A method for predicting the remaining life of an offshore wind turbine considering the influence of typhoon

Abstract

The application relates to a kind of offshore wind turbine remaining life prediction methods considering the influence of typhoon, comprising: based on the influence mechanism of typhoon on the degradation of unit component, build component failure rate increment model under typhoon impact;Combined with the operation state monitoring data of unit, build the degradation and impact correlation model of component operating state, quantify the ability to withstand external impact;Based on the time-varying randomness characteristics of the performance degradation of unit caused by random impact of typhoon, build offshore wind turbine component time-varying failure rate model considering the dependence of degradation and impact, and based on the model, build unit component remaining life prediction model under typhoon impact scenario, adopt particle swarm optimization algorithm to optimize parameters with minimum prediction error, obtain optimized remaining life prediction model;Based on real-time data, correct the optimized remaining life prediction model and realize remaining life prediction.Compared with prior art, the application has the advantages of high life prediction accuracy.

Description

Technical Field

[0001] This invention relates to the field of offshore wind turbine life prediction, and in particular to a method for predicting the remaining life of offshore wind turbines that takes into account the impact of typhoons. Background Technology

[0002] Offshore wind power is a crucial force in accelerating my country's "3060" dual-carbon strategy, and the high reliability of offshore wind turbines is a key factor for the sustainable development of large-scale offshore wind power. Accurately predicting the remaining lifespan of offshore wind turbines has significant theoretical and engineering application value for reducing the risk of sudden failures and making decisions on high-reliability operation and efficient maintenance. Coastal areas in my country where large-scale, clustered offshore wind power development is severely affected by severe weather such as typhoons. The strong winds and giant waves brought by typhoons pose a significant challenge to the accurate prediction of the remaining lifespan of offshore wind turbines.

[0003] Existing research on offshore wind turbine life prediction mainly suffers from the following two problems: 1) Previous studies on wind turbine remaining life prediction typically rely on the natural degradation process of the turbine, assuming that the turbine's performance gradually deteriorates over time, neglecting the random impact of objectively existing offshore typhoon weather, making it difficult to accurately describe the actual degradation evolution of the turbine under harsh offshore operating conditions. 2) In the context of multi-source, multi-dimensional offshore operation and maintenance big data, accurately characterizing the correlation between random impacts and degradation states of the turbine, precisely modeling the turbine failure process, and quantifying the uncertainty of remaining life prediction still need to be addressed. Summary of the Invention

[0004] The purpose of this invention is to provide a method for predicting the remaining life of offshore wind turbines that takes into account the impact of typhoons. Under typhoon conditions, this method minimizes the mean square error of the prediction and achieves interactive linkage between the status of turbine components and the stochastic degradation model to obtain accurate predictions of the life of offshore wind turbines.

[0005] The objective of this invention can be achieved through the following technical solutions:

[0006] A method for predicting the remaining life of offshore wind turbines considering the impact of typhoons includes the following steps:

[0007] Based on the mechanism of typhoon impact on unit component degradation, the effects of direct and indirect impact on component failure rate are analyzed, and an incremental model of component failure rate under typhoon impact is constructed.

[0008] To address the issue of quantifying the correlation between component degradation process and impact, a correlation model between component operating status degradation and impact is constructed by combining unit operating status monitoring data to quantify the ability to withstand external impacts.

[0009] Based on the time-varying randomness of the performance degradation of the unit caused by the random impact of typhoons, a time-varying failure rate model of offshore wind turbine components is constructed by combining the component failure rate increment model under typhoon impact and the degradation-impact correlation model.

[0010] Based on the time-varying failure rate model of offshore wind turbine components, a prediction model for the remaining life of turbine components is constructed under the typhoon impact scenario. The particle swarm optimization algorithm is used to optimize the parameters of the component failure rate increment model and the degradation and impact correlation model under the typhoon impact with the minimum prediction error, so as to obtain the optimized remaining life prediction model.

[0011] Based on the optimized remaining life prediction model, the remaining life prediction model is corrected and optimized by combining typhoon impact monitoring data and real-time unit operation status monitoring data, and the remaining life prediction is realized.

[0012] The construction of the component failure rate increment model under typhoon impact includes the following steps:

[0013] Obtain typhoon data from typhoon-prone areas;

[0014] Constructing a Batts typhoon wind field model;

[0015] Assuming that the number of times an offshore wind turbine is hit by a typhoon during operation follows a Poisson process with parameter λ, and that the typhoon wind field model remains unchanged before and after a typhoon passes through the wind farm, the Monte Carlo sampling method is used to obtain the impact probability of each typical typhoon and a typhoon stochastic impact model is established.

