within the fault frequency range can be captured;

[0012] Collect historical data sets from wind turbine transmission systems, including various features and corresponding timestamps.

[0013] As a preferred solution of the wind turbine transmission chain safety protection and monitoring method of the present invention, the dynamic model of the wind turbine transmission system includes mathematical modeling of the main shaft, gearbox, and bearings to obtain the dynamic equations of the main shaft, gearbox, and bearings;

[0014] The dynamic equation of the main axis is expressed as:

[0015]

[0016] Among them, J s Represents the rotational inertia of the main shaft; θ s (t) represents the angular displacement of the main axis; and denote angular velocity and angular acceleration respectively; c s represents the spindle damping coefficient; k s Indicates the spindle stiffness; T w (t) represents the input torque applied by the wind wheel on the main shaft; T g (t) represents the torque output from the main shaft to the gearbox;

[0017] The dynamic equation of the gearbox, assuming that the transmission ratio of the gearbox is r, the dynamic equation of the low-speed shaft of the gearbox is expressed as:

[0018]

[0019] The dynamic equation of the high-speed shaft of the gearbox is expressed as:

[0020]

[0021] Among them, J g1 Indicates the moment of inertia of the low-speed shaft of the gearbox; J g2 Represents the moment of inertia of the high-speed shaft of the gearbox; θ g1 (t) represents the angular displacement of the low-speed shaft; θ g2 (t) represents the angular displacement of the high-speed shaft; and represents angular velocity; cg1 and c g2 represents the damping coefficient; k g1 and k g2 represents the stiffness coefficient; T g1 (t) represents the torque of the low-speed shaft; T g2 (t) represents the torque of the high-speed shaft; T gen (t) represents the torque output from the gearbox to the generator;

[0022] The dynamic equation of the generator is:

[0023]

[0024] Among them, J gen represents the inertia moment of the generator; θ gen (t) represents the angular displacement of the generator rotor; c gen represents the damping coefficient of the generator; T elec (t) represents the electromagnetic torque generated by the generator.

[0025] As a preferred embodiment of the wind turbine transmission chain safety protection and monitoring method of the present invention

[0026] The signal processing method includes performing denoising on the collected data and performing denoising on the signal using wavelet transform; assuming that the original signal is x(t) and the signal after wavelet transform is W x (a, b), the formula is:

[0027]

[0028] Where a and b represent the scale and translation parameters respectively, and ψ represents the mother wavelet function; perform empirical mode decomposition denoising to decompose the signal x(t) into several intrinsic mode functions c i (t), reconstruct the signal after removing high-frequency noise:

[0029]

[0030] Among them, r n (t) represents the residual term;

[0031] Frequency domain feature extraction, Fourier transform converts the signal from time domain to frequency domain, and the spectrum calculation formula is expressed as:

[0032]

[0033] Where X(f) represents the spectrum, f represents the frequency; time-frequency domain feature extraction, characteristic frequency extraction: extract key frequency components, gear meshing frequency f g , bearing failure frequency f b; Use short-time Fourier transform to analyze the time-frequency characteristics of the signal, use the moving window h(t) to split it in time, and get the short-time spectrum X(t,f) formula as follows:

[0034]

[0035] The time-frequency information of the signal is obtained through wavelet transform.

[0036] As a preferred solution of the wind turbine transmission chain safety protection and monitoring method of the present invention, the fault feature extraction includes using the linear discriminant analysis intra-class scatter matrix S W and the between-class scatter matrix

[0037] Intra-class scatter matrix S W , the formula is:

[0038]

[0039] Where C represents the total number of categories; χ c represents the sample set in category c; x i represents the i-th sample in category c; μ c Represents the mean vector of category c:

[0040]

[0041] Among them, N c represents the number of samples in category c; (x i -μ c ) represents the sample x i Deviation from its class mean; inter-class scatter matrix S B The formula is:

[0042]

[0043] Where μ represents the population mean vector of all samples:

[0044]

[0045] Where N represents the total number of samples; (μ c -μ) represents the deviation between the mean vector of category c and the overall mean vector;

[0046] The formula of the LDA transformation matrix W is expressed as:

[0047]

[0048] Calculate the intra-class scatter matrix S W and the inter-class scatter matrix S B, by solving the generalized eigenvalue problem, the formula is expressed as:

[0049]

[0050] Where Λ represents the eigenvalue diagonal matrix, and W represents the corresponding eigenvector matrix; the eigenvector corresponding to the largest eigenvalue is selected to form the matrix W as the optimal projection direction of LDA.

