Method for predicting service life of multi-source data driven wind power equipment based on acceleration effect

Abstract

The invention provides a multi-source data driven wind power equipment service life prediction method based on an acceleration effect, and relates to the field of wind power equipment service life prediction.The method comprises the steps that operation environment data and degradation fault information of wind power equipment are collected, and a wind power equipment basic performance degradation model is constructed; based on the primary acceleration stress, the secondary acceleration stress and the comprehensive acceleration equation, constructing a first wind power equipment accelerated degradation model; constructing a wind power equipment state multi-source data set based on the degradation fault information, and performing data updating on the first wind power equipment accelerated degradation model according to the wind power equipment state multi-source data set, a Bayesian function and an EM algorithm to obtain a second wind power equipment accelerated degradation model; and according to the degradation process multi-condition hypothesis function and the second wind power equipment accelerated degradation model, a wind power equipment multi-stage degradation process model is constructed to predict the residual life of the wind power equipment, the self-adaptive capability of the accelerated degradation model is enhanced, and the prediction accuracy of the residual life of the wind power equipment is improved.

Description

[0001] Cross-reference to related applications This application claims priority to Chinese application CN202411533001.2 entitled “Method and System for Predicting the Lifetime of Wind Turbine Equipment Based on Multi-Source Data Driven by Acceleration Effect”, filed on October 30, 2024, which is incorporated herein by reference in its entirety. Technical Field

[0002] This invention relates to the field of wind power equipment life prediction, and in particular to a multi-source data-driven method for wind power equipment life prediction based on acceleration effects. Background Technology

[0003] Social progress and development cannot be separated from the support of new energy sources. However, the contradiction between humanity's ever-increasing energy demand and the increasingly severe environmental situation and the dwindling reserves of non-renewable energy is becoming increasingly acute. Since the beginning of the 21st century, energy security issues and the new energy revolution have attracted widespread attention from the international community. In order to meet energy supply demands and achieve sustainable energy development, the transformation from a fossil fuel system to a sustainable energy system has become a basic trend in energy development worldwide. Wind energy, as one of the earliest developed and most important renewable energy sources in the integrated energy system, has enormous potential in greenhouse gas emission reduction due to its abundant reserves, renewability, and clean, pollution-free characteristics. It has become the mainstream direction of global renewable energy development and an important force driving industrial upgrading and adjusting the energy system.

[0004] Chinese Patent Publication No. CN117743939A discloses a method, system, and electronic device for predicting the remaining service life of wind turbines. The method includes acquiring current monitoring data of the wind turbine to be predicted; determining the remaining service life of the wind turbine to be predicted using a remaining service life prediction model based on the current monitoring data; wherein the remaining service life prediction model is obtained by training a parallel multi-scale fusion network using a training dataset; the training dataset includes processed monitoring data of the wind turbine used for training and corresponding remaining service life labels. However, the method provided in the above application assumes that the unit performance gradually deteriorates over time and does not consider new operating conditions and data changes, making it difficult to accurately describe the actual degradation evolution of wind turbines, resulting in inaccurate predictions of the remaining service life of wind turbines. Therefore, it is essential to provide a multi-source data-driven wind turbine life prediction method based on acceleration effects to improve the accuracy of remaining service life predictions. Summary of the Invention

[0005] In view of this, this invention proposes a multi-source data-driven method for predicting the remaining lifespan of wind power equipment based on the acceleration effect. By systematically collecting and analyzing the operating environment data and degradation fault information of wind power equipment, and combining a second wind power accelerated degradation model and a multi-condition assumption function, the method helps to improve the accuracy of predicting the remaining lifespan of wind power equipment.

[0006] This invention provides a multi-source data-driven method for predicting the lifespan of wind power equipment based on acceleration effects, the method comprising: Collect operating environment data and degradation fault information of wind power equipment, and construct a basic performance degradation model of wind power equipment based on the operating environment data and degradation fault information; Mechanism analysis was performed on the blade crack propagation data in the degradation fault information to obtain the main accelerating stress and secondary accelerating stress of the wind power equipment. Based on the main accelerating stress, the secondary accelerating stress and the comprehensive acceleration equation, a first wind power equipment accelerated degradation model was constructed. Based on the degradation fault information, a multi-source dataset of wind power equipment status is constructed. According to the multi-source dataset of wind power equipment status, Bayesian function and EM algorithm, the first wind power accelerated degradation model is updated to obtain the second wind power accelerated degradation model. Based on the multi-condition hypothesis function of the degradation process and the second accelerated degradation model of wind power equipment, a multi-stage degradation process model of wind power equipment is constructed to predict the remaining life of wind power equipment.

[0007] Based on the above technical solutions, preferably, the step of constructing a basic performance degradation model for wind power equipment based on the operating environment data and the degradation fault information specifically includes: Based on the degradation failure mechanism and failure mode in the operating environment data and degradation fault information, the main influencing indicators of wind power equipment performance degradation are determined; Based on the main influencing indicators, the nonlinear Gamma correction function, and the blade crack propagation data in the degradation fault information, a basic performance degradation model for wind power equipment is constructed.

[0008] Based on the above technical solutions, preferably, after constructing the basic performance degradation model of wind power equipment, the method further includes: All unknown parameters in the basic performance degradation model of the wind power equipment are given initial values, wherein the initial values ​​are derived from the prior distribution, maximum likelihood estimation and method moments; The unknown parameters are divided into multiple parameter groups according to conditional independence. In each iteration, an update operation is performed on each parameter group in sequence. The update operation includes, under the condition that the current values ​​of other parameters are fixed, sampling from the conditional posterior distribution of the first parameter group to update the parameter group; based on the updated first parameter group, sampling from the conditional posterior distribution of the second parameter group to update the parameter group; and sampling from the corresponding conditional posterior distribution and updating the remaining parameter groups in a preset order until all parameter groups have completed one update cycle. Repeat the update operation until the preset number of iterations is reached. Extract the posterior distribution, posterior mean, posterior standard deviation and confidence interval of each unknown parameter from the final posterior sample. Use the posterior sample to perform interval evaluation of key quantities in the wind power equipment life prediction method.