[0016] Based on the typhoon stochastic impact model, the impact of direct and indirect impacts on component failure rate is analyzed, and an incremental model of component failure rate under typhoon impact is constructed.

[0017] The typhoon stochastic impact model is as follows:

[0018]

[0019]

[0020]

[0021]

[0022] Where P(N(t)=n) represents the probability of a typhoon making n impacts within the interval (0, t), λ is the typhoon arrival rate, and f G (t si Let t be the time t required for the i-th impact to reach its first destination. si The probability density, Γ(i) is the gamma function value; V RmaxLet R be the gradient wind speed at the maximum wind radius, K be an empirical parameter ranging from 6.93 to 6.97, Δp be the pressure difference at the typhoon center, and R be the gradient wind speed at the maximum wind radius. max The radius of maximum wind speed is given by f, where f is the Coriolis parameter; v represents the instantaneous wind speed of the typhoon. s denoted as typhoon speed; r is the distance from any position of the typhoon to its center; x is an empirical parameter with a value ranging from 0.5 to 0.7.

[0023] The incremental model for component failure rate under typhoon impact is as follows:

[0024]

[0025] in, Let be the correlation function between component degradation and impact, representing the state of component k in operation. The impact of shocks on the incremental failure rate, t si and t ei α and ε represent the initial arrival time and the end time of the typhoon's impact on the unit, respectively. α and ε are the correction coefficients for the direct or indirect impact of the typhoon on the components, and v is the instantaneous wind speed of the typhoon.

[0026] Distinguishing between the effects of direct and indirect impacts, the incremental model for component failure rate under direct impact is expressed as follows:

[0027]

[0028]

[0029] The incremental model of component failure rate under indirect impact is expressed as follows:

[0030]

[0031]

[0032] Where, Δλ sk1i α represents the failure rate of external components of the generator unit caused by direct impact from a typhoon. k1 ε k1 Here, represents the correction factor for direct typhoon impact, w represents the wind load perpendicular to the wind turbine surface, ρ represents the air density, and C represents the wind load. s S is the shape factor. α Δλ is the projected area on the plane perpendicular to the wind direction. sk2i α represents the failure rate of internal components of the generator set caused by the indirect impact of a typhoon. k2 ε k2 These are the correction factors for indirect typhoon impact, where p is the power of the wind turbine, and C is the power of the wind turbine. p λ is the wind energy conversion efficiency coefficient. a For the tip speed ratio of the fan blades, β aThe pitch angle of the wind turbine blades, r a The radius of the wind turbine blades.

[0033] The construction of the degradation and impact correlation model based on component operating status includes the following steps:

[0034] Obtain historical operational monitoring data of offshore wind turbines;

[0035] Principal component clustering analysis was used to extract the health status characteristics of offshore wind turbine components under different operating conditions, and different component states were classified according to the difference values ​​between other states and the healthy state.

[0036] Based on the component state classification results under different working conditions and the historical state evolution process, a stochastic state model is established based on the Markov chain state transition process to describe the historical state transition process of the component, and a degradation and impact correlation model is obtained.

[0037] The degradation-impact correlation model is expressed as follows:

[0038]

[0039] R = X'X / n q

[0040] Y = XU = [Y1, Y2, ..., Y] q ]

[0041]

[0042]

[0043]

[0044] in, The operating state of component k. Let μ be the correlation function between component degradation and impact, and μ be the state variable correction coefficient. The input variable matrix X, after standardization of SCADA data related to a certain component of the wind turbine, contains q-dimensional initial parameters, with each dimension containing n parameters. q Let X' represent the transpose of X, and let the characteristic equation of the correlation coefficient matrix R have q eigenvalues, where λ1≥λ2≥…≥λ q U = (U1, U2, ..., U q Let Y be the eigenvectors corresponding to the eigenvalues, and let Y be the principal components, sorted in descending order of principal component variance. q λ is the qth principal component. ji Let ω be the eigenvalue corresponding to the i-th principal component under working condition j, and δ be the cumulative variance contribution rate. The principal component analysis results for each working condition j are selected based on the cumulative variance contribution rate being greater than δ. jOne principal component, π(t) g ) represents monitoring point t g The probability distribution of each state of the wind turbine components at any given time, A is the state transition matrix, Δt is an arbitrary time interval, t0 is the state transition time step length, and S... k This is the vector matrix of quantized values ​​for each state.