[0051] As a preferred solution of the wind turbine transmission chain safety protection and monitoring method of the present invention, the early warning model of the system stability degradation trend includes: setting the target to predict the output value y through the regression model i ; The regression model formula is expressed as:

[0052] f(x)=w·φ(x)+b

[0053] Among them, w represents the weight vector, φ(x) represents the feature mapping function, which maps the input feature x to a high-dimensional space, and b represents the bias term;

[0054] Select a loss function so that when the error ∈ is within the allowable range, the error is ignored and the model is as smooth as possible. The loss function formula is expressed as:

[0055]

[0056] Among them, ||w|| 2 represents the regularization term; C represents the regularization parameter, which is used to balance the model complexity and training error; ξ i , represents the slack variable, which is used to deal with intolerable deviations;

[0057] For each training sample (x i ,y i )The constraints satisfied are expressed as:

[0058] y i -(w·φ(x i )+b)≤∈+ξ i

[0059]

[0060] Among them, ∈ represents the tolerance of the model; let A=∈+ξ i -y i +w·φ(x i )+b,

[0061] Introducing the Lagrange multiplier α i and The optimization problem is transformed into a dual problem, and the formula is expressed as:

[0062]

[0063] By solving the dual problem, we get the optimal values ​​of w and b, and introduce the kernel function to deal with the nonlinear feature mapping problem. i ,x j ) is defined as:

[0064] K(x i ,x j )=φ(x i )·φ(x j )

[0065] Using training data Perform model training and obtain the optimal w and b, as well as the support vector, by optimizing the dual problem.

[0066] As a preferred embodiment of the wind turbine transmission chain safety protection and monitoring method of the present invention, the predictive maintenance of the equipment includes using a radial basis function kernel for improvement and selecting the optimal γ parameter value through cross-validation; γ determines the size of the influence range in high-dimensional space, and the improved kernel function formula is expressed as:

[0067] K(x i ,x j )=exp(-γ||x i -x j || 2 )

[0068] The trained support vector machine model can be used to predict the output corresponding to the new input feature x The formula is:

[0069]

[0070] in, Represents the predicted output corresponding to time t; α i and represents the Lagrange multiplier obtained through training;

[0071] The degradation trend of the wind turbine transmission system is analyzed based on the predicted values ​​y(t) at multiple time points. If the predicted y(t) increases over time, it means that the system status is deteriorating, and maintenance should be performed in advance based on the trend.

[0072] A wind turbine transmission chain safety protection and monitoring system using any of the methods described in the present invention, wherein:

[0073] The data acquisition module is responsible for collecting various data during the operation of the wind turbine, including sensor data, operating parameters, and environmental conditions;

[0074] The dynamic modeling module builds a dynamic model of the wind turbine transmission system based on the collected operating data to describe the dynamic behavior and response of the system;

[0075] The signal processing and feature extraction module pre-processes the collected signals and extracts system fault features, providing a basis for building a degradation trend model;

[0076] The early warning and maintenance strategy formulation module builds an early warning model of degradation trends based on the extracted fault characteristics, and formulates a predictive maintenance strategy for the equipment based on the early warning results.

[0077] A computer device comprises: a memory and a processor; the memory stores a computer program, comprising: the steps of implementing any one of the methods of the present invention when the processor executes the computer program.

[0078] A computer-readable storage medium stores a computer program thereon, comprising: steps of implementing any one of the methods of the present invention when the computer program is executed by a processor.

[0079] The present invention has the following beneficial effects: Through systematic design and scientific analysis, it significantly improves the operational reliability of wind turbines, reduces failure rates and maintenance costs, and possesses significant practical application value and broad prospects for promotion. Based on multi-point prediction analysis, the present invention dynamically adjusts maintenance strategies and implements preventive maintenance before system conditions deteriorate, effectively reducing unplanned downtime and improving the overall operational efficiency and service life of wind turbines. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0081] Figure 1 This is an overall flow chart of a safety protection and monitoring method for a wind turbine transmission chain provided by the first embodiment of the present invention. DETAILED DESCRIPTION

[0082] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0083] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a safety protection and monitoring method for a wind turbine transmission chain, comprising:

[0084] S1: Collect wind turbine operation data.

