Wind turbine control method and system based on component interaction influence analysis

Abstract

The invention relates to the technical field of wind turbine generator management, in particular to a wind turbine generator control method and system based on component interaction influence analysis, and the method comprises the steps: collecting the multi-source operation data of a wind turbine generator impeller system and a gear box system and the turbulence intensity of a wind field in real time; based on the multi-source operation data, calculating a dynamic interaction influence index between the impeller and the gearbox; according to the dynamic interaction influence index and the turbulence intensity of the wind field, judging whether the wind turbine is in a low-interaction influence working condition or a high-interaction influence working condition currently; when it is judged that the working condition is the low-interaction-influence working condition, a first control strategy based on historical optimal parameters is adopted to control the wind turbine; when it is judged that the working condition is the high-interaction-influence working condition, a second control strategy based on the prediction model is adopted to control the wind turbine; the dynamic interaction influence of the impeller and the gearbox can be evaluated in real time, and grading self-adaptive control is achieved accordingly.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind turbine management, and particularly relates to a wind turbine control method and system based on component interaction influence analysis. BACKGROUND

[0002] With the wide application of wind power generation technology, the reliability and operation efficiency of wind turbine are increasingly concerned. In actual operation, there is strong dynamic coupling and mechanical interaction between the impeller system and the gearbox system of the wind turbine. Especially in the wind field environment with fluctuating wind speed or high turbulence intensity, the fluctuation of the aerodynamic load generated by the impeller will significantly affect the dynamic response of the gearbox, aggravate the vibration and stress level, and further affect the service life of the transmission system and the stability of the whole machine power generation.

[0003] The control strategy of the existing wind turbine is mostly designed based on the wind speed, power or state parameters of a single component, such as using standard variable pitch control or linear control method based on transmission chain torsional vibration suppression. However, such methods generally do not fully consider the real-time dynamic interaction between the impeller and the gearbox. Especially when the wind condition changes dramatically, it is difficult to achieve precise and adaptive control according to the coupling relationship between the two, resulting in reduced operation reliability of the unit in complex working conditions, excessive mechanical load, and even causing premature fatigue damage of key components. SUMMARY

[0004] The present application provides a wind turbine control method and system based on component interaction influence analysis, which can real-time evaluate the dynamic interaction influence between the impeller and the gearbox, and realize hierarchical adaptive control accordingly, and can effectively solve the problems in the background art.

[0005] In order to achieve the above purpose, in a first aspect, the present application provides a wind turbine control method based on component interaction influence analysis, comprising: real-time collecting multi-source operation data of the impeller system and the gearbox system of the wind turbine and the turbulence intensity of the wind field; based on the multi-source operation data, calculating a dynamic interaction influence index between the impeller and the gearbox; determining whether the wind turbine is currently in a low interaction influence working condition or a high interaction influence working condition according to the dynamic interaction influence index and the turbulence intensity of the wind field; when it is determined to be a low interaction influence working condition, using a first control strategy based on historical optimal parameters to control the wind turbine; when it is determined to be a high interaction influence working condition, using a second control strategy based on a prediction model to control the wind turbine.

[0006] In a possible design, the multi-source operation data at least includes impeller strain distribution data, impeller vibration data, gearbox vibration data and gearbox speed data.

[0007] With reference to the first aspect, in a possible design, the dynamic interaction influence indicator comprises a load fluctuation coefficient and a response correlation degree.

[0008] With reference to the first aspect, in a possible design, the calculation of the load fluctuation coefficient comprises: calculating a real-time aerodynamic load of the impeller based on impeller strain distribution data; calculating a load transmission correction factor based on transmission system parameters; generating the load fluctuation coefficient according to statistical characteristic values of the impeller aerodynamic load and the load transmission correction factor.