[0045] Assume the failure rate λ of system component k k Failure rate λ due to natural degradation 0k and the cumulative failure rate increment Δλ of components caused by multiple typhoon impacts ski (t) is composed of components whose lifespan is a non-negative continuous random variable, and whose natural degradation process is described by a Weibull distribution. Combining the component failure rate increment model under typhoon impact and the degradation-impact correlation model, a time-varying failure rate model for offshore wind turbine components considering the dependence of degradation and impact is obtained:

[0046]

[0047]

[0048]

[0049] Where i represents the number of typhoon impacts; N(t) represents the cumulative number of typhoon impacts up to time t; D i (t) represents the impact amplitude at time t during the i-th typhoon; Let be the correlation function between component degradation and impact, representing the state of component k in operation. The impact of shocks on the incremental failure rate; t si and t ei These represent the initial arrival time and the end time of the typhoon's impact on the generator unit, respectively; β is the shape parameter, and η is the scale parameter.

[0050] The remaining life prediction model for unit components that considers both degradation and impact is as follows:

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057] Among them, component reliability R k(t) represents the probability that the normal operating time of the component is greater than t, and P represents the probability of the event. Let n represent the expected failure rate function of components considering the impact of random typhoons. z Let Z represent the total number of typhoon impact scenarios. A single typhoon impact scenario is a set of the typhoon's maximum wind speed radius, moving speed, central pressure difference, wind speed at the maximum wind speed radius, and typhoon's moving direction. Let z be a variable representing the typhoon impact scenario, and T be the moment the component fails. The remaining lifespan of the component at time t is... Let x be the expected value of the remaining lifespan, and let x be the remaining lifespan time variable. Represents the expected value of remaining lifespan. The distribution function, To determine the expected reliability value function of components under historical typhoon impacts, This represents the reliability function considering a random typhoon impact scenario. for The derivative of t s To make the probability density function of the expected remaining lifetime at t=0 The moment corresponding to the maximum value, This is the reliability failure threshold.

[0058] By statistically analyzing the lifespan frequency of a component of a wind turbine in an offshore wind farm, the parameters of a natural degradation failure rate model are estimated using the maximum likelihood estimation method. Based on the natural degradation model, n is then extracted. w Taiwanese generator set performs parameter estimation for incremental component failure rate model under typhoon impact:

[0059]

[0060]

[0061]

[0062]

[0063]

[0064] Where β is the shape parameter, η is the scale parameter, and f 0k Let t be the probability density function of the Weibull distribution, and L(β,η) be the likelihood function. z Let n be the sample values ​​of component lifespan, lnL(β,η) be the log-likelihood function, and n be the value of component lifespan. c For the sample size, U j Let be the root mean square error of unit j. and y j (t gi ) represent the monitoring points t and j of unit j respectively. giThe estimated and actual remaining lifetime at time n g n is the number of monitoring points. w This represents the number of experimental fan samples drawn. For n w The unit's root mean square error.

[0065] Compared with the prior art, the present invention has the following beneficial effects:

[0066] (1) This invention takes into account the impact of random typhoon impacts on the prediction results and establishes a time-varying failure rate model and an incremental failure rate model for offshore wind turbine components that are dependent on degradation and impacts. These models can accurately describe the evolution of actual degradation of the units under harsh offshore operating conditions.

[0067] (2) The present invention establishes a correlation model between component degradation and impact, accurately describes the relationship between random impact and degradation state, and uses actual component monitoring data to correct the remaining life prediction model, so that the component degradation prediction process is more consistent with the actual degradation process and the life prediction accuracy is improved. Attached Figure Description

[0068] Figure 1 This is a flowchart of the method of the present invention;

[0069] Figure 2 A schematic diagram illustrating the evolution of failure rates of offshore wind turbine components under typhoon impact;

[0070] Figure 3 A model of the Batts typhoon wind field and a schematic diagram of its impact on offshore wind farms;

[0071] Figure 4 A schematic diagram showing the state division of components;

[0072] Figure 5 A diagram showing the impact of operating condition classification on the results of principal component cluster analysis.

[0073] Figure 6 A schematic diagram of the remaining lifespan of unit components under random typhoon impact scenarios;

[0074] Figure 7 A flowchart for parameter estimation in particle swarm optimization algorithm;

[0075] Figure 8 The results are the generator current variation and condition assessment.

[0076] Figure 9 Quantify the change of the expected value of the generator's random state over operating time;

[0077] Figure 10 The probability density distribution of the expected remaining lifespan of the generator;

[0078] Figure 11 The reliability change process of three lifetime prediction methods;

[0079] Figure 12 This refers to the reliability change process corresponding to the correction process of Method 3;

[0080] Figure 13 Comparison of remaining lifetime prediction results from three methods;

[0081] Figure 14 The probability density curves of the expected remaining lifetime under different typhoon arrival rates λ;

[0082] Figure 15 The impact of the number of random typhoon impacts on prediction error. Detailed Implementation

[0083] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0084] A method for predicting the remaining life of offshore wind turbines considering the impact of typhoons, such as Figure 1 As shown, it includes the following steps:

[0085] 1) Based on the mechanism of typhoon impact on unit component degradation, analyze the impact of direct and indirect impact on component failure rate, and construct an incremental model of component failure rate under typhoon impact.

