Wind driven generator main bearing fault prediction method, system and equipment and storage medium

Abstract

The invention relates to a wind driven generator main bearing fault prediction method, system and device, and a storage medium. The method comprises the following steps: collecting operation data of a wind driven generator main bearing and environmental data of an environment where a wind driven generator is located according to a time sequence, and preprocessing the data; classifying the operation data of the main bearing of the wind driven generator according to the temperature, and correcting the operation data of each type of main bearing of the wind driven generator; constructing a wind driven generator main bearing state prediction model, inputting the data set into the wind driven generator main bearing state prediction model for training, and predicting vibration, strain and temperature data of the wind driven generator main bearing through the trained wind driven generator main bearing state prediction model; and based on the predicted vibration, strain and temperature data of the main bearing of the wind driven generator, calculating the posterior probability of a fault of the main bearing of the wind driven generator in the future, judging the health state of the main bearing of the wind driven generator according to the posterior probability, and performing early warning.

Description

Technical Field

[0001] The invention relates to a method, system, equipment and storage medium for predicting a main bearing fault of a wind generator, and belongs to the technical field of wind power generation. Background Art

[0002] The main bearing of a wind turbine is an important component of a wind turbine. Its operating environment usually has the following characteristics:

[0003] Large ambient temperature difference: Fans are usually installed at high altitudes or in cold areas. Low temperature environments may cause increased material brittleness. At the same time, in summer or under long-term operating conditions, bearings will face problems of material softening and lubricant deterioration due to high temperatures.

[0004] Complex operating stress: The main bearing of the fan needs to bear multi-axial loads, including radial loads and axial loads generated by the rotation of the fan blades.

[0005] High requirements for long-term stability: The operation time of fans is often as long as decades, and their main bearings need to maintain stable performance for a long time to avoid failures caused by the environment or fatigue.

[0006] Existing fan main bearing monitoring technologies mostly use simple linear models or empirical formulas to correct the impact of temperature changes on bearing operation. This method has the following shortcomings:

[0007] Ignoring the nonlinear characteristics of materials: Under low temperature conditions, the change in the elastic modulus of materials is usually nonlinear, and existing technologies are difficult to accurately reflect the actual impact of low temperature on vibration and strain.

[0008] Inability to fully describe the impact of high temperature: Under high temperature conditions, the thermal expansion and softening of materials and the degradation of lubricants will work together, but existing monitoring schemes usually only consider a single factor among them, which makes it difficult to fully reflect the complex coupling effects in high temperature environments.

[0009] Lack of response mechanism to dynamic temperature changes: The temperature may fluctuate dramatically during fan operation, and existing technologies are unable to dynamically adjust and correct the model, resulting in insufficient prediction accuracy. Summary of the invention

[0010] In order to solve the above problems existing in the prior art, the present invention proposes a method, system, device and storage medium for predicting main bearing faults of a wind turbine.

[0011] The technical solution of the present invention is as follows:

[0012] In one aspect, the present invention provides a method for predicting a fault of a main bearing of a wind turbine generator, comprising the following steps:

[0013] Collect vibration data, strain data, temperature data of the main bearing of the wind turbine and environmental data of the environment in which the wind turbine is located in chronological order and pre-process the data;

[0014] The vibration data and strain data of the main bearing of the wind turbine generator are classified according to the temperature data, and the vibration data and strain data of the main bearing of each type of wind turbine generator are corrected respectively;

[0015] constructing a data set based on the corrected vibration data, the corrected strain data, the temperature data, and the environmental data of the environment in which the wind turbine is located;

[0016] Construct a wind turbine main bearing state prediction model, input the data set into the wind turbine main bearing state prediction model for training, and predict the vibration, strain and temperature data of the wind turbine main bearing through the trained wind turbine main bearing state prediction model;

[0017] Based on the predicted vibration, strain and temperature data of the wind turbine main bearing, the posterior probability of future wind turbine main bearing failure is calculated, and the health status of the wind turbine main bearing is judged according to the posterior probability and an early warning is issued.

[0018] As a preferred embodiment of the present invention, the temperature data is sorted from large to small, the sorted temperature data is divided into three equal parts, and the vibration data and strain data corresponding to the temperature data are classified into high temperature data, normal temperature data and low temperature data according to the quantile, as shown in the following formula:

[0019]

[0020] Where: Q represents the location of the current temperature data.