[0009] More preferably, the step of using the posterior sample to perform interval evaluation of key quantities in the wind power equipment life prediction method specifically includes: The basic performance degradation model of the wind power equipment is calculated based on the MAE index to obtain the mean absolute error of the basic performance degradation model of the wind power equipment. The basic performance degradation model of the wind power equipment is calculated based on the MSE index to obtain the mean square error of the basic performance degradation model of the wind power equipment. The basic performance degradation model of the wind power equipment is calculated based on the MAPE index, and the average absolute error percentage of the basic performance degradation model of the wind power equipment is calculated. The distance between the training sample bias and the test sample bias in the basic performance degradation model of the wind power equipment is calculated based on the DIC criterion.

[0010] More preferably, the expression function corresponding to the first wind power accelerated degradation model includes:

[0011]

[0012]

[0013] in, for Length of blade crack at any given time This represents the comprehensive acceleration stress index value. This represents the quantified value of the tensile stress, which indicates the principal accelerating stress. Denotes the Gamma distribution function, with shape parameter . , Represents the basic shape function, with the scale parameter being... , Represents the comprehensive acceleration equation, Represents Boltzmann's constant. Indicates the secondary acceleration stress coefficient. Indicates the principal accelerating stress coefficient. K This indicates the temperature in the operating environment of the wind power equipment. Indicates tensile stress. This represents the comprehensive acceleration stress index value. E a Indicates activation energy. Indicates the first accelerating stress coefficient. Indicates the second accelerating stress coefficient. This represents the Boltzmann constant.

[0014] More preferably, the multi-source dataset of wind power equipment status includes degradation cause data and degradation result data. The degradation cause data includes temperature monitoring data, stress intensity, and corrosive medium content. The degradation result data includes crack propagation length, material residual strength, and tensile strength.

[0015] More preferably, the step of updating the first wind turbine accelerated degradation model based on the multi-source dataset of wind turbine status, the Bayesian function, and the EM algorithm specifically includes: Based on the prior information and sample information in the multi-source dataset of the wind power equipment status, the degradation parameters in the first wind power equipment accelerated degradation model are updated in real time through a Bayesian function to obtain the posterior distribution parameters of the degradation parameters. The maximum likelihood estimate of the degradation parameters in the first wind power accelerated degradation model is obtained by iteratively using the EM algorithm, and the update values ​​of the remaining parameters in the first wind power accelerated degradation model corresponding to the maximum likelihood estimate are calculated according to the parameter update function.

[0016] More preferably, the principal accelerating stress characterizes the core stress that directly drives crack propagation, and the principal accelerating stress includes wind load and vibration load. The wind load and the vibration load are weighted and calculated to obtain the tensile stress quantification value corresponding to the principal accelerating stress. The secondary accelerating stress characterizes the environmental stress that indirectly affects crack propagation, and the secondary accelerating stress includes the blade service environment temperature, ambient relative humidity, and salt spray concentration.

[0017] A second aspect of this application provides a multi-source data-driven wind turbine life prediction system based on acceleration effects. The multi-source data-driven wind turbine life prediction system includes an information acquisition module, a degradation integration module, and a life prediction module. The information acquisition module is used to collect operating environment data and degradation fault information of wind power equipment, and to construct a basic performance degradation model of wind power equipment based on the operating environment data and degradation fault information. The degradation integration module is used to perform mechanism analysis on the blade crack propagation data in the degradation fault information, obtain the main accelerating stress and secondary accelerating stress of the wind turbine, and construct a first wind turbine accelerated degradation model based on the main accelerating stress, the secondary accelerating stress and the comprehensive acceleration equation; construct a wind turbine state multi-source dataset based on the degradation fault information, and update the first wind turbine accelerated degradation model according to the wind turbine state multi-source dataset, Bayesian function and EM algorithm to obtain a second wind turbine accelerated degradation model; The lifetime prediction module is used to construct a multi-stage degradation process model for wind power equipment based on the multi-condition assumption function of the degradation process and the second accelerated degradation model of wind power equipment, so as to predict the remaining lifetime of wind power equipment.

[0018] A third aspect of this application provides an electronic device including a processor, a memory, a user interface, and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory.

[0019] The multi-source data-driven wind power equipment lifetime prediction method based on acceleration effect provided by this invention has the following advantages over existing technologies: (1) By systematically collecting and analyzing the operating environment data and degradation fault information of wind power equipment, the basic performance degradation model can reflect the actual performance changes of the equipment under different environmental conditions, thereby improving the accuracy and practicality of the model. At the same time, by analyzing the mechanism of blade crack propagation data, the main accelerating stress and secondary accelerating stress affecting crack propagation can be effectively identified. The accelerated degradation model is dynamically updated using Bayesian function and EM algorithm to enhance the adaptive capability of the accelerated degradation model. Furthermore, by combining the second wind power equipment accelerated degradation model and multi-condition hypothesis function, it helps to improve the prediction accuracy of the remaining life of wind power equipment.

[0020] (2) The model performance is comprehensively evaluated by using multiple indicators such as MAE, MSE, MAPE and DIC, so that the analysis of prediction quality is not limited to a single indicator, which enhances the depth and effectiveness of the interpretation of the results. Combined with the calculation results of different indicators, it supports the continuous monitoring and optimization of the wind power equipment performance degradation model, ensuring that the model can adapt to new operating conditions and data changes in a timely manner and enhance its prediction ability.

[0021] (3) By using the Bayesian function to update the degradation parameters in real time, the model can continuously adjust according to the new observation data during the operation of the equipment, which significantly enhances the flexibility and response speed of the model to changes in the operating environment, thereby reducing the prediction error caused by environmental changes. The EM algorithm can effectively handle missing values ​​in the sample data, ensuring that even if the data is incomplete, the model can still provide reasonable parameter estimates and state assessments, improving the integrity and reliability of data processing. At the same time, the application of the parameter update function enables the model to adjust relevant parameters other than the degradation parameters, which enhances the overall coordination of the model, makes the relationship between the various parameters more accurate, and thus improves the model's prediction ability. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart illustrating a multi-source data-driven wind power equipment lifetime prediction method based on acceleration effect provided by the present invention; Figure 2 A schematic diagram of the performance degradation simulation results of the basic performance degradation model of wind power equipment provided by this invention; Figure 3 A schematic diagram of the framework of the multi-source data-driven wind power equipment life prediction system provided by the present invention; Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention.

[0024] Explanation of reference numerals in the attached figures: 1. Multi-source data-driven wind power equipment life prediction system; 11. Information acquisition module; 12. Degradation integration module; 13. Life prediction module; 2. Electronic equipment; 21. Processor; 22. Communication bus; 23. User interface; 24. Network interface; 25. Memory. Detailed Implementation

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0026] This invention discloses a multi-source data-driven method for predicting the lifespan of wind power equipment based on acceleration effects, with reference to... Figure 1 The steps of this method include S1 to S4.