[0085] Furthermore, the wind turbine operation data includes selecting sensors and placing them at key positions in the transmission chain, including the gearbox, bearings, and main shaft, for data collection; selecting a suitable sampling frequency to ensure that signals within the fault frequency range can be captured.

[0086] Furthermore, historical data sets of wind turbine transmission systems are collected, including various features and corresponding timestamps.

[0087] It should be noted that the collection of historical data sets provides the foundation for long-term system monitoring and trend analysis. By collecting historical data and combining it with timestamp information, the system can perform more accurate fault prediction and condition assessment. This step not only enhances the system's foresight but also provides important data support for subsequent signal processing, feature extraction, and trend analysis. By comparing historical data with current operating data, the system can identify potential abnormal trends and take preventative measures before problems develop into serious failures.

[0088] S2: Establish a dynamic model of the wind turbine transmission system.

[0089] Furthermore, the dynamic model of the wind turbine transmission system includes mathematical modeling of the main shaft, gearbox, and bearings to obtain the dynamic equations of the main shaft, gearbox, and bearings.

[0090] Furthermore, the dynamic equation of the main axis is expressed as:

[0091]

[0092] Among them, J s Represents the rotational inertia of the main shaft; θ s (t) represents the angular displacement of the main axis; and denote angular velocity and angular acceleration respectively; c s represents the spindle damping coefficient; k sIndicates the spindle stiffness; T w (t) represents the input torque applied by the wind wheel on the main shaft; T g (t) represents the torque output from the main shaft to the gearbox.

[0093] Furthermore, the dynamic equation of the gearbox, assuming that the transmission ratio of the gearbox is r, the dynamic equation of the low-speed shaft of the gearbox is expressed as:

[0094]

[0095] Furthermore, the dynamic equation of the high-speed shaft of the gearbox is expressed as:

[0096]

[0097] Among them, J g1 Indicates the moment of inertia of the low-speed shaft of the gearbox; J g2 Represents the moment of inertia of the high-speed shaft of the gearbox; θ g1 (t) represents the angular displacement of the low-speed shaft; θ g2 (t) represents the angular displacement of the high-speed shaft; and represents angular velocity; c g1 and c g2 represents the damping coefficient; k g1 and k g2 represents the stiffness coefficient; T g1 (t) represents the torque of the low-speed shaft; T g2 (t) represents the torque of the high-speed shaft; T gen (t) represents the torque output from the gearbox to the generator.

[0098] Furthermore, the dynamic equation of the generator is:

[0099]

[0100] Among them, J gen represents the inertia moment of the generator; θ gen (t) represents the angular displacement of the generator rotor; c gen represents the damping coefficient of the generator; T elec (t) represents the electromagnetic torque generated by the generator.

[0101] Furthermore, in order to obtain the response of the system under different working conditions, a numerical solution method is used for calculation, and the formula is expressed as:

[0102] k1=f(t n ,θ n )

[0103]

[0104] k4=f(t n +h,θ n +h·k3)

[0105]

[0106] Among them, t n represents the current time point; h represents the time step, θ n represents the state variable; f represents the derivative function of the dynamic system.

[0107] It should be noted that dynamic solution of the system using numerical methods (such as the Euler method) can simulate the system's response in the time dimension. This method can handle complex nonlinear dynamic equations and derive the system's state variable x(t) at different time steps Δt, thereby enabling precise control and real-time monitoring of the wind turbine drive system. In practical applications, this improvement can significantly improve the system's response speed and prediction accuracy, further optimizing the overall performance of the wind turbine.

[0108] S3: Extract fault characteristics through signal processing methods and build an early warning model for system stability degradation trends.

[0109] Furthermore, the signal processing method includes performing denoising on the collected data and performing denoising on the signal using wavelet transform; assuming that the original signal is x(t) and the signal after wavelet transform is W x (a, b), the formula is:

[0110]

[0111] Where a and b represent the scale and translation parameters respectively, and ψ represents the mother wavelet function; perform empirical mode decomposition denoising to decompose the signal x(t) into several intrinsic mode functions c i (t), reconstruct the signal after removing high-frequency noise:

[0112]

[0113] Among them, r n (t) represents the residual term.