[0009] With reference to the first aspect, in a possible design, the load transmission correction factor is calculated by the product of a gearbox speed correction coefficient, a transmission ratio correction coefficient and a main shaft stiffness correction coefficient; wherein the main shaft stiffness correction coefficient is dynamically updated based on a real-time elastic modulus of the main shaft and a real-time cross-sectional moment of inertia.

[0010] With reference to the first aspect, in a possible design, determining that the wind turbine is currently in a low interaction influence working condition or a high interaction influence working condition according to the dynamic interaction influence indicator and a wind farm turbulence intensity comprises: determining dynamic threshold values corresponding to the load fluctuation coefficient and the response correlation degree respectively based on historical operation data and real-time operation state; calculating differences between the load fluctuation coefficient and the response correlation degree and their respective dynamic threshold values; performing weighted fusion on the differences by taking the wind farm turbulence intensity as a weight factor to obtain a working condition determination value; determining that the wind turbine is in a low interaction influence working condition or a high interaction influence working condition according to the working condition determination value.

[0011] With reference to the first aspect, in a possible design, the first control strategy is a historical optimal control parameter combination most similar to a current working condition characteristic, which is retrieved from a historical database.

[0012] With reference to the first aspect, in a possible design, the second control strategy comprises: predicting vibration and stress states of the gearbox in a future time period by a gearbox response prediction model, and generating a control parameter combination based on a prediction result.

[0013] With reference to the first aspect, in a possible design, the control parameter combination comprises a variable pitch angle, a generator torque reference value and a transmission chain damping coefficient.

[0014] Secondly, the application further provides a wind turbine control system based on component interaction influence analysis, comprising: A multi-source data acquisition module is configured to acquire, in real time, operation data of a wind turbine impeller system, operation data of a gear box system, and wind farm turbulence intensity data; A dynamic interaction calculation module is configured to receive the multi-source operation data and calculate an index of dynamic interaction between the impeller and the gear box based on the multi-source operation data; A working condition determination module is configured to receive the index of dynamic interaction and the wind farm turbulence intensity data, and determine whether the wind turbine is currently in a low interaction condition or a high interaction condition based on the index of dynamic interaction and the wind farm turbulence intensity; A first execution module is configured to, when the working condition determination module determines that the wind turbine is currently in the low interaction condition, call a first control strategy based on historical optimal parameters, and control the wind turbine according to the first control strategy; A second execution module is configured to, when the working condition determination module determines that the wind turbine is currently in the high interaction condition, call a second control strategy based on a prediction model, and control the wind turbine according to the second control strategy.

[0015] The technical scheme of the present application can achieve the following technical effects: The present application effectively solves the problem that the existing control method cannot cope with the dynamic coupling effect between the impeller and the gear box by multi-source data sensing, dynamic interaction quantitative evaluation, operation condition determination, and differential control strategy execution. Firstly, by simultaneously acquiring multi-dimensional operation data of the impeller and the gear box and calculating an index of dynamic interaction, the interaction between the components is quantified. Secondly, based on the index of dynamic interaction and the wind farm turbulence intensity, the working condition is determined, and the control strategy is differentiated. In the low interaction condition, the historical optimal parameters are used for control, ensuring the control efficiency and system stability. In the high interaction condition, the prediction control is started, and the control parameters are generated in advance through the gear box response prediction model, realizing the transition from passive response to active intervention. Finally, the cooperative operation of the two control strategies enables the system to maintain efficient operation in stable conditions and effectively suppress vibration and stress peaks in complex conditions, improving the operation reliability and equipment life of the unit. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A logic flow chart of the wind turbine control method based on component interaction influence analysis in the embodiment of the present application; Figure 2 A structure block diagram of the wind turbine control system based on component interaction influence analysis in the embodiment of the present application. DETAILED DESCRIPTION

[0017] The present application will be described below in conjunction with the drawings in the present application.