[0021] As a preferred embodiment of the present invention, the vibration data in the low temperature data is corrected, as shown in the following formula:

[0022]

[0023] Where: V(t) represents the vibration data at time t; V ′ (t) represents the corrected vibration data at time t; α represents the low temperature variation function; T t Represents the temperature data at time t; represents the gradient of temperature data; β represents the attenuation factor of low temperature vibration;

[0024] The strain data in the low temperature data is corrected as shown in the following formula:

[0025]

[0026] Where: ∈(t) represents the strain data at time t; ∈ ′ (t) represents the corrected strain data at time t; γ represents the elastic modulus correction function; σ represents the material stress; κ represents the attenuation factor of low temperature strain.

[0027] As a preferred embodiment of the present invention, the vibration data in the high temperature data is corrected, as shown in the following formula:

[0028]

[0029] Where: δ represents the lubricant degradation sensitivity coefficient; η0 represents the lubricant base viscosity; η(T t ) indicates T t The viscosity of the lubricating oil at the temperature; φ represents the high temperature lubrication degradation correction function; ψ represents the high temperature vibration adjustment coefficient; ΔT represents the temperature change;

[0030] The strain data in the high temperature data is corrected as shown in the following formula:

[0031]

[0032] Where: ω represents the coupling term of temperature and stress; χ represents the correction factor of thermal expansion.

[0033] As a preferred embodiment of the present invention, the corrected vibration data is subjected to joint wavelet transform to obtain a corresponding vibration spectrum sequence, which is specifically shown in the following formula:

[0034]

[0035] Where: a represents the scale parameter; b represents the translation parameter; ψ represents the mother wavelet; W x Represents the modified vibration data after joint wavelet transformation.

[0036] As a preferred embodiment of the present invention, the wind turbine main bearing state prediction model is constructed based on a long short-term memory network model.

[0037] As a preferred embodiment of the present invention, the predicted vibration, strain and temperature data of the main bearing of the wind turbine generator are input into a Bayesian network, and several nodes are defined in the Bayesian network, including a temperature node, a vibration node, a strain node and a main bearing health status node;

[0038] Construct the causal relationship between nodes, set the temperature node, vibration node, and strain node as the parent node of the health state node, and use the prior probability distribution for vibration, strain, and temperature data. For the probability that the health state node state takes a specific value as a fault, the posterior probability that the health state node state is a fault is obtained by updating the following formula:

[0039]

[0040] Where: F = 1 indicates that the healthy node state is faulty; T indicates the predicted temperature node data; V indicates the vibration node data; ∈ indicates the strain node data; P(F = 1|T, V, ∈) indicates the posterior probability that the healthy node state is faulty; P(F = 1) indicates the prior probability that the healthy node state is faulty; P(T, V, ∈) indicates the prior probability of temperature, vibration and strain data; P(T, V, ∈|F = 1) indicates the likelihood that the healthy node state is faulty.

[0041] On the other hand, the present invention also provides a fault prediction system for a main bearing of a wind turbine generator, comprising a data acquisition module, a data correction module, a wind turbine generator main bearing state prediction module and a wind turbine generator main bearing fault warning module;

[0042] The data acquisition module is used to collect vibration data, strain data, temperature data of the main bearing of the wind turbine generator and environmental data of the environment in which the wind turbine generator is located in chronological order and pre-process the data;

[0043] The data correction module is used to classify the vibration data and strain data of the main bearing of the wind turbine generator according to the temperature data, and respectively correct the vibration data and strain data of each type of the main bearing of the wind turbine generator;

[0044] The wind turbine main bearing state prediction module is used to construct a data set based on the corrected vibration data, the corrected strain data, the temperature data and the environmental data of the environment in which the wind turbine is located; construct a wind turbine main bearing state prediction model, input the data set into the wind turbine main bearing state prediction model for training, and predict the vibration, strain and temperature data of the wind turbine main bearing through the trained wind turbine main bearing state prediction model;

[0045] The wind turbine main bearing fault warning module is used to calculate the posterior probability of future wind turbine main bearing failure based on the predicted vibration, strain and temperature data of the wind turbine main bearing, judge the health status of the wind turbine main bearing according to the posterior probability and issue a warning.

[0046] On the other hand, the present invention further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any embodiment of the present invention when executing the program.