[0114] Furthermore, frequency domain feature extraction, Fourier transform converts the signal from time domain to frequency domain, and the spectrum calculation formula is expressed as:

[0115]

[0116] Where X(f) represents the spectrum, f represents the frequency; time-frequency domain feature extraction, characteristic frequency extraction: extract key frequency components, gear meshing frequency f g , bearing failure frequency f b; Use short-time Fourier transform to analyze the time-frequency characteristics of the signal, use the moving window h(t) to split it in time, and get the short-time spectrum X(t,f) formula as follows:

[0117]

[0118] Furthermore, the time-frequency information of the signal is obtained through wavelet transform.

[0119] Furthermore, the fault feature extraction includes using the linear discriminant analysis intra-class scatter matrix S W and the between-class scatter matrix

[0120] Furthermore, the intra-class scatter matrix S W , the formula is:

[0121]

[0122] Where C represents the total number of categories; x c represents the sample set in category c; x i represents the i-th sample in category c; μ c Represents the mean vector of category c:

[0123]

[0124] Among them, N c represents the number of samples in category c; (x i -μ c ) represents the sample x i Deviation from its class mean; inter-class scatter matrix S B The formula is:

[0125]

[0126] Where μ represents the population mean vector of all samples:

[0127]

[0128] Where N represents the total number of samples; (μ c -μ) represents the deviation between the mean vector of category c and the overall mean vector.

[0129] Furthermore, the formula of the LDA transformation matrix W is expressed as:

[0130]

[0131] Furthermore, the intra-class scatter matrix S is calculated W and the inter-class scatter matrix S B, by solving the generalized eigenvalue problem, the formula is expressed as:

[0132]

[0133] Where Λ represents the eigenvalue diagonal matrix, and W represents the corresponding eigenvector matrix; the eigenvector corresponding to the largest eigenvalue is selected to form the matrix W as the optimal projection direction of LDA.

[0134] It should be noted that in LDA, the calculation of the intra-class and inter-class scatter matrices ensures maximum separation between features from different categories. By selecting the eigenvector with the largest eigenvalue for projection, the separability of fault features is significantly improved, allowing different fault modes to be clearly distinguished in low-dimensional space. This feature projection method addresses the issues of information redundancy and classification difficulties in high-dimensional data, improving the accuracy and efficiency of the classifier.

[0135] S4: Formulate maintenance strategies based on early warning results to perform predictive maintenance on equipment.

[0136] Furthermore, the early warning model of the system stability degradation trend includes setting the goal to predict the output value y through the regression model i ; The regression model formula is expressed as:

[0137] f(x)=w·φ(x)+b

[0138] Among them, w represents the weight vector, φ(x) represents the feature mapping function, which maps the input feature x to a high-dimensional space, and b represents the bias term.

[0139] Select a loss function so that when the error ∈ is within the allowable range, the error is ignored and the model is as smooth as possible. The loss function formula is expressed as:

[0140]

[0141] Among them, ||w|| 2 represents the regularization term; C represents the regularization parameter, which is used to balance the model complexity and training error; ξ i , Represents the slack variable, which is used to deal with intolerable deviations.

[0142] Furthermore, for each training sample (x i ,y i )The constraints satisfied are expressed as:

[0143] y i -(w·φ(x i )+b)≤∈+ξ i

[0144]

[0145] Among them, ∈ represents the tolerance of the model; let A=∈+ξ i -y i +w·φ(x i )+b,

[0146] Furthermore, the Lagrange multiplier α is introduced i and The optimization problem is transformed into a dual problem, and the formula is expressed as:

[0147]

[0148] Furthermore, by solving the dual problem, we can get the optimal values ​​of w and b, and introduce the kernel function to deal with the nonlinear feature mapping problem. i ,x j ) is defined as:

[0149] K(x i ,x j )=φ(x i )·φ(x j )

[0150] Furthermore, using training data Perform model training and obtain the optimal w, b and support vector by optimizing the dual problem.

[0151] Furthermore, predictive maintenance of equipment involves using a radial basis function kernel for improvement and selecting the optimal γ parameter value through cross-validation; γ determines the size of the influence range in high-dimensional space. The improved kernel function formula is expressed as:

[0152] K(x i ,x j )=exp(-γ||x i -x j || 2 )

[0153] Furthermore, the trained support vector machine model can be used to predict the output corresponding to the new input feature x The formula is:

[0154]

[0155] in, Represents the predicted output corresponding to time t; α i and represents the Lagrange multiplier obtained through training.