[0018] As Figure 1As shown, the wind turbine control method based on component interaction influence analysis of the application specifically comprises the following steps: Step S1, real-time collection of multi-source operation data of the wind turbine impeller system and the gear box system and the wind farm turbulence intensity; the multi-source operation data at least includes impeller strain distribution data, impeller vibration data, gear box vibration data and gear box rotating speed data; Step S2, calculation of the dynamic interaction influence index between the impeller and the gear box based on the multi-source operation data; Step S3, determination of whether the wind turbine is currently in a low interaction influence working condition or a high interaction influence working condition according to the dynamic interaction influence index and the wind farm turbulence intensity; Step S4, when the low interaction influence working condition is determined, a first control strategy based on historical optimal parameters is adopted to control the wind turbine; Step S5, when the high interaction influence working condition is determined, a second control strategy based on a prediction model is adopted to control the wind turbine.

[0019] In the embodiment, the existing method only designs a control strategy based on wind speed, power or a single component parameter, and cannot reflect the mutual influence of impeller aerodynamic load fluctuation and gear box dynamic response; the method collects multi-source operation data such as impeller strain distribution data, impeller vibration data, gear box vibration data and gear box rotating speed data in real time, combines with the wind farm turbulence intensity, further calculates the load fluctuation coefficient and the response correlation degree, quantifies the dynamic interaction relationship between the impeller and the gear box from the data dimension to the index dimension, changes the control basis from the single component state to the system coupling state, and solves the fundamental problem of ignoring the interaction influence in the existing method; the method divides the low interaction influence working condition and the high interaction influence working condition based on the dynamic interaction influence index and the turbulence intensity: the historical optimal parameters are called under the low working condition to avoid the waste of resources of complex real-time calculation; the vibration and stress state is predicted in advance through the gear box response prediction model under the high working condition, the predictive control is realized, the low working condition control redundancy problem is solved, the high working condition response lag problem is solved, and the on-demand control is achieved; in the strategy control logic of the method, the historical optimal parameters called under the low working condition are the verified parameters considering the power generation efficiency and the load safety, and the optimal control parameter combination generated by the prediction model under the high working condition avoids the gear box overload risk in advance and avoids the power loss caused by excessive control, realizes the cooperative balance of safety and efficiency, and solves the contradiction between safety and efficiency in the prior art.

[0020] More specifically, in the method, under the condition of low interaction influence, calling the historical optimal parameters can greatly reduce the real-time calculation load of the control unit, and the saved computing resources can support the accurate operation of the gearbox response prediction model under the condition of high interaction influence; At the same time, the prediction results under high working condition can update the historical database, optimize the optimal parameters called under low working condition, so that the computing resources of the control unit can be allocated according to the needs, and the overall operation efficiency of the control system is improved without increasing the hardware cost, which cannot be achieved by single real-time control or historical parameter control; At the same time, the dynamic interaction index and the two-dimensional working condition judgment of the turbulence intensity can improve the working condition classification accuracy and avoid strategy mis-triggering. The existing control only judges the working condition according to the wind speed or single component vibration value, which is easy to misjudge the working condition and trigger unnecessary protection strategies due to occasional interference. The method combines the dynamic interaction index reflecting the inherent coupling strength of the impeller and the gearbox with the wind field turbulence intensity reflecting the external environmental disturbance to form a two-dimensional judgment standard. For example, when the instantaneous gust causes abnormal impeller vibration data, but the response correlation degree in the interaction influence index does not exceed the threshold value and the turbulence intensity is low, it is judged as a low interaction working condition, and the complex prediction control is not triggered, avoiding the power loss caused by strategy mis-triggering. On the contrary, when the turbulence intensity is high and the response correlation degree reaches the threshold value, it is judged as a high interaction working condition, ensuring that the prediction control intervenes in time, improving the working condition classification accuracy, reducing strategy mis-triggering compared with single-dimensional judgment, ensuring the safety of components, and avoiding the influence of unnecessary control adjustment on power generation efficiency. In addition, the existing method is prone to control parameter mutation when the wind condition changes suddenly, which causes the impeller and the gearbox transmission chain to bear additional impact load due to the lack of prediction of the change trend of the interaction influence. In the method, the wind field turbulence intensity can perceive the change trend of the wind condition in advance, the dynamic interaction index quantifies the change rate of the interaction relationship, and combined with the high working condition prediction model, it can adjust the control parameters in advance based on the predicted gearbox state at the initial stage of the transition from low working condition to high working condition, avoid control parameter mutation, and greatly reduce the impact load of the transmission chain when the wind condition changes suddenly, thereby reducing the accumulation of mechanical fatigue. This advantage can only be achieved by the cooperation of the three.