[0047] In yet another aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any embodiment of the present invention.

[0048] The present invention has the following beneficial effects:

[0049] Based on the complex variation law of material properties under low and high temperature conditions, the present invention proposes an improved correction formula, which can accurately reflect the nonlinear influence of low and high temperature conditions on the performance of the fan main bearing, making the monitoring data more in line with the actual working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] It should be understood that the step numbers used in this document are only for convenience of description and are not intended to limit the order in which the steps are executed.

[0053] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0054] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0055] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0056] Embodiment 1:

[0057] See also Figure 1 , a fault prediction method for a main bearing of a wind turbine generator comprises the following steps:

[0058] Collect vibration data, strain data, temperature data of the main bearing of the wind turbine and environmental data of the environment in which the wind turbine is located in chronological order and pre-process the data;

[0059] The environmental data of the environment in which the wind turbine is located includes gas content and magnetic field data, and the vibration data is obtained by arranging a vibration sensor on the bearing end cover of the main bearing of the wind turbine;

[0060] The temperature data is obtained by arranging a temperature sensor on the bearing end cover of the fan main bearing, which is mainly to obtain the temperature data around the vibration sensor, considering the influence of temperature on the vibration sensor;

[0061] The strain data is obtained by arranging strain gauges on the bearing seat of the fan main bearing;

[0062] The gas data is obtained by arranging a pipeline on the bearing seat of the fan main bearing to collect gas, and then analyzing the gas content data using a gas chromatograph;

[0063] The magnetic field data is obtained by arranging a magnetic sensor on the bearing seat of the main bearing of the wind turbine near the generator or other electromagnetic equipment to obtain magnetic field changes;

[0064] The vibration data and strain data of the main bearing of the wind turbine generator are classified according to the temperature data, and the vibration data and strain data of the main bearing of each type of wind turbine generator are corrected respectively;

[0065] constructing a data set based on the corrected vibration data, the corrected strain data, the temperature data, and the environmental data of the environment in which the wind turbine is located;

[0066] Construct a wind turbine main bearing state prediction model, input the data set into the wind turbine main bearing state prediction model for training, and predict the vibration, strain and temperature data of the wind turbine main bearing through the trained wind turbine main bearing state prediction model;

[0067] Based on the predicted vibration, strain and temperature data of the wind turbine main bearing, the posterior probability of future wind turbine main bearing failure is calculated, and the health status of the wind turbine main bearing is judged according to the posterior probability and an early warning is issued.

[0068] As a preferred implementation of this embodiment, the statistical features of vibration data and strain data are extracted by the following formula:

[0069]

[0070]

[0071]

[0072] Use Z-score normalization to normalize the gas content and magnetic field data:

[0073]

[0074] As a preferred implementation of this embodiment, the temperature data is sorted from large to small, the sorted temperature data is divided into three equal parts, and the vibration data and strain data corresponding to the temperature data are classified into high temperature data, normal temperature data and low temperature data according to the quantile, as shown in the following formula:

[0075]

[0076] Where: Q represents the location of the current temperature data.

[0077] As a preferred implementation of this embodiment, the vibration data in the low temperature data is corrected, as shown in the following formula:

[0078]

[0079] Where: V(t) represents the vibration data at time t; V ′ (t) represents the corrected vibration data at time t; α represents the low temperature variation function; T t Represents the temperature data at time t; It represents the gradient of temperature data, describing the influence of temperature non-uniformity inside the bearing material or in the surrounding environment on vibration; β represents the attenuation factor of low-temperature vibration, which controls the attenuation of vibration changes caused by temperature gradient. When it is large, it indicates that the vibration signal is less sensitive to extreme temperature differences. When it is smaller, the attenuation effect is weaker and the vibration change is more significant;

[0080] α(T t )=α0·(0.5+k1T 2 +k2ΔT)

[0081] Among them: α0 represents the influence coefficient at the reference temperature; k1 and k2 represent the nonlinear adjustment coefficients of temperature response, which are obtained based on experimental data; ΔT represents the change of current temperature compared with the reference temperature;

[0082] The strain data in the low temperature data is corrected as shown in the following formula:

[0083]

[0084] Where: ∈(t) represents the strain data at time t; ∈ ′ (t) represents the corrected strain data at time t; γ represents the elastic modulus correction function; σ represents the material stress; κ represents the attenuation factor of low temperature strain;