[0086] 1-1) Obtain typhoon data in typhoon-prone areas;

[0087] 1-2) Construct the Batts typhoon wind field model; the typhoon wind field model and its impact on offshore wind farms are illustrated in the diagram below. Figure 3 As shown;

[0088] 1-3) Assume that the number of times an offshore wind turbine is hit by a typhoon during operation follows a Poisson process with parameter λ. The typhoon wind field model remains unchanged before and after the typhoon passes through the wind farm. The Monte Carlo sampling method is used to obtain the impact probability of each typical typhoon and a typhoon random impact model is established.

[0089] The typhoon stochastic impact model is as follows:

[0090]

[0091]

[0092]

[0093]

[0094] Where P(N(t)=n) represents the probability of a typhoon making n impacts within the interval (0, t), λ is the typhoon arrival rate, and f G (t si Let t be the time t required for the i-th impact to reach its first destination. si The probability density, Γ(i) is the gamma function value; V Rmax Let R be the gradient wind speed at the maximum wind radius, K be an empirical parameter ranging from 6.93 to 6.97, Δp be the pressure difference at the typhoon center, and R be the gradient wind speed at the maximum wind radius. max The radius of maximum wind speed is given by f, which is the Coriolis parameter; v represents the instantaneous wind speed of the typhoon. s denoted as typhoon speed; r is the distance from any position of the typhoon to its center; x is an empirical parameter with a value ranging from 0.5 to 0.7.

[0095] 1-4) Based on the typhoon stochastic impact model, analyze the impact of direct and indirect impacts on component failure rates, and construct an incremental model of component failure rates under typhoon impact:

[0096]

[0097] in, Let be the correlation function between component degradation and impact, representing the state of component k in operation. The impact of shocks on the incremental failure rate, t si and t ei α and ε represent the initial arrival time and the end time of the typhoon's impact on the unit, respectively. α and ε are the correction coefficients for the direct or indirect impact of the typhoon on the components, and v is the instantaneous wind speed of the typhoon.

[0098] Distinguishing between the effects of direct and indirect impacts, the incremental model for component failure rate under direct impact is expressed as follows:

[0099]

[0100]

[0101] The incremental model of component failure rate under indirect impact is expressed as follows:

[0102]

[0103]

[0104] Where, Δλ sk1i α represents the failure rate of external components of the generator unit caused by direct impact from a typhoon. k1 ε k1 Here, represents the correction factor for direct typhoon impact, w represents the wind load perpendicular to the wind turbine surface, ρ represents the air density, and C represents the wind load. sS is the shape factor. α Δλ is the projected area on the plane perpendicular to the wind direction. sk2i α represents the failure rate of internal components of the generator set caused by the indirect impact of a typhoon. k2 ε k2 These are the correction factors for indirect typhoon impact, where p is the power of the wind turbine, and C is the power of the wind turbine. p λ is the wind energy conversion efficiency coefficient. a For the tip speed ratio of the fan blades, β a The pitch angle of the wind turbine blades, r a The radius of the wind turbine blade.

[0105] This embodiment takes a certain offshore wind farm in my country as an example. The offshore wind farm includes 36 3MW wind turbines, which are planned and laid out in 4 rows and 9 columns. The north-south spacing of the wind turbines is 0.5km and the east-west spacing is 1km.

[0106] Based on historical typhoon yearbooks and statistical data on historical typhoon events monitored by the wind farm, the estimated parameters of the typhoon stochastic impact model are shown in Table 1.

[0107] Table 1. Estimated values ​​of typhoon model parameters

[0108]

[0109] 2) To address the issue of quantifying the correlation between component degradation process and impact, a correlation model between component operating status degradation and impact is constructed by combining unit operating status monitoring data to quantify the ability to withstand external impacts.

[0110] 2-1) Obtain historical operational monitoring data of offshore wind turbines;

[0111] 2-2) Principal component clustering analysis was used to extract the health status characteristics of offshore wind turbine components under different operating conditions, and different component states were classified according to the difference values ​​between other states and the healthy state.

[0112] 2-3) Based on the component state classification results under different working conditions and the historical state evolution process, a stochastic state model is established based on the Markov chain state transition process to describe the historical state transition process of the component, and a degradation and impact correlation model is obtained.