[0027] Step S1: Collect operating environment data and degradation fault information of wind power equipment, and construct a basic performance degradation model of wind power equipment based on the operating environment data and degradation fault information.

[0028] In this step, the operating environment data includes environmental parameters such as wind speed, wind direction, temperature, humidity, and vibration. This data can be monitored and recorded in real time using sensor equipment. Degradation fault information can be collected by establishing a comprehensive fault information database to record various degradation failure modes and modes, such as blade crack propagation and gear wear. Fault information is analyzed to determine the main degradation failure mechanisms, such as fatigue, corrosion, and wear, and detailed information about the failure process is collected, including the time of failure, the extent of failure, and the cause of failure.

[0029] This step also includes steps S11 to S12.

[0030] Step S11: Based on the degradation failure mechanism and failure mode in the operating environment data and degradation fault information, determine the main influencing indicators of wind power equipment performance degradation.

[0031] In this step, based on the collected degradation failure information, the main degradation failure modes are identified, such as blade fatigue cracks, gear wear, and bearing failure. The physical mechanisms of each failure mode are analyzed, including stress concentration, surface wear, and material fatigue. For each failure mode, key operating environment factors influencing its occurrence and development are determined, such as wind, vibration, temperature, and humidity. Through experimental analysis or engineering experience, the quantitative impact of these environmental factors on the failure mechanism is determined. A comprehensive analysis of the key influencing factors under various failure modes is conducted to screen out the indicators most sensitive to overall performance degradation. For example, for blade fatigue cracks, wind and vibration may be the most significant influencing indicators, while for gear wear, temperature and load may be more critical.

[0032] Step S12: Based on the main influencing indicators, the nonlinear Gamma correction function, and the blade crack propagation data in the degradation fault information, a basic performance degradation model of wind power equipment is constructed.

[0033] In this step, the main influencing indicators and blade crack propagation data are first quantified. Based on the main influencing indicators determined in step S11, such as the main influencing indicators of "wind force" and "vibration" for blade fatigue cracks or "temperature" and "load" for gear wear, the operating environment data and degradation fault information are quantified.

[0034] First, the key influencing indicators are quantified: wind force is converted into wind speed load (unit: N / m²), and vibration is converted into vibration acceleration (unit: m / s²). Daily average values ​​are calculated using real-time data collected by sensors as input variables. Then, blade crack propagation data is quantified: blade crack length (unit: mm) and corresponding monitoring time (unit: h) are extracted from degradation fault information to form a crack propagation sequence, such as: crack length a1 at time t1, crack length a2 at time t2, ..., t n Crack length a at time n .

[0035] In this step, the nonlinear Gamma correction function satisfies the following condition: 1) X (0) = 0; 2) It has independent increments; 3) For any and , ,in It is a non-decreasing continuous function. Define the function. satisfy ,and , is called a function For shape function, β> 0 is the scale parameter. When selecting a shape function that makes the shape convex, a value of 0 is typically used. ,in b Used to describe the nonlinear process described above; the scaling parameter of the nonlinear Gamma degradation model β Let it be a random variable, and for ease of subsequent reliability calculations, let β It follows a Gamma distribution, that is... Therefore, the model can be represented as:

[0036] Based on the law of total probability for continuous random variables, obtain the following... t Degradation over time The distribution density is calculated using the following formula:

[0037]

[0038] in, for The amount of degradation at any given time, i.e., the blade crack length, is expressed in mm; for The distribution function, whose shape parameter is The scale parameter is ; Shape function: This is a proportionality coefficient, which is determined by the properties of the blade material; for example, it is 0.02 to 0.05 for glass fiber composite materials. The mapping function for the main influencing indicators, such as wind load. and vibration acceleration The combined effects: The weights are calibrated using crack propagation experimental data; Runtime, in hours; This represents the nonlinear coefficient, characterizing the degree of nonlinearity in the degradation rate. It is obtained by fitting blade crack propagation data and is typically taken as 1.2 to 1.8. It is a scale parameter that follows a Gamma distribution: The shape parameter is determined by the blade's factory quality inspection data and is taken as 5 to 8. The scaling parameter, estimated from historical degradation data of the same batch of blades, is taken as 0.1~0.3. F Describing the degrees of freedom as and of F distributed.

[0039] In this embodiment, after constructing the basic performance degradation model of wind power equipment, the following is also included: All unknown parameters in the basic performance degradation model of wind power equipment are given initial values, which are derived from prior distribution, maximum likelihood estimation and method moments. The unknown parameters are divided into multiple parameter groups based on conditional independence. In each iteration, update operations are performed on each parameter group in sequence. The update operation includes: drawing samples from the conditional posterior distribution of the first parameter group to update the parameter group while keeping the current values ​​of other parameters fixed; drawing samples from the conditional posterior distribution of the second parameter group to update the parameter group based on the updated first parameter group; and sampling and updating the remaining parameter groups from their corresponding conditional posterior distributions in a preset order until all parameter groups have completed one update cycle. Repeat the update operation until the preset number of iterations is reached. Extract the posterior distribution, posterior mean, posterior standard deviation and confidence interval of each unknown parameter from the final posterior sample. Use the posterior sample to perform interval evaluation on the key quantities in the wind power equipment life prediction method.

[0040] Furthermore, WinBUGS was used to simulate and estimate the parameters of a wind power degradation model based on the Gibbs sampling method, obtaining the posterior distribution of the parameters, their mean, standard deviation, confidence interval, and parameter changes during the iteration process, such as... Figure 2 As shown, Figure 2 This is a schematic diagram of the performance degradation simulation results of the basic performance degradation model of wind power equipment.

[0041] 1) Initialize samples ; 2) From the full conditional distribution Samples drawn from ; 3) From the full conditional distribution Samples drawn from ; 4) From the full conditional distribution Samples drawn from .

[0042] In this embodiment, the basic performance degradation model of wind power equipment is calculated based on the MAE index to obtain the mean absolute error of the basic performance degradation model of wind power equipment. The calculation formula is expressed as follows:

[0043] in, Indicates sample size. Indicates the first i The true value of each sample Indicates the first i One predicted output.