[0085]

[0086] Where: γ0 represents the reference influence coefficient; m1 represents the linear response coefficient of temperature to material deformation; m2 represents the nonlinear enhancement coefficient of stress to material deformation;

[0087] In this embodiment, the nonlinear enhancement coefficient of stress to material deformation is specifically the bearing crack growth rate, which is specifically shown in the following formula:

[0088]

[0089] in: It represents the crack growth rate per cycle; ΔK represents the stress intensity factor range; C and m represent material constants;

[0090] As a preferred implementation of this embodiment, the vibration data in the high temperature data is corrected, as shown in the following formula:

[0091]

[0092] Where: δ represents the lubricant degradation sensitivity coefficient; η0 represents the lubricant base viscosity; η(T t ) indicates T t The viscosity of the lubricating oil at the temperature; φ represents the high temperature lubrication degradation correction function; ψ represents the high temperature vibration adjustment coefficient; ΔT represents the temperature change;

[0093]

[0094] Where: ε represents the viscosity-temperature coefficient; T0 represents the reference temperature (such as 25°C); A represents the cross-sectional area of ​​the bearing; Ea represents the activation energy of the lubricating oil;

[0095]

[0096] Where: φ0 represents the high-temperature lubrication degradation benchmark coefficient; c1 represents the degradation factor of lubricating oil under high temperature conditions, reflecting the rapid disintegration of the molecular structure of lubricating oil under high temperature. The larger the value, the stronger the degree of degradation of lubricating oil to high temperature; c2 represents the local performance change of lubricating oil caused by temperature difference, describing the weakening effect of temperature non-uniformity on lubrication performance; C k It indicates the concentration of pollutants in the lubricating oil, reflecting the proportion of solid particles or chemical impurities in the lubricating oil; c3 indicates the influence coefficient of pollutants on lubrication degradation;

[0097] The strain data in the high temperature data is corrected as shown in the following formula:

[0098]

[0099] Where: ω represents the coupling term of temperature and stress; χ represents the correction factor of thermal expansion, which represents the comprehensive influence of thermal expansion effect on material properties (such as strain, vibration, etc.);

[0100]

[0101] Where: μ represents the adjustment coefficient of high temperature to stress sensitivity; It represents a nonlinear adjustment term, indicating that the local temperature non-uniformity aggravates the damage accumulation of the material;

[0102] As a preferred implementation of this embodiment, the corrected vibration data is subjected to joint wavelet transform to obtain a corresponding vibration spectrum sequence, as shown in the following formula:

[0103]

[0104] Where: a represents the scale parameter; b represents the translation parameter; ψ represents the mother wavelet; W x represents the corrected vibration data after joint wavelet transformation;

[0105] The vibration spectrum sequence includes three different frequencies: low frequency, medium frequency and high frequency, which are divided into signals of three different spectrums: low frequency spectrum, medium frequency spectrum and high frequency spectrum by a multi-band filter;

[0106] Multiband filters consist of multiple bandpass filters, each responsible for a specific frequency range;

[0107] The low frequencies in the vibration spectrum sequence reflect macroscopic motion, while the high frequencies are sensitive to local defects;

[0108] As a preferred implementation of this embodiment, the wind turbine main bearing state prediction model is constructed based on a long short-term memory network model, and the data set is divided into: 70% training set, 15% validation set, and 15% test set according to a general proportion;

[0109] The training set is input into the LSTM model, and the model is trained to predict future time series signals to obtain an initial LSTM model;

[0110] Using the validation set and the test set, the initial LSTM model is optimized, and the appropriate number of LSTM layers, number of units, and batch size are adjusted to achieve optimal working efficiency of the model and improve the generalization ability of the model and the robustness of the output;

[0111] The adjusted value can be expressed as: h t =LSTM(x t ,h t-1 ), where x tis the input feature vector at time t, which contains the normalized gas content and magnetic field data as well as the vibration spectrum sequence and strain data, h t-1 is the hidden state of the previous time step;

[0112] Since we normalized the target variable during training, we need to perform an inverse transformation to restore it to its original scale as follows:

[0113]

[0114] Where σ and μ are the mean and standard deviation of the target variable (such as temperature, vibration, strain) calculated during the training phase, and y t+n is the future temperature T after inverse change, vibration time series V, strain time series S;