[0113] The stochastic state model of the component is described by a Markov chain state transition process. Under the random disturbances of typhoon weather, as the random failure states of unit components intensify, the components become increasingly sensitive to external shocks, and their shock-bearing capacity gradually decreases. Correlation function between component degradation and shock. It is a monotonically increasing function.

[0114] The degradation-impact correlation model is expressed as follows:

[0115]

[0116] R = X'X / n q

[0117] Y = XU = [Y1, Y2, ..., Y] q ]

[0118]

[0119]

[0120]

[0121] in, The operating state of component k. Let μ be the correlation function between component degradation and impact, and μ be the state variable correction coefficient. The input variable matrix X, after standardization of SCADA data related to a certain component of the wind turbine, contains q-dimensional initial parameters, with each dimension containing n parameters. q Let X' represent the transpose of X, and let the characteristic equation of the correlation coefficient matrix R have q eigenvalues, where λ1≥λ2≥…≥λ q U = (U1, U2, ..., U q Let Y be the eigenvectors corresponding to the eigenvalues, and let Y be the principal components, sorted in descending order of principal component variance. q Y1 is the q-th principal component, and Y2 is the first principal component, possessing the largest variance and capable of explaining most of the information in the data. λ ji Let ω be the eigenvalue corresponding to the i-th principal component under working condition j, and δ be the cumulative variance contribution rate. The principal component analysis results for each working condition j are selected based on the cumulative variance contribution rate being greater than δ. j One principal component, π(t) g ) represents monitoring point t g The probability distribution of each state of the wind turbine components at any given time, A is the state transition matrix, Δt is an arbitrary time interval, t0 is the state transition time step length, and S... k This is the vector matrix of quantized values ​​for each state.

[0122] This embodiment takes a generator with a high failure rate and serious consequences as an example, and divides the generator's state into four states: healthy, slightly abnormal, abnormal, and faulty. Figure 4 As shown, the initial probability distribution is π(t) g =0) =

[1000] .

[0123] Based on the generator-related SCADA monitoring parameters and the initial principal component clustering analysis results, the operating conditions were divided into three categories under low, medium, and high wind speeds. Principal component clustering analysis was performed on the data for each of the three operating conditions, and principal components were selected based on a cumulative variance contribution rate greater than 90%. The clustering results of generator health characteristics under different operating conditions after principal component analysis are shown below. Figure 5 As shown.

[0124] 3) Based on the time-varying randomness of the performance degradation of the unit caused by the random impact of typhoons, and combining the component failure rate increment model under typhoon impact and the degradation-impact correlation model, a time-varying failure rate model of offshore wind turbine components considering the dependence of degradation and impact is constructed.

[0125] Assume the failure rate λ of system component k k Failure rate λ due to natural degradation 0k and the cumulative failure rate increment Δλ of components caused by multiple typhoon impacts ski (t) is composed of components whose lifespan is a non-negative continuous random variable, and whose natural degradation process is described by a Weibull distribution. Combining the component failure rate increment model under typhoon impact and the degradation-impact correlation model, a time-varying failure rate model for offshore wind turbine components considering the dependence of degradation and impact is obtained:

[0126]

[0127]

[0128]

[0129] Where i represents the number of typhoon impacts; N(t) represents the cumulative number of typhoon impacts up to time t; D i (t) represents the impact amplitude at time t during the i-th typhoon; Let be the correlation function between component degradation and impact, representing the state of component k in operation. The impact of shocks on the incremental failure rate; t si and t ei These represent the initial arrival time and the end time of the typhoon's impact on the generator unit, respectively; β is the shape parameter, and η is the scale parameter.

[0130] A schematic diagram illustrating the evolution of failure rates of offshore wind turbine components under typhoon impact is shown below. Figure 2 As shown.

[0131] 4) Based on the time-varying failure rate model of offshore wind turbine components, a prediction model for the remaining life of turbine components is constructed under the typhoon impact scenario. The particle swarm optimization algorithm is used to optimize the parameters of the component failure rate increment model and the degradation and impact correlation model under the typhoon impact with the minimum prediction error, so as to obtain the optimized remaining life prediction model.