[0044] The mean square error (MSE) of the basic performance degradation model of wind power equipment is calculated based on the MSE index. The calculation formula is as follows:

[0045] The basic performance degradation model of wind power equipment is calculated based on the MAPE index. The mean absolute error percentage of the basic performance degradation model of wind power equipment is expressed by the following formula:

[0046] The distance between the training sample bias and the test sample bias in the basic performance degradation model of wind power equipment is calculated based on the DIC criterion. The calculation formula is expressed as follows:

[0047] in, express The posterior distribution mean, Indicates known posterior parameters The likelihood function E under the given conditions θ|y Represents the time scale parameter of a given sample θ The expectation of the posterior distribution. This represents the posterior mean of the bias.

[0048] By comprehensively evaluating model performance using multiple indicators such as MAE, MSE, MAPE, and DIC, the analysis of prediction quality is not limited to a single indicator, enhancing the depth and effectiveness of result interpretation. Combining the calculation results of different indicators supports continuous monitoring and optimization of wind power equipment performance degradation models, ensuring that the models can adapt to new operating conditions and data changes in a timely manner, thus enhancing their predictive capabilities. Quantifying model performance provides more scientific data support for equipment operation and maintenance decisions, improving the efficiency and reliability of wind farm management. By reducing error indicators and ensuring the model's generalization ability, the service life of equipment can be effectively extended, maintenance cycles optimized, and ultimately, the long-term economic and environmental benefits of wind farms improved.

[0049] Step S2: Perform mechanism analysis on the blade crack propagation data in the degradation fault information, obtain the main accelerating stress and secondary accelerating stress of the wind turbine, and construct the first wind turbine accelerated degradation model based on the main accelerating stress, secondary accelerating stress and comprehensive acceleration equation.

[0050] In this step, the primary accelerating stress includes common and critical factors that affect the failure of wind power equipment, such as basic performance parameters like tensile strength; the secondary accelerating stress includes minor factors that affect the failure of wind power equipment, such as equipment maintenance frequency, overall equipment structure, enterprise management model, and personnel quality.

[0051] The principal accelerating stress characterizes the core stress that directly drives crack propagation. The principal accelerating stress includes wind load and vibration load. A weighted calculation of the wind load and vibration load is performed to obtain the quantified tensile stress value corresponding to the principal accelerating stress. The secondary accelerating stress characterizes the environmental stress that indirectly affects crack propagation. The secondary accelerating stress includes the blade's service environment temperature, relative humidity, and salt spray concentration. First, based on the mechanism analysis of blade crack propagation data, the principal and secondary accelerating stresses are quantified: Principal accelerating stress is the core stress that directly drives crack propagation. For blade fatigue cracks, it mainly includes: wind load ( (Unit: N / m²) includes the blade load calculated from real-time wind speed, collected and converted by an anemometer; vibration load ( (Unit: m / s²) includes the vibration acceleration during blade operation, collected by vibration sensors. The tensile stress quantification value of the principal accelerating stress is expressed as tensile stress. This indicates that it was obtained through weighted calculation: This weight is determined by crack propagation experimental data.

[0052] Secondary accelerating stresses are environmental stresses that indirectly affect crack propagation. For blade crack propagation, they mainly include: temperature ( ,unit: This includes the ambient temperature of the blades during service, collected by a temperature sensor; humidity ( ,unit: This includes ambient relative humidity, collected by a humidity sensor; and salt spray concentration (…). (Unit: mg / m³) includes salt spray corrosion factors from coastal wind fields, collected by corrosion sensors. Secondary accelerated stress occurs through temperature ( Integrating into the comprehensive acceleration equation, humidity and salt spray affect the principal acceleration stress through correction factors. For example, when the salt spray concentration C > 0.5 mg / m³, the tensile stress... Revised to .

[0053] The comprehensive acceleration equation is the core equation relating principal and secondary accelerating stresses to the degradation rate. It is used to quantify the accelerating effect of stress on crack propagation, and the formula is as follows:

[0054] in, It represents the comprehensive accelerated stress index value, characterizing the degree of degradation acceleration under the combined action of principal and secondary stresses; The temperature in the secondary accelerating stress is directly used as an input variable in the equation. The tensile stress quantification value (N / m²) representing the principal accelerating stress is directly used as the input variable of the equation; The activation energy of the blade material (J / mol) is determined by the properties of the blade material. Represents the Boltzmann constant; This represents the secondary acceleration stress coefficient (1 / ℃), obtained from experiments on the effect of temperature on material properties, and is usually taken as 0.015; This represents the principal accelerating stress coefficient (m² / N), obtained from experiments on the effect of tensile stress on crack propagation, and is typically taken as... .

[0055] The first accelerated degradation model is based on nonlinearity. The degradation process will be accelerated by a combination of stress indices. Incorporating the degradation rate parameter, the core formula is as follows:

[0056] in: express The length of the blade crack at any given moment, i.e., the amount of degradation; Denotes the Gamma distribution function, with shape parameter . The scale parameter is Its basic performance degradation model ); denoted by the basic shape function, k represents the scaling factor, and b represents the nonlinearity factor, which is taken from the calibration results of the basic performance degradation model. The comprehensive acceleration stress index is calculated in step S2 by multiplying it by the shape parameter. The degradation rate is adjusted, where a larger A value results in a faster degradation rate.

[0057] Furthermore, the mechanism analysis of crack propagation failure in wind turbine blades reveals the key principal and secondary accelerating stresses influencing this degradation process. Specifically, for the failure mechanism of crack propagation, the main accelerating stress factors are identified—the stresses that directly lead to crack propagation. For fatigue failure, principal accelerating stresses may include wind loads and vibration loads; for corrosion failure, they may include chemical corrosion environments. In addition to the direct principal accelerating stresses, it is also necessary to analyze some stress factors that indirectly affect crack propagation, i.e., secondary accelerating stresses. For blade crack propagation, secondary accelerating stresses may include environmental factors such as temperature, humidity, and salt spray, which may indirectly promote crack propagation by affecting material properties and surface conditions.

[0058] The expression for the comprehensive acceleration equation is:

[0059] in, Represents the comprehensive acceleration equation, K This indicates the temperature in the operating environment of the wind power equipment. Indicates tensile stress. This represents the comprehensive acceleration stress index value. E a Indicates activation energy. Indicates the first accelerating stress coefficient. Indicates the second accelerating stress coefficient. Represents the Boltzmann constant, Boltzmann constant κ, and tensile stress. The primary stress, along with secondary stresses such as various environmental indicators, the structural characteristics of wind turbines, the factory performance of individual products, and the safety management effectiveness at wind turbine operation sites, constitutes the comprehensive accelerating stress index value. This study integrates the comprehensive accelerated stress evaluation factors of wind power equipment with the various performance levels of the products at the factory, as well as factors such as the working environment, operating status, overall equipment structure, and safety management effectiveness during service. Based on the composition of the influencing factors of each major category of indicators, the study analyzes and supplements the underlying calculation indicators, thereby forming a comprehensive accelerated stress index system for wind power equipment performance degradation.