[0115] As a preferred implementation of this embodiment, the predicted vibration, strain and temperature data of the main bearing of the wind turbine generator are input into a Bayesian network, and several nodes are defined in the Bayesian network, including a temperature node, a vibration node, a strain node and a main bearing health status node;

[0116] Construct the causal relationship between nodes, set the temperature node, vibration node, and strain node as the parent node of the health state node, and use the prior probability distribution for vibration, strain, and temperature data. For the probability that the health state node state takes a specific value as a fault, the posterior probability that the health state node state is a fault is obtained by updating the following formula:

[0117]

[0118] Where: F = 1 means that the healthy node state is faulty (binary variable, 0 means normal, 1 means faulty); T means predicted temperature node data; V means vibration node data; ∈ means strain node data; P(F = 1 | T, V, ∈) means the posterior probability that the healthy node state is faulty; P(F = 1) means the prior probability that the healthy node state is faulty; P(T, V, ∈) means the prior probability of temperature, vibration and strain data; P(T, V, ∈ | F = 1) means the likelihood that the healthy node state is faulty;

[0119] By constructing a validation set and a test set to compare the predicted posterior probability with the actual fault event, the network structure and conditional probability table are continuously adjusted to obtain a constructed Bayesian network structure;

[0120] When P(F=1|T、V、∈)≥50%, an early warning warning is issued for the health status of the fan main bearing; when P(F=1|T、V、∈)<50%, the fan main bearing is judged to be in a healthy state.

[0121] Embodiment 2:

[0122] A fault prediction system for a main bearing of a wind turbine generator, comprising a data acquisition module, a data correction module, a wind turbine generator main bearing state prediction module and a wind turbine generator main bearing fault early warning module;

[0123] The data acquisition module is used to collect vibration data, strain data, temperature data of the main bearing of the wind turbine generator and environmental data of the environment in which the wind turbine generator is located in chronological order and pre-process the data;

[0124] The data correction module is used to classify the vibration data and strain data of the main bearing of the wind turbine generator according to the temperature data, and respectively correct the vibration data and strain data of each type of the main bearing of the wind turbine generator;

[0125] The wind turbine main bearing state prediction module is used to construct a data set based on the corrected vibration data, the corrected strain data, the temperature data and the environmental data of the environment in which the wind turbine is located; construct a wind turbine main bearing state prediction model, input the data set into the wind turbine main bearing state prediction model for training, and predict the vibration, strain and temperature data of the wind turbine main bearing through the trained wind turbine main bearing state prediction model;

[0126] The wind turbine main bearing fault warning module is used to calculate the posterior probability of future wind turbine main bearing failure based on the predicted vibration, strain and temperature data of the wind turbine main bearing, judge the health status of the wind turbine main bearing according to the posterior probability and issue a warning.

[0127] The system is used to implement the method in Example 1, which will not be described in detail here.

[0128] Embodiment three:

[0129] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any embodiment of the present invention when executing the program.

[0130] Embodiment 4:

[0131] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in any embodiment of the present invention is implemented.

[0132] In the embodiments of the present application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can be represented by: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c can be single or multiple.

[0133] Those of ordinary skill in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented in a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0134] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0135] In several embodiments provided in the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, 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, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), disk or optical disk, and other media that can store program codes.

[0136] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for predicting main bearing faults of a wind turbine, characterized in that: The following steps are involved: Collect vibration data, strain data, temperature data of the main bearing of the wind turbine and environmental data of the environment in which the wind turbine is located in chronological order and pre-process the data; Classifying the vibration data and strain data of the main bearing of the wind turbine generator according to the temperature data, and respectively correcting the vibration data and strain data of the main bearing of each type of wind turbine generator; constructing a data set based on the corrected vibration data, the corrected strain data, the temperature data, and the environmental data of the environment in which the wind turbine is located; Construct a wind turbine main bearing state prediction model, input the data set into the wind turbine main bearing state prediction model for training, and predict the vibration, strain and temperature data of the wind turbine main bearing through the trained wind turbine main bearing state prediction model; Based on the predicted vibration, strain and temperature data of the wind turbine main bearing, the posterior probability of future wind turbine main bearing failure is calculated, and the health status of the wind turbine main bearing is judged according to the posterior probability and an early warning is issued.