[0132] Based on the historical typhoon distribution probability, Monte Carlo sampling was used to randomly sample and establish N(t) random typhoon impact scenarios within the range (0, t). Each typhoon impact scenario is a set of the typhoon's maximum wind speed radius, moving speed, central pressure difference, wind speed at the maximum wind speed radius, and typhoon's moving direction. Based on historical unit condition monitoring data, a stochastic state model of unit components based on Markov transition processes was constructed to obtain the historical state transition processes of the components. This was combined with any operating time t. g Previous actual typhoon weather data and component operation status monitoring data were used to correct the typhoon impact scenario and component status transition process, thereby obtaining a prediction model for the remaining life of unit components that considers the interdependence of degradation and impact:

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139] Among them, component reliability R k (t) represents the probability that the normal operating time of the component is greater than t, and P represents the probability of the event. Let n represent the expected failure rate function of components considering the impact of random typhoons. z Let Z represent the total number of typhoon impact scenarios, z represent the typhoon impact scenario variable, and T represent the time when the component fails. The remaining lifespan of the component at time t is... Let x be the expected value of the remaining lifespan, and let x be the remaining lifespan time variable. Represents the expected value of remaining lifespan. The distribution function, To determine the expected reliability value function of components under historical typhoon impacts, This represents the reliability function considering a random typhoon impact scenario. for The derivative of t s To make the probability density function of the expected remaining lifetime at t=0 The moment corresponding to the maximum value, This is the reliability failure threshold.

[0140] By statistically analyzing the lifespan frequency of a component of a wind turbine in an offshore wind farm, the parameters of a natural degradation failure rate model are estimated using the maximum likelihood estimation method. Based on the natural degradation model, n is then extracted. w Taiwanese generator set performs parameter estimation for incremental component failure rate model under typhoon impact:

[0141]

[0142]

[0143]

[0144]

[0145]

[0146] Where β is the shape parameter, η is the scale parameter, and f 0k Let t be the probability density function of the Weibull distribution, and L(β,η) be the likelihood function. z Let n be the sample values ​​of component lifespan, lnL(β,η) be the log-likelihood function, and n be the value of component lifespan. c For the sample size, U j Let be the root mean square error of unit j. and y j (t gi ) represent the monitoring points t and j of unit j respectively. gi The estimated and actual remaining lifetime at time n g n is the number of monitoring points. w U represents the number of experimental fan samples drawn, where n is the number of samples. w The unit's root mean square error.

[0147] A schematic diagram of the remaining lifespan of generator components under random typhoon impact scenarios is shown below. Figure 6 As shown. Figure 7 The flowchart shows the particle swarm optimization algorithm. In this embodiment, the initial number of particles is set to 40, the learning factor is 1.494, and the maximum particle velocity is 0.8. The parameter estimation results from the particle swarm optimization algorithm are shown in Table 2.

[0148] Table 2. Parameter estimation results for the remaining life prediction model.

[0149]

[0150] 5) Based on the optimized remaining life prediction model, the remaining life prediction model is corrected and optimized by combining typhoon impact monitoring data and real-time unit operation status monitoring data, and the remaining life prediction is realized.

[0151] Figure 8This indicates the generator status assessment results for a period of time prior to the unit shutdown due to abnormal generator current. Figure 9 This represents the change in the expected value of the generator's stochastic state quantization at any monitoring point under different initial states as a function of operating time, based on the Markov state transition process.

[0152] Figure 10 This is a graph showing the probability density distribution function of the expected value of the generator's remaining lifespan, taking into account the random impact of a typhoon.

[0153] Taking the prediction of the remaining life of a generator unit as an example, the effectiveness of the model is verified by combining historical monitoring data. Three methods are compared: Method 1: a natural degradation prediction model that does not consider the impact of typhoons; Method 2: a prediction model considering the dependence of degradation on typhoon impact in a random typhoon impact scenario; and Method 3: a prediction model based on Method 2, modified by incorporating actual typhoon impacts. The model is revised every 50 days based on the historical intervals between typhoons and the duration of each state.

[0154] Figure 11 The results show that Method 1, which does not consider the impact of typhoons and only takes into account the natural degradation process of unit components, yields the most optimistic prediction results, overestimating the reliability and remaining life of the components. Compared to Method 1, Methods 2 and 3 consider the impact of typhoons on the degradation process of unit components, and the reliability evolution process is closer to reality. The root mean square error (RMSE) of the remaining life prediction is reduced by 17.4% and 25.1%, respectively. Compared with Method 2, Method 3 further reduces the prediction error by 7.7% by incorporating a correction model based on actual typhoon impacts.

[0155] Figure 12 It was stated that the unit had experienced the impact of Typhoon Haikui and Typhoon Bolaven during its operation, which led to the correction process of the model in Method 3.

[0156] Figure 13 The results show that as the amount of monitoring data increases, the prediction results of Method 3 are closer to the actual lifespan value, and the prediction accuracy is improved.