[0156] Furthermore, by observing the forecast trend over a period of time, if Continues to increase and exceeds a predetermined threshold y th , it can trigger the early warning mechanism and perform maintenance in advance:

[0157]

[0158] Among them, y th It represents the degradation index threshold set according to historical data or expert experience; [t0, t1] represents the observation time interval.

[0159] Furthermore, the degradation trend of the wind turbine transmission system is analyzed based on the predicted values ​​y(t) at multiple time points. If the predicted y(t) increases over time, it means that the system status is deteriorating, and maintenance can be performed in advance based on the trend.

[0160] It should be noted that by introducing Lagrange multipliers and converting them into dual problems, the difficulty of solving nonlinear optimization problems is greatly simplified. The use of kernel functions enables the model to handle complex nonlinear features. By performing feature mapping in high-dimensional space, the model can more accurately capture the changing trend of the system state. The radial basis function kernel (RBF kernel) is particularly suitable for state prediction of complex systems such as wind turbines because it can accurately perform classification and regression in nonlinear feature space. The improvement of this method significantly improves the flexibility and adaptability of the model, and solves the technical obstacle that traditional linear models cannot handle complex nonlinear problems.

[0161] On the other hand, this embodiment also provides a safety protection and monitoring system for a wind turbine transmission chain, which includes:

[0162] The data acquisition module is responsible for collecting various data during the operation of the wind turbine, including sensor data, operating parameters, and environmental conditions.

[0163] The dynamic modeling module establishes a dynamic model of the wind turbine transmission system based on the collected operating data to describe the dynamic behavior and response of the system.

[0164] The signal processing and feature extraction module pre-processes the collected signals and extracts system fault features, providing a basis for building a degradation trend model.

[0165] The early warning and maintenance strategy formulation module builds an early warning model of degradation trends based on the extracted fault characteristics, and formulates a predictive maintenance strategy for the equipment based on the early warning results.

[0166] If the above 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 and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0167] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0168] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be 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 processing in another suitable manner as necessary, and then stored in a computer memory.

[0169] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0170] Example 2: The following is an embodiment of the present invention, which provides a safety protection and monitoring method for a wind turbine transmission chain. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0171] In this example, the purpose is to verify the effectiveness of a wind turbine transmission chain safety protection and monitoring method. The focus is on the ability to achieve predictive maintenance by collecting wind turbine operating data, establishing a dynamic model of the transmission system, extracting fault characteristics, and constructing a system stability degradation trend warning model. For the purpose of the experiment, three wind turbines (labeled A, B, and C) in a wind farm were selected as test objects. These turbines have different degrees of transmission chain failure risks in actual operation. The key steps of the experiment are as follows:

[0172] First, the test team installed high-precision vibration sensors, temperature sensors, and pressure sensors in key parts of the wind turbine (including the gearbox, main shaft, bearings, etc.) to ensure that operating data closely related to the status of the transmission system could be collected in real time. The sensor sampling frequency was set to 10kHz to ensure that signals could be accurately captured within the frequency range of various faults. In addition, based on historical operating data and combined with data collected by the sensors, a dynamic model of the transmission system was constructed. This model includes the dynamic equations of the main shaft, gearbox, and generator to simulate the system response under different operating conditions.

[0173] Secondly, the collected data is denoised using wavelet transform to remove noise interference from the operating environment and retain the key features of the fault signal. Furthermore, frequency domain features are extracted using Fourier transform and empirical mode decomposition (EMD) techniques to further identify key fault characteristics such as gear meshing frequency and bearing fault frequency. Furthermore, linear discriminant analysis (LDA) is used to classify the extracted features to construct an early warning model for the stability degradation trend of the wind turbine transmission system.

[0174] Finally, a corresponding maintenance strategy is developed based on the output of the early warning model. Specifically, when the system predicts that fault characteristics continue to intensify and exceed a set threshold, it automatically triggers an early warning and generates a detailed maintenance plan, including the areas to be inspected, the recommended maintenance time, and the required spare parts. This test lasted six months, with key parameters of each wind turbine recorded and analyzed monthly. Some of the test data is shown in Table 1.

[0175]

[0176]

[0177] Analysis of the test data in Table 1 reveals that all three wind turbines exhibited some degree of degradation during the test period, based on the trends in key parameters such as vibration frequency, temperature, gear meshing frequency, and bearing failure frequency. In particular, in the third month, key parameters for turbines A, B, and C reached elevated levels. This phenomenon indicates that with continued operation of the wind turbines, wear and fatigue in the drivetrain gradually intensified, leading to a decrease in system stability.