[0060] Continuous-time stochastic processes This is a nonlinear accelerated Gamma process considering random effects. The role of the acceleration equation in the nonlinear Gamma process is to adjust the size of the shape parameter in the incremental distribution of the degradation process according to the magnitude of various accelerating stresses, thereby changing the degradation rate. This random process should satisfy the following conditions: 1) ; 2) It has independent increments; 3) For any and , ,in It is a non-decreasing continuous function; in, ,and .

[0061] Based on the analysis of primary accelerating stress and the construction and calculation of secondary accelerating stress index system, the final accelerated degradation model is obtained, and the calculation formula is as follows:

[0062] in: express The length of the blade crack at any given moment, i.e., the amount of degradation; Let Gamma be the distribution function, with shape parameter . The scale parameter is Its basic performance degradation model ); denoted by the basic shape function, k represents the scaling factor, and b represents the nonlinearity factor, which is taken from the calibration results of the basic performance degradation model. The comprehensive acceleration stress index, calculated in step 2, is obtained by multiplying it by the shape parameter. The degradation rate is adjusted, where a larger A value results in a faster degradation rate. Indicates the temperature in the service environment. Indicates tensile stress. and The primary stress is the environmental performance, while secondary stresses include various environmental indicators, the structural characteristics of wind turbines, the factory performance of individual products, and the safety management effectiveness at wind turbine operation sites. The comprehensive accelerated stress index value represents the evaluation factors of the comprehensive accelerated stress value of wind power equipment, as well as the various performance levels of the product at the factory, and factors such as the working environment, working status, overall equipment structure, and safety management effectiveness during service. According to the composition of the influencing factors of each major category of indicators, the underlying calculation indicators are analyzed and supplemented to form a comprehensive accelerated stress index system for wind power equipment performance degradation.

[0063] To ensure the acceleration factor remains constant, the parameters in the accelerated degradation model of wind turbine blade crack propagation are set appropriately. The calculation formula is as follows:

[0064] in, Let t represent any degenerate quantity. α,i Indicates the first iAt stress level, corresponding confidence level α lifespan, t α,j Indicates the first j At stress level, corresponding confidence level α lifespan, Indicates the first i At stress level, corresponding confidence level lifespan, Indicates the first j At stress level, corresponding confidence level Lifespan.

[0065] Based on the principle of constant acceleration factor, the life distribution prediction of wind power equipment under normal operating stress is obtained, and the calculation formula is as follows:

[0066] in, The failure threshold for wind power equipment. Let be the probability distribution function of the remaining service life of the wind turbine. T Let b represent the remaining lifespan random variable, and b represent the nonlinear coefficient.

[0067] Step S3: Construct a multi-source dataset of wind power equipment status based on degradation fault information. Update the data of the first wind power accelerated degradation model according to the multi-source dataset of wind power equipment status, Bayesian function and EM algorithm to obtain the second wind power accelerated degradation model.

[0068] The multi-source dataset of wind power equipment status includes degradation cause data and degradation result data. The degradation cause data includes temperature monitoring data, stress intensity, and corrosive medium content, while the degradation result data includes crack propagation length, material residual strength, and tensile strength.

[0069] This step also includes steps S31 to S32.

[0070] Step S31: Based on the prior information and sample information in the multi-source dataset of wind power equipment status, the degradation parameters in the first wind power accelerated degradation model are updated in real time using a Bayesian function to obtain the posterior distribution parameters of the degradation parameters.

[0071] The conditional distribution of actual degradation data is obtained based on the equipment drift coefficient given in the degradation model. The calculation formula is as follows:

[0072] in, Indicates a time interval.

[0073] make and X and The complete log-likelihood function corresponding to all samples is obtained by treating them as observables, and the calculation formula is as follows:

[0074] The parameter update process is described using Bayesian principles to obtain the posterior distribution expression of the device drift coefficient. The calculation formula is as follows:

[0075] in, , and For ease of description, the unknown parameters that are dynamically updated in the detection data of this model are represented by a vector. Represents unknown parameters, i.e. ; The number of degradation information items for each product sample. At discrete time points Measure its degradation value and obtain A degraded data point, namely To facilitate model establishment and update calculations, the same measurement time is used for all samples. ;remember This dataset includes the complete performance degradation data of both prototype and service products. Indicates the first A vector composed of the performance degradation data of each individual unit over time, i.e. and ;make Indicates from time At the time During the process The degradation increment of each monomer; parameters Follow the mean variance is The sample follows a normal distribution; at this point, the sample information is known. and prior information In the case of The posterior distribution is still a normal distribution. Let the mean and variance of the posterior distribution be respectively... and ; in, , , This indicates that during the parameter update process of the wind turbine blade crack propagation accelerated degradation model based on Bayesian update and EM algorithm, the first... In the next iteration, the drift coefficient is obtained by fusing prior information on the drift coefficient with real-time degradation sample information of the wind power equipment. The posterior distribution mean, This represents the stochastic parameter describing the accelerated degradation rate of wind turbine blade crack propagation in the Wiener degradation model considering stochastic effects, and , The mean of the prior distribution of the drift coefficients. Prior distribution variance of drift coefficient; This indicates the drift coefficient under the same iterative scenario described above. The variance of the posterior distribution; This represents the number of iterations in the EM algorithm, used to gradually approximate the optimal estimate of the drift coefficient.

[0076] Step S32: Obtain the maximum likelihood estimate of the degradation parameters in the first wind power accelerated degradation model through EM algorithm iteration, and calculate the update values ​​of the remaining parameters in the first wind power accelerated degradation model corresponding to the maximum likelihood estimate according to the parameter update function.