[0157] The impact of typhoon arrival rate λ on remaining life prediction and the impact of random typhoon impact frequency on prediction error are analyzed separately. Figure 14 The data shows that as the frequency of typhoon impacts increases, the peak value of the probability density curve shifts to the upper left, and the curve shape becomes narrower and the probability distribution becomes more concentrated. This indicates that the increase in the frequency of typhoon impacts accelerates the rate at which the reliability of unit components decreases, the expected value of remaining life prediction decreases, the probability distribution becomes more concentrated, and the uncertainty of remaining life prediction decreases. Figure 15The results show that the Remaining Life Prediction Error (RMSE) is minimized when the number of typhoon impacts matches the actual number of occurrences, and corrections to monitoring data are even more beneficial in reducing the RMSE. When the number of typhoon impacts deviates from the actual number of typhoon occurrences, the RMSE prediction error caused by this deviation is reduced.

[0158] As can be seen from this embodiment, the method proposed in this invention is effective and feasible, and can provide a reference for predicting the remaining life of offshore wind turbines considering the impact of typhoons.

[0159] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for predicting the remaining life of offshore wind turbines considering the impact of typhoons, characterized in that, Includes the following steps: Based on the mechanism of typhoon-induced degradation of unit components, the impact of direct and indirect impacts on component failure rates is analyzed, and an incremental model of component failure rates under typhoon impact is constructed. The incremental model of component failure rates under typhoon impact is as follows: in, Let be the correlation function between component degradation and impact, representing the component k In running status The impact of shocks on the increase in failure rate and These represent the initial arrival time and the end time of the typhoon's impact on the generator unit, respectively. , This is a correction factor for the direct or indirect impact of typhoons on components. v This refers to the instantaneous wind speed of a typhoon. To address the issue of quantifying the correlation between component degradation process and impact, a correlation model between component operating status degradation and impact is constructed by combining unit operating status monitoring data to quantify the ability to withstand external impacts. Based on the time-varying randomness of turbine performance degradation caused by typhoon impacts, and combining the incremental component failure rate model under typhoon impacts and the degradation-impact correlation model, a time-varying failure rate model for offshore wind turbine components considering the dependence of degradation and impact is constructed; wherein, the degradation-impact correlation model is expressed as: in, For components k The running status, This is a function relating component degradation to impact. The state variable correction coefficients are the input variable matrix after standardization of SCADA data related to a certain component of the wind turbine. Include Initial parameters for each dimension, each parameter contains One sample, express The transpose of the correlation coefficient matrix The characteristic equations are eigenvalues ,in, , The eigenvectors corresponding to the eigenvalues. Principal components are sorted in descending order of their variance. For the first principal component, For working conditions Next The eigenvalues ​​corresponding to each principal component For the cumulative variance contribution rate, each type of working condition Principal component analysis results are arranged according to cumulative variance contribution rate greater than Select Principal components, For monitoring points The probability distribution of each state of the wind turbine components at any given time. Here is the state transition matrix. For any time interval, Due to the long state transition time step, S k The vector matrix of quantized values ​​for each state; Based on the time-varying failure rate model of offshore wind turbine components, a prediction model for the remaining life of turbine components is constructed under the typhoon impact scenario. The particle swarm optimization algorithm is used to optimize the parameters of the component failure rate increment model and the degradation and impact correlation model under the typhoon impact with the minimum prediction error, so as to obtain the optimized remaining life prediction model. Based on the optimized remaining life prediction model, the remaining life prediction model is corrected and optimized by combining typhoon impact monitoring data and real-time unit operation status monitoring data, and the remaining life prediction is realized.

2. The method for predicting the remaining life of offshore wind turbines considering the impact of typhoons according to claim 1, characterized in that, The construction of the component failure rate increment model under typhoon impact includes the following steps: Obtain typhoon data from typhoon-prone areas; Constructing a Batts typhoon wind field model; Assume that the number of times an offshore wind turbine is hit by a typhoon during operation follows a parameter. The Possion process is such that the typhoon wind field model remains unchanged before and after the typhoon passes through the wind farm. The Monte Carlo sampling method is used to obtain the impact probability of each typical typhoon and to establish a typhoon stochastic impact model. Based on the typhoon stochastic impact model, the impact of direct and indirect impacts on component failure rate is analyzed, and an incremental model of component failure rate under typhoon impact is constructed.

3. The method for predicting the remaining life of offshore wind turbines considering the impact of typhoons according to claim 2, characterized in that, The typhoon stochastic impact model is as follows: in, Indicates in (0, t Typhoon occurs in the interior n The probability of a secondary impact. For typhoon arrival rate, For the first Time required for the first arrival of the secondary impact The probability density, The value is the gamma function value. The gradient wind speed at the maximum wind radius. This is an empirical parameter, with a value range of 6.93 to 6.