2. A method for predicting main bearing faults of a wind turbine according to claim 1, characterized in that: The temperature data is sorted from large to small, and the sorted temperature data is divided into three equal parts. The vibration data and strain data corresponding to the temperature data are classified into high temperature data, normal temperature data, and low temperature data according to the quantiles, as shown in the following formula: Where: Q represents the location of the current temperature data.

3. A method for predicting main bearing faults of a wind turbine according to claim 2, characterized in that: The vibration data in the low temperature data is corrected as shown in the following formula: Where: V(t) represents the vibration data at time t; V ′ (t) represents the corrected vibration data at time t; α represents the low temperature variation function; T t Represents the temperature data at time t; represents the gradient of temperature data; β represents the attenuation factor of low temperature vibration; The strain data in the low temperature data is corrected as shown in the following formula: Where: ∈(t) represents the strain data at time t; ∈′(t) represents the corrected strain data at time t; γ represents the elastic modulus correction function; σ represents the material stress; κ represents the attenuation factor of low temperature strain.

4. A method for predicting main bearing faults of a wind turbine according to claim 3, characterized in that: The vibration data in the high temperature data is corrected as shown in the following formula: Where: δ represents the lubricant degradation sensitivity coefficient; η0 represents the lubricant base viscosity; η(T t ) indicates T t The viscosity of the lubricating oil at the temperature; φ represents the high temperature lubrication degradation correction function; ψ represents the high temperature vibration adjustment coefficient; ΔT represents the temperature change; The strain data in the high temperature data is corrected as shown in the following formula: Where: ω represents the coupling term of temperature and stress; χ represents the correction factor of thermal expansion.

5. A method for predicting main bearing faults of a wind turbine according to claim 1, characterized in that: The corrected vibration data is subjected to joint wavelet transform to obtain the corresponding vibration spectrum sequence, as shown in the following formula: Where: a represents the scale parameter; b represents the translation parameter; ψ represents the mother wavelet; W x Represents the modified vibration data after the joint wavelet transformation.

6. A method for predicting main bearing faults of a wind turbine according to claim 1, characterized in that: The wind turbine main bearing state prediction model is constructed based on a long short-term memory network model.

7. A method for predicting main bearing faults of a wind turbine according to claim 1, characterized in that: Input the predicted vibration, strain and temperature data of the main bearing of the wind turbine into a Bayesian network, and define several nodes in the Bayesian network including a temperature node, a vibration node, a strain node and a main bearing health status node; Construct the causal relationship between nodes, set the temperature node, vibration node, and strain node as the parent node of the health state node, and use the prior probability distribution for vibration, strain, and temperature data. For the probability that the health state node state takes a specific value as a fault, the posterior probability that the health state node state is a fault is obtained by updating the following formula: Where: F = 1 indicates that the healthy node state is faulty; T indicates the predicted temperature node data; V indicates the vibration node data; ∈ indicates the strain node data; P(F = 1|T, V, ∈) indicates the posterior probability that the healthy node state is faulty; P(F = 1) indicates the prior probability that the healthy node state is faulty; P(T, V, ∈) indicates the prior probability of temperature, vibration and strain data; P(T, V, ∈|F = 1) indicates the likelihood that the healthy node state is faulty.

8. A wind turbine main bearing fault prediction system, characterized in that: It includes a data acquisition module, a data correction module, a wind turbine main bearing state prediction module and a wind turbine main bearing fault warning module; The data acquisition module is used to collect vibration data, strain data, temperature data of the main bearing of the wind turbine generator and environmental data of the environment in which the wind turbine generator is located in chronological order and pre-process the data; The data correction module is used to classify the vibration data and strain data of the main bearing of the wind turbine generator according to the temperature data, and respectively correct the vibration data and strain data of each type of the main bearing of the wind turbine generator; The wind turbine main bearing state prediction module is used to construct a data set based on the corrected vibration data, the corrected strain data, the temperature data and the environmental data of the environment in which the wind turbine is located; construct a wind turbine main bearing state prediction model, input the data set into the wind turbine main bearing state prediction model for training, and predict the vibration, strain and temperature data of the wind turbine main bearing through the trained wind turbine main bearing state prediction model; The wind turbine main bearing fault warning module is used to calculate the posterior probability of future wind turbine main bearing failure based on the predicted vibration, strain and temperature data of the wind turbine main bearing, judge the health status of the wind turbine main bearing according to the posterior probability and issue a warning.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images