[0077] First, the parameter system and classification of the first wind power equipment accelerated degradation model: First Wind Power Equipment Accelerated Degradation Model The complete parameter system includes: The core degradation parameter, also known as the maximum likelihood estimation object, is the equipment drift coefficient. It directly determines the rate of degradation, that is ; The remaining parameters, i.e., the distribution parameters that need to be updated synchronously: because Due to random effects, it is assumed that it follows a normal distribution. Meanwhile, the model's random error follows Therefore, the remaining parameters include: This represents the mean of the drift coefficients, which follow a normal distribution and reflect the average degradation rate of the population. The variance of the drift coefficient indicates that it follows a normal distribution, reflecting individual differences in degradation. The variance of the diffusion coefficient indicates that it follows a normal distribution, reflecting the magnitude of random error.

[0078] Obtaining latent variables under conditions of degraded data and prior parameter estimates. The expectation of the complete log-likelihood function is calculated using the following formula:

[0079] in, This represents the objective function used to quantify the expectation of the "complete log-likelihood function of latent variables" during the parameter update process of the wind power accelerated degradation model based on Bayesian update and EM algorithm.

[0080] Specifically defined as: given the degradation data of oil-specific pipelines and the prior parameter estimates of the current iteration step, such as the prior mean of the drift coefficients in the Wiener degradation model. ,variance Prior values ​​of shape parameters in the Gamma degradation model Prior values ​​of scale parameters Under the same conditions, the expectation of the complete log-likelihood function of the latent variables with respect to the posterior distribution of the latent variables; The function represents the mathematical expectation operator, which in the above scenario specifically refers to the "conditional expectation calculation operator based on the posterior distribution of latent variables".

[0081] Maximizing the expectation of the complete log-likelihood function is as follows:

[0082] The expression for the parameter update function is:

[0083] in, This indicates that the drift coefficient follows a normal distribution with mean. k Indicates the number of iterations. n Indicates the number of products. Indicates the first i The mean of the posterior distribution of each individual, The variance represents the drift coefficient following a normal distribution. Indicates the first i The variance of the posterior distribution of each individual, The variance represents the normal distribution of the diffusion coefficient. Indicates time interval, Indicates from time At the time During the process i The degradation increment of each monomer, m Indicates the number of degraded data. The mean drift coefficient update value for the (k+1)th iteration (mm / h) b ); The drift coefficient variance update value for the (k+1)th iteration ((mm / h) b ) 2 ); : The updated value of the diffusion coefficient variance in the (k+1)th iteration (mm²); k is the iteration step, initially k=0; n is the number of individual devices, such as the number of blades in the same wind farm; Let be the posterior mean of the drift coefficients of the i-th individual in the k-th iteration, which is obtained by Bayesian update, as follows: ; The posterior variance of the drift coefficient of the i-th individual in the k-th iteration is obtained by Bayesian update, as follows: m represents the number of degradation data points for each individual, such as the number of monitoring times per leaf. For the j-th time interval, such as the time difference between two monitoring sessions, the unit is hours (h). ; The degradation increment of the i-th monomer during the j-th time interval, i.e., the change in crack length, is expressed in mm. b is a nonlinear coefficient, which is taken from the previous model.

[0084] By using Bayesian functions to update degradation parameters in real time, the model can continuously adjust based on new observation data during equipment operation, significantly enhancing the model's flexibility and response speed to changes in the operating environment. This reduces prediction errors caused by environmental changes. The EM algorithm can effectively handle missing values ​​in sample data, ensuring that the model can still provide reasonable parameter estimates and state assessments even when data is incomplete, improving the completeness and reliability of data processing. At the same time, the application of parameter update functions allows the model to adjust relevant parameters other than degradation parameters, which enhances the overall coordination of the model, makes the relationships between various parameters more accurate, and thus improves the model's predictive ability.

[0085] Step S4: Based on the multi-condition hypothesis function of the degradation process and the second accelerated degradation model of wind power equipment, construct a multi-stage degradation process model of wind power equipment to predict the remaining life of wind power equipment.

[0086] The expression for the multi-condition hypothesis function of the degradation process is:

[0087] in, X ( t ) represents the multi-condition hypothesis function of the degradation process. X (0) represents the initial degradation amount. Indicates the first l Drift parameters in each degradation stage For the product in the l The total amount of degradation in each degradation stage Indicate the first l The variance of the diffusion coefficient in each degradation stage follows a normal distribution.

[0088] Based on multiple updates of accelerated stress, the device's degradation path has entered the [stage / phase]. At this stage, the degradation process satisfies the following assumptions:

[0089] in, .

[0090] The real-time remaining life distribution prediction results of wind turbines during the degradation process of accelerated stress updates are obtained, and the calculation formula is as follows:

[0091] in, Represents random variables The probability density function; This represents the observed or evaluated degradation-related features. In the context of wind power equipment, these are the actual monitored degradation indicators, such as blade crack length and equipment performance degradation. They are key observations for connecting the model with actual data. The number of sample categories; The summation index is used for traversal. arrive Grouping; For the first The normal distribution function of the group, where, For the first The mean of a group's normal distribution characterizes the central tendency of the group's degradation features; For the first The variance of a group's normal distribution reflects the degree of dispersion of its degradation characteristics; overall Used to describe the Statistical distribution patterns of group degradation data; The mean parameter is related to the overall distribution; For the corresponding The variance parameter reflects the dispersion of the core random variable and shows the fluctuation range of the core parameter under different equipment or different operating conditions. The variance parameter is a supplement.

[0092] By systematically collecting and analyzing the operating environment data and degradation fault information of wind power equipment, the constructed basic performance degradation model can reflect the actual performance changes of the equipment under different environmental conditions, thereby improving the accuracy and practicality of the model. At the same time, through the mechanism analysis of blade crack propagation data, the principal and secondary accelerating stresses affecting crack propagation can be effectively identified. The accelerated degradation model is dynamically updated using Bayesian functions and EM algorithms to enhance its adaptive capability. Furthermore, by combining the second wind power accelerated degradation model and multi-condition assumption functions, high-precision prediction of the remaining life of wind power equipment can be achieved.

[0093] Based on the above method, this application discloses a multi-source data-driven wind power equipment life prediction system based on acceleration effect, referencing... Figure 3 The multi-source data-driven wind power equipment life prediction system 1 includes an information acquisition module 11, a degradation integration module 12, and a life prediction module 13, wherein... The information acquisition module 11 is used to collect operating environment data and degradation fault information of wind power equipment, and to construct a basic performance degradation model of wind power equipment based on the operating environment data and degradation fault information. The degradation integration module 12 is used to perform mechanism analysis on the blade crack propagation data in the degradation fault information, obtain the main accelerating stress and secondary accelerating stress of the wind turbine, and construct a first wind turbine accelerated degradation model based on the main accelerating stress, secondary accelerating stress and comprehensive acceleration equation; construct a wind turbine state multi-source dataset based on the degradation fault information, and update the first wind turbine accelerated degradation model according to the wind turbine state multi-source dataset, Bayesian function and EM algorithm to obtain a second wind turbine accelerated degradation model; The lifetime prediction module 13 is used to construct a multi-stage degradation process model of wind power equipment based on the multi-condition assumption function of the degradation process and the second accelerated degradation model of wind power equipment, so as to predict the remaining lifetime of wind power equipment.