[0021] In some embodiments of the present application, the wind turbine impeller system and the gearbox system have a strong dynamic coupling relationship. The fluctuation of the aerodynamic load borne by the impeller is transmitted to the gearbox through the transmission chain, causing the gearbox vibration to intensify and the stress to be abnormal, and the vibration feedback of the gearbox reacts on the impeller, accelerating the fatigue damage of the impeller. The existing control strategy cannot capture the essential characteristics of this dynamic interaction because it only collects single component parameters or basic environmental data. For example, relying only on wind speed data cannot reflect the correlation between impeller strain distribution and gearbox vibration, and collecting only gearbox speed cannot predict the influence of aerodynamic load fluctuation on gearbox stress. Therefore, multiple source operation data of the impeller system and the gearbox system and the wind field turbulence intensity need to be collected simultaneously.

[0022] Specifically, combined with the structural characteristics and data requirements of the impeller and the gearbox, data collection is carried out through the establishment of a hierarchical sensor network, as follows: Impeller system sensor deployment: distributed optical fiber sensors are arranged at the root, middle, and tip of the three blades of the impeller, and multiple sensing nodes are uniformly arranged along the length of each blade to collect real-time strain distribution data of the impeller; multiple MEMS accelerometers are arranged at the connection between the impeller hub and the main shaft to collect vibration data of the impeller, including vibration acceleration and vibration frequency; the sensors are packaged in corrosion-resistant and wind-load-resistant sealed housings to adapt to the harsh environment of the wind farm.

[0023] Gearbox system sensor deployment: multiple MEMS accelerometers are arranged at the bearing seats of the input shaft, intermediate shaft, and output shaft of the gearbox to collect vibration data of the gearbox; a non-contact speed sensor is installed at the end of the output shaft of the gearbox to obtain real-time speed data of the gearbox; temperature sensors and oil contamination sensors are arranged in the gearbox oil tank to assist in determining the operating state of the gearbox; all sensors are connected to the data acquisition module through shielded cables to reduce electromagnetic interference.

[0024] Wind farm turbulence intensity sensor deployment: a three-dimensional ultrasonic anemometer is installed on the windward surface of the top of the wind turbine nacelle to collect real-time wind speed, wind direction, and turbulence intensity data; the turbulence intensity is calculated based on the ratio of the standard deviation of wind speed to the average wind speed within a predetermined time period; multiple auxiliary weather stations are uniformly arranged within a predetermined range around the wind farm to synchronously collect turbulence data for cross-validation with the data from the nacelle ultrasonic anemometer to avoid errors caused by single-point data anomalies.

[0025] More specifically, a high-speed data acquisition card based on FPGA is used, which integrates multiple-channel analog signal input interfaces and can simultaneously connect distributed optical fiber sensors, MEMS accelerometers, speed sensors, and ultrasonic anemometers to realize synchronous acquisition of multi-source data; the acquisition card is synchronized with the wind turbine main control system clock through IEEE 1588v2 protocol to ensure that the timestamp error of all data is ≤1ms, avoiding the problem of asynchronous time and space of multiple sensors; the following measures are taken to ensure stable data transmission: The sensor signal transmission adopts a differential signal transmission method with shielded twisted pair to reduce the influence of electromagnetic field interference of the generator on the signal; a 5G private network communication link is built between the data acquisition module and the main control system to realize real-time data transmission using the low latency and high reliability characteristics of 5G; local buffering is set up in the acquisition module to supplement the transmission of buffered data when the network is temporarily interrupted to avoid data loss; real-time preprocessing is performed on the collected raw data, including filtering and normalization, to ensure the quality of the data transmitted to the main control system.