97. The pressure difference at the center of the typhoon. The radius of maximum wind speed. For the parameters of the formula; v This indicates the instantaneous wind speed of a typhoon. This refers to the typhoon's movement speed; The distance from any location of the typhoon to its center; This is an empirical parameter, with a value range of 0.5 to 0.

7.

4. The method for predicting the remaining life of offshore wind turbines considering the impact of typhoons according to claim 1, characterized in that, Distinguishing between the effects of direct and indirect impacts, the incremental model for component failure rate under direct impact is expressed as follows: The incremental model of component failure rate under indirect impact is expressed as follows: in, The failure rate of external components of the generator unit caused by direct impact from the typhoon. , These are the correction factors for the direct impact of typhoons. The wind load is perpendicular to the surface of the wind turbine. ρ air density, C s The shape factor, The projected area is the area perpendicular to the wind direction plane. This indicates the failure rate of internal components of the generator set caused by the indirect impact of the typhoon. , These are the correction factors for the indirect impact of typhoons. The power of the fan. The wind energy conversion efficiency coefficient. This refers to the tip speed ratio of the fan blades. The pitch angle of the wind turbine blades. The radius of the wind turbine blade.

5. The method for predicting the remaining life of offshore wind turbines considering the impact of typhoons according to claim 1, characterized in that, The construction of the degradation and impact correlation model based on component operating status includes the following steps: Obtain historical operational monitoring data of offshore wind turbines; Principal component clustering analysis was used to extract the health status characteristics of offshore wind turbine components under different operating conditions, and different component states were classified according to the difference values ​​between other states and the healthy state. Based on the component state classification results under different working conditions and the historical state evolution process, a stochastic state model is established to describe the historical state transition process of the component based on the Markov chain state transition process, and a degradation and impact correlation model is obtained.

6. The method for predicting the remaining life of offshore wind turbines considering the impact of typhoons according to claim 1, characterized in that, Assuming system components k Failure rate λ k Failure rate due to natural degradation λ 0k and the cumulative increase in component failure rate due to repeated typhoon impacts The lifespan of the turbine components is a non-negative continuous random variable, and their natural degradation process is described by a Weibull distribution. Combining the component failure rate increment model under typhoon impact and the degradation-impact correlation model, a time-varying failure rate model for offshore wind turbine components considering the dependence of degradation and impact is obtained: in, Number of typhoon impacts; For The cumulative number of times a typhoon has impacted the area; For the first During the second typhoon The magnitude of the impact at any given moment; Let be the correlation function between component degradation and impact, representing the component k In running status The impact of impacts on the failure rate increment; and These represent the initial arrival time and the end time of the typhoon's impact on the generator unit, respectively; β For shape parameters, This is the scale parameter.

7. The method for predicting the remaining life of offshore wind turbines considering the impact of typhoons according to claim 6, characterized in that, The remaining life prediction model for the unit components is as follows: Among them, component reliability This indicates that the normal operating time of the component is greater than probability, P Represents the probability of an event. This represents the expected value function of the component failure rate considering the impact of random typhoons. n z This indicates the total number of typhoon impact scenarios. This is a typhoon impact scenario. A single typhoon impact scenario is a set of the typhoon's maximum wind speed radius, moving speed, central pressure difference, wind speed at the maximum wind speed radius, and the typhoon's direction of movement. Variables for typhoon impact scenarios The moment when the component fails, running until The remaining lifespan of the component at any given time is , The remaining life expectancy is... For the remaining lifespan time variable, Represents the expected value of remaining lifespan. The distribution function, To determine the expected reliability value function of components under historical typhoon impacts, This represents the reliability function considering a random typhoon impact scenario. for The derivative function, To make take Probability density function of expected remaining life The moment corresponding to the maximum value, This is the reliability failure threshold.

8. The method for predicting the remaining life of offshore wind turbines considering the impact of typhoons according to claim 1, characterized in that, By statistically analyzing the lifespan frequency of a component of a wind turbine in an offshore wind farm, the parameters of a natural degradation failure rate model are estimated using the maximum likelihood estimation method. Furthermore, based on natural degradation, parameters are extracted... Taiwanese generator set performs parameter estimation for incremental component failure rate model under typhoon impact: in, β For shape parameters, For scale parameters, Let Weibull be the probability density function. Let be the likelihood function. This is a sample value of component lifespan. Let be the log-likelihood function. For sample size, For the unit The root mean square error, and The units At the monitoring point The estimated and actual remaining lifetime values ​​at each time point. The number of monitoring points This represents the number of experimental fan samples drawn. for The unit's root mean square error.

Images