[0094] In one example, the information acquisition module 11 is used to determine the main influencing indicators of wind power equipment performance degradation based on the degradation failure mechanism and failure mode in the operating environment data and degradation fault information; and to construct a basic performance degradation model of wind power equipment based on the main influencing indicators, the nonlinear Gamma correction function and the blade crack propagation data in the degradation fault information.

[0095] In one example, after constructing the basic performance degradation model for wind turbines, the following is also included: All unknown parameters in the basic performance degradation model of wind power equipment are given initial values, which are derived from prior distribution, maximum likelihood estimation and method moments. The unknown parameters are divided into multiple parameter groups based on conditional independence. In each iteration, update operations are performed on each parameter group in sequence. The update operation includes: drawing samples from the conditional posterior distribution of the first parameter group to update the parameter group while keeping the current values ​​of other parameters fixed; drawing samples from the conditional posterior distribution of the second parameter group to update the parameter group based on the updated first parameter group; and sampling and updating the remaining parameter groups from their corresponding conditional posterior distributions in a preset order until all parameter groups have completed one update cycle. Repeat the update operation until the preset number of iterations is reached. Extract the posterior distribution, posterior mean, posterior standard deviation and confidence interval of each unknown parameter from the final posterior sample. Use the posterior sample to perform interval evaluation on the key quantities in the wind power equipment life prediction method.

[0096] In one example, posterior samples are used to perform interval evaluations of key quantities in wind turbine lifetime prediction methods, specifically including: The basic performance degradation model of wind power equipment is calculated based on the MAE index to obtain the mean absolute error of the basic performance degradation model of wind power equipment. The mean square error of the basic performance degradation model of wind power equipment is calculated based on the MSE index. The basic performance degradation model of wind power equipment is calculated based on the MAPE index, and the average absolute error percentage of the basic performance degradation model of wind power equipment is calculated. The distance between the training sample bias and the test sample bias in the basic performance degradation model of wind power equipment is calculated based on the DIC criterion.

[0097] In one example, the expression function corresponding to the accelerated degradation model of the first wind turbine includes:

[0098]

[0099]

[0100] in, for Length of blade crack at any given time This represents the comprehensive acceleration stress index value. This represents the quantified value of the tensile stress, which indicates the principal accelerating stress. Denotes the Gamma distribution function, with shape parameter . , Represents the basic shape function, with the scale parameter being... , Represents the comprehensive acceleration equation, Represents Boltzmann's constant. Indicates the secondary acceleration stress coefficient. Indicates the principal accelerating stress coefficient. K This indicates the temperature in the operating environment of the wind power equipment. Indicates tensile stress. This represents the comprehensive acceleration stress index value. E a Indicates activation energy. Indicates the first accelerating stress coefficient. Indicates the second accelerating stress coefficient. This represents the Boltzmann constant.

[0101] In one example, the multi-source dataset of wind turbine status includes degradation cause data and degradation result data. The degradation cause data includes temperature monitoring data, stress intensity, and corrosive medium content, while the degradation result data includes crack propagation length, residual material strength, and tensile strength.

[0102] In one example, the degradation integration module 12 is used to update the degradation parameters in the first wind power accelerated degradation model in real time using a Bayesian function based on prior information and sample information in the multi-source dataset of wind power status, so as to obtain the posterior distribution parameters of the degradation parameters; it iteratively obtains the maximum likelihood estimate of the degradation parameters in the first wind power accelerated degradation model through the EM algorithm, and calculates the update values ​​of the remaining parameters in the first wind power accelerated degradation model corresponding to the maximum likelihood estimate according to the parameter update function.

[0103] In one example, the principal accelerating stress characterizes the core stress that directly drives crack propagation. The principal accelerating stress includes wind load and vibration load. The wind load and vibration load are weighted and calculated to obtain the tensile stress quantification value corresponding to the principal accelerating stress. The secondary accelerating stress characterizes the environmental stress that indirectly affects crack propagation. The secondary accelerating stress includes the blade service environment temperature, relative humidity, and salt spray concentration.

[0104] Please see Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 4 As shown, the electronic device 2 may include: at least one processor 21, at least one network interface 24, user interface 23, memory 25, and at least one communication bus 22.

[0105] The communication bus 22 is used to enable communication between these components.

[0106] The user interface 23 may include a display screen and a camera. Optionally, the user interface 23 may also include a standard wired interface and a wireless interface.

[0107] The network interface 24 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0108] The processor 21 may include one or more processing cores. The processor 21 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 25, and by calling data stored in the memory 25. Optionally, the processor 21 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 21 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 21 and may be implemented as a separate chip.

[0109] The memory 25 may include random access memory (RAM) or read-only memory. Optionally, the memory 25 may include non-transitory computer-readable storage medium. The memory 25 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 25 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 25 may also be at least one storage device located remotely from the aforementioned processor 21. Figure 4 As shown, the memory 25, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for a multi-source data-driven wind power equipment life prediction method based on acceleration effects.

[0110] exist Figure 4In the electronic device 2 shown, the user interface 23 is mainly used to provide an input interface for the user and obtain the user input data; while the processor 21 can be used to call the application stored in the memory 25, which is a multi-source data-driven wind power equipment life prediction method based on acceleration effect. When executed by one or more processors, the electronic device performs one or more methods as described in the above embodiments.

[0111] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-source data-driven method for predicting the lifespan of wind power equipment based on acceleration effects, characterized in that, The method includes: Collect operating environment data and degradation fault information of wind power equipment, and construct a basic performance degradation model of wind power equipment based on the operating environment data and degradation fault information; Mechanism analysis was performed on the blade crack propagation data in the degradation fault information to obtain the main accelerating stress and secondary accelerating stress of the wind power equipment. Based on the main accelerating stress, the secondary accelerating stress and the comprehensive acceleration equation, a first wind power equipment accelerated degradation model was constructed. Based on the degradation fault information, a multi-source dataset of wind power equipment status is constructed. According to the multi-source dataset of wind power equipment status, Bayesian function and EM algorithm, the first wind power accelerated degradation model is updated to obtain the second wind power accelerated degradation model. Based on the multi-condition hypothesis function of the degradation process and the second accelerated degradation model of wind power equipment, a multi-stage degradation process model of wind power equipment is constructed to predict the remaining life of wind power equipment.