[0178] Secondly, the evolution of the early warning index also demonstrates the inventiveness of this invention. For Unit A, for example, the early warning index gradually increased from 20% in the first month to 50% in the third month, ultimately triggering a maintenance plan. This early warning mechanism successfully identified potential risks before a failure occurred, averting a potentially serious malfunction. Compared to traditional scheduled maintenance, this invention, through real-time monitoring and analysis, more accurately determines the actual operating status of the equipment, avoiding unnecessary downtime and maintenance operations, thereby improving equipment utilization.

[0179] Furthermore, a comparison of the data from units A, B, and C demonstrates that the proposed model is adaptable to the individual differences between wind turbines. Even under similar operating conditions, the three units exhibited varying warning indices and maintenance trigger times, demonstrating the proposed model's adaptability to complex industrial environments. This personalized predictive maintenance strategy significantly improves the accuracy of maintenance work and reduces the resource waste and equipment damage risks associated with over- or delayed maintenance, which are common with traditional technologies.

[0180] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. 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 wind turbine transmission chain safety protection and monitoring method, characterized in that: include: Collect wind turbine operation data; Establish a dynamic model of the wind turbine transmission system; Extract fault characteristics through signal processing methods and build an early warning model for system stability degradation trends; Formulate maintenance strategies based on early warning results to perform predictive maintenance on equipment; The wind turbine operation data includes selecting sensors and placing them at key locations in the transmission chain, including the gearbox, bearings, and main shaft, to collect data; selecting an appropriate sampling frequency to ensure that signals within the fault frequency range can be captured; Collect historical data sets from wind turbine drive systems, including various features and corresponding timestamps; The dynamic model of the wind turbine transmission system includes mathematical modeling of the main shaft, gear box, and bearings to obtain dynamic equations of the main shaft, gear box, and bearings; The dynamic equation of the main axis is expressed as: Among them, J s Represents the rotational inertia of the main shaft; θ s (t) represents the angular displacement of the main axis; and denote angular velocity and angular acceleration respectively; c s represents the spindle damping coefficient; k s Indicates the spindle stiffness; T w (t) represents the input torque applied by the wind wheel on the main shaft; T g (t) represents the torque output from the main shaft to the gearbox; The dynamic equation of the gearbox, assuming that the transmission ratio of the gearbox is r, the dynamic equation of the low-speed shaft of the gearbox is expressed as: The dynamic equation of the high-speed shaft of the gearbox is expressed as: Among them, J g1 Indicates the moment of inertia of the low-speed shaft of the gearbox; J g2 Represents the moment of inertia of the high-speed shaft of the gearbox; θ g1 (t) represents the angular displacement of the low-speed shaft; θ g2 (t) represents the angular displacement of the high-speed shaft; and represents angular velocity; c g1 and c g2 represents the damping coefficient; k g1 and k g2 represents the stiffness coefficient; T g1 (t) represents the torque of the low-speed shaft; T g2 (t) represents the torque of the high-speed shaft; T gen (t) represents the torque output from the gearbox to the generator; The dynamic equation of the generator is: Among them, J gen represents the inertia moment of the generator; θ gen (t) represents the angular displacement of the generator rotor; c gen represents the damping coefficient of the generator; T elec (t) represents the electromagnetic torque generated by the generator.

2. The wind turbine transmission chain safety protection and monitoring method according to claim 1, characterized in that: The signal processing method includes performing denoising on the collected data and performing denoising on the signal using wavelet transform; assuming that the original signal is x(t) and the signal after wavelet transform is W x (a, b), the formula is: Where a and b represent the scale and translation parameters respectively, and ψ represents the mother wavelet function; perform empirical mode decomposition denoising to decompose the signal x(t) into several intrinsic mode functions c i (t), reconstruct the signal after removing high-frequency noise: Among them, r n (t) represents the residual term; Frequency domain feature extraction, Fourier transform converts the signal from time domain to frequency domain, and the spectrum calculation formula is expressed as: Where X(f) represents the spectrum, f represents the frequency; time-frequency domain feature extraction, characteristic frequency extraction: extract key frequency components, gear meshing frequency f g , bearing failure frequency f b ; Use short-time Fourier transform to analyze the time-frequency characteristics of the signal, use the moving window h(t) to split it in time, and get the short-time spectrum X(t,f) formula as follows: The time-frequency information of the signal is obtained through wavelet transform.