[0026] In order to facilitate data management, a data integration module is constructed in the wind turbine main control system. The collected impeller strain distribution data, impeller vibration data, gearbox vibration data, gearbox speed data and wind farm turbulence intensity data are structured and integrated in the format of "time stamp + component identification + data type" to form a unified data frame. At the same time, a real-time database is established to store the original data with a time granularity of 1ms.

[0027] In this embodiment, through the combination of distributed optical fiber sensor and MEMS accelerometer, full-dimensional acquisition of impeller strain distribution, vibration and gearbox vibration, speed is realized, combined with wind farm turbulence intensity data, data input is provided for step S2 dynamic interaction influence index calculation, avoiding misjudgment of interaction characteristics caused by data missing; IEEE 1588v2 protocol is used to realize multi-sensor clock synchronization, and through the combination of 5G private network and local cache, data transmission stability is ensured under extreme turbulence conditions, signal loss or delay is avoided; the sensor package adopts anti-corrosion and wind load resistance design, the transmission link realizes anti-interference through differential signal and 5G private network, and the filtering and normalization operation in the data preprocessing link further improves the data quality, ensuring that the collected data can truly reflect the running state and interaction characteristics of the impeller and the gearbox.

[0028] In some embodiments of the present application, the dynamic interaction influence index includes a load fluctuation coefficient and a response correlation degree; the load fluctuation coefficient is used to quantify the impact strength of the fluctuation degree of the impeller aerodynamic load on the gearbox; the response correlation degree is used to quantify the internal correlation between the impeller state change and the gearbox dynamic response; therefore, the multi-source operation data needs to be processed to convert the scattered impeller strain, vibration and gearbox vibration, speed data into quantifiable interaction indicators.

[0029] Specifically, based on the multi-source operation data collected in step S1, including impeller strain distribution data, impeller vibration data, gearbox vibration data, gearbox speed data, the data is preprocessed to eliminate interference and unify data dimensions; specifically, statistical outlier identification criteria are used to identify and eliminate outliers in the impeller strain and gearbox vibration data, and then the data is denoised by smoothing processing to reduce the influence of high-frequency noise on index calculation; based on the time synchronization realized by the clock synchronization protocol in step S1, the time dimension of other component data is aligned with the time stamp of any key parameter of the core component as the reference; in the spatial dimension, the spatial mapping relationship between the impeller monitoring site and the gearbox monitoring site is established by combining the impeller blade sensor node position and the gearbox sensor position, to ensure that the data corresponds to the physical position of the component interaction.

[0030] More specifically, the load fluctuation coefficient is used to represent the fluctuation degree of the impeller aerodynamic load and the impact strength of the transmission chain to the gearbox, and the calculation steps are as follows: In step S211, based on the pretreated impeller strain distribution data, the inherent material parameters and structural parameters of the impeller blade are combined, and the real-time aerodynamic load of the impeller is calculated through the material mechanics formula. The blade material parameters include the elastic modulus and the Poisson's ratio, and the structural parameters include the blade length and the cross-sectional area. The material mechanics formula is F=E·ε·S, wherein E is the elastic modulus, ε is the strain value of the monitoring site, F is the load value of the monitoring site, and S is the cross-sectional area of the monitoring site. The formula can be derived from the basic Hooke's law σ=E·ε and the stress definition σ=F / S. The stress definition is substituted into the Hooke's law to obtain F / A=E·ε, and after rearrangement, F=E·ε·A is obtained. In step S212, the statistical characteristic value of the aerodynamic load of the impeller in each period