2. The method as described in claim 1, characterized in that, The construction of a basic performance degradation model for wind power equipment based on the operating environment data and the degradation fault information specifically includes: Based on the degradation failure mechanism and failure mode in the operating environment data and degradation fault information, the main influencing indicators of wind power equipment performance degradation are determined; Based on the main influencing indicators, the nonlinear Gamma correction function, and the blade crack propagation data in the degradation fault information, a basic performance degradation model for wind power equipment is constructed.

3. The method as described in claim 1, characterized in that, After constructing the basic performance degradation model for wind power equipment, the following is also included: All unknown parameters in the basic performance degradation model of the wind power equipment are given initial values, wherein the initial values ​​are derived from the prior distribution, maximum likelihood estimation and method moments; The unknown parameters are divided into multiple parameter groups according to conditional independence. In each iteration, an update operation is performed on each parameter group in sequence. The update operation includes, under the condition that the current values ​​of other parameters are fixed, sampling from the conditional posterior distribution of the first parameter group to update the parameter group; based on the updated first parameter group, sampling from the conditional posterior distribution of the second parameter group to update the parameter group; and sampling from the corresponding conditional posterior distribution and updating the remaining parameter groups in a preset order until all parameter groups have completed one update cycle. Repeat the update operation until the preset number of iterations is reached. Extract the posterior distribution, posterior mean, posterior standard deviation and confidence interval of each unknown parameter from the final posterior sample. Use the posterior sample to perform interval evaluation of key quantities in the wind power equipment life prediction method.

4. The method as described in claim 3, characterized in that, The process of using the posterior sample to perform interval evaluation of key quantities in the wind power equipment life prediction method specifically includes: The basic performance degradation model of the wind power equipment is calculated based on the MAE index to obtain the mean absolute error of the basic performance degradation model of the wind power equipment. The basic performance degradation model of the wind power equipment is calculated based on the MSE index to obtain the mean square error of the basic performance degradation model of the wind power equipment. The basic performance degradation model of the wind power equipment is calculated based on the MAPE index, and the average absolute error percentage of the basic performance degradation model of the wind power equipment is calculated. The distance between the training sample bias and the test sample bias in the basic performance degradation model of the wind power equipment is calculated based on the DIC criterion.

5. The method as described in claim 1, characterized in that, The expression functions corresponding to the first wind power accelerated degradation model include: in, for Length of blade crack at any given time This represents the comprehensive acceleration stress index value. This represents the quantified value of the tensile stress, which indicates the principal accelerating stress. Denotes the Gamma distribution function, with shape parameter . , Represents the basic shape function, with the scale parameter being... , Represents the comprehensive acceleration equation, Represents Boltzmann's constant. Indicates the secondary acceleration stress coefficient. Indicates the principal accelerating stress coefficient. K This indicates the temperature in the operating environment of the wind power equipment. Indicates tensile stress. This represents the comprehensive acceleration stress index value. E a Indicates activation energy. Indicates the first accelerating stress coefficient. Indicates the second accelerating stress coefficient. This represents the Boltzmann constant.

6. The method as described in claim 1, characterized in that, The multi-source dataset of wind power equipment status includes degradation cause data and degradation result data. The degradation cause data includes temperature monitoring data, stress intensity, and corrosive medium content. The degradation result data includes crack propagation length, material residual strength, and tensile strength.

7. The method as described in claim 1, characterized in that, The step of updating the first wind turbine accelerated degradation model based on the multi-source dataset of wind turbine status, Bayesian function, and EM algorithm specifically includes: Based on the prior information and sample information in the multi-source dataset of the wind power equipment status, the degradation parameters in the first wind power equipment accelerated degradation model are updated in real time through a Bayesian function to obtain the posterior distribution parameters of the degradation parameters. The maximum likelihood estimate of the degradation parameters in the first wind power accelerated degradation model is obtained by iteratively using the EM algorithm, and the update values ​​of the remaining parameters in the first wind power accelerated degradation model corresponding to the maximum likelihood estimate are calculated according to the parameter update function.

8. The method as described in claim 1, characterized in that, The primary accelerating stress characterizes the core stress that directly drives crack propagation. The primary accelerating stress includes wind load and vibration load. The wind load and vibration load are weighted and calculated to obtain the tensile stress quantification value corresponding to the primary accelerating stress. The secondary accelerating stress characterizes the environmental stress that indirectly affects crack propagation. The secondary accelerating stress includes the blade service environment temperature, relative humidity, and salt spray concentration.

9. A multi-source data-driven wind power equipment life prediction system based on acceleration effect, characterized in that, The multi-source data-driven wind power equipment life prediction system (1) includes an information acquisition module (11), a degradation integration module (12), and a life prediction module (13), wherein, The information acquisition module (11) is used to collect operating environment data and degradation fault information of wind power equipment, and to construct a basic performance degradation model of wind power equipment based on the operating environment data and the degradation fault information. The degradation integration module (12) is used to perform mechanism analysis on the blade crack propagation data in the degradation fault information, obtain the main accelerating stress and secondary accelerating stress of the wind power equipment respectively, and construct a first wind power equipment accelerated degradation model based on the main accelerating stress, the secondary accelerating stress and the comprehensive acceleration equation; construct a wind power equipment state multi-source dataset based on the degradation fault information, and update the first wind power equipment accelerated degradation model according to the wind power equipment state multi-source dataset, Bayesian function and EM algorithm to obtain a second wind power equipment accelerated degradation model; The life prediction module (13) is used to construct a multi-stage degradation process model of wind power equipment based on the multi-condition assumption function of the degradation process and the second wind power accelerated degradation model, so as to predict the remaining life of wind power equipment.

10. An electronic device, characterized in that, The device includes a processor (21), a memory (25), a user interface (23), and a network interface (24). The memory (25) is used to store instructions. The user interface (23) and the network interface (24) are used to communicate with other devices. The processor (21) is used to execute the instructions stored in the memory (25) to cause the electronic device (2) to perform the method as described in any one of claims 1-8.

Images