3. The wind turbine transmission chain safety protection and monitoring method according to claim 2, characterized in that: The extraction of fault features includes using the linear discriminant analysis intra-class scatter matrix S W and the between-class scatter matrix Intra-class scatter matrix S W , the formula is: Where C represents the total number of categories; represents the sample set in category c; x i represents the i-th sample in category c; μ c Represents the mean vector of category c: Among them, N c represents the number of samples in category c; (x i -μ c ) represents the sample x i Deviation from its class mean; inter-class scatter matrix S B The formula is: Where μ represents the population mean vector of all samples: Where N represents the total number of samples; (μ c -μ) represents the deviation between the mean vector of category c and the overall mean vector; The formula of the LDA transformation matrix W is expressed as: Calculate the intra-class scatter matrix S W and the inter-class scatter matrix S B , by solving the generalized eigenvalue problem, the formula is expressed as: Where Λ represents the eigenvalue diagonal matrix, and W represents the corresponding eigenvector matrix; the eigenvector corresponding to the largest eigenvalue is selected to form the matrix W as the optimal projection direction of LDA.

4. The wind turbine transmission chain safety protection and monitoring method according to claim 3, characterized in that: The early warning model of the system stability degradation trend includes setting the target to predict the output value y through the regression model i ; The regression model formula is expressed as: f(x)=w·φ(x)·+b Among them, w represents the weight vector, φ(x) represents the feature mapping function, which maps the input feature x to a high-dimensional space, and b represents the bias term; Select a loss function so that when the error ∈ is within the allowable range, the error is ignored and the model is as smooth as possible. The loss function formula is expressed as: Among them, ||w|| 2 represents the regularization term; C represents the regularization parameter, which is used to balance the model complexity and training error; ξ i , represents the slack variable, which is used to deal with intolerable deviations; For each training sample (x i ,y i )The constraints satisfied are expressed as: y i -(w·φ(x i )+b)≤∈+ξ i Among them, ∈ represents the tolerance of the model; let A=∈+ξ i -y i +w·φ(x i )+b, Introducing the Lagrange multiplier α i and The optimization problem is transformed into a dual problem, and the formula is expressed as: By solving the dual problem to obtain the optimal values ​​of w and b, the kernel function is introduced to deal with the nonlinear feature mapping problem, K(x i ,x j ) is defined as: K(x i ,x j )=φ(x i )·φ(x j ) Using training data Perform model training and obtain the optimal w and b, as well as the support vector, by optimizing the dual problem.

5. The wind turbine transmission chain safety protection and monitoring method according to claim 4, characterized in that: The equipment predictive maintenance includes using a radial basis function kernel for improvement and selecting an optimal γ parameter value through cross-validation; γ determines the size of the influence range in high-dimensional space, and the improved kernel function formula is expressed as: K(x i ,x j )=exp(-γ||x i -x j || 2 ) The trained support vector machine model can be used to predict the output corresponding to the new input feature x The formula is: in, Represents the predicted output corresponding to time t; α i and represents the Lagrange multiplier obtained through training; The degradation trend of the wind turbine transmission system is analyzed based on the predicted values ​​y(t) at multiple time points. If the predicted y(t) increases over time, it means that the system status is deteriorating, and maintenance should be carried out in advance based on the trend.

6. A wind turbine transmission chain safety protection and monitoring system using the method according to any one of claims 1 to 5, characterized in that: The data acquisition module is responsible for collecting various data during the operation of the wind turbine, including sensor data, operating parameters, and environmental conditions; The dynamic modeling module builds a dynamic model of the wind turbine transmission system based on the collected operating data to describe the dynamic behavior and response of the system; The signal processing and feature extraction module pre-processes the collected signals and extracts system fault features, providing a basis for building a degradation trend model; The early warning and maintenance strategy formulation module builds an early warning model of degradation trends based on the extracted fault characteristics, and formulates a predictive maintenance strategy for the equipment based on the early warning results.

7. 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 wind turbine transmission chain safety protection and monitoring method according to any one of claims 1 to 5 are implemented.

8. 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 a wind turbine transmission chain safety protection and monitoring method according to any one of claims 1 to 5 are implemented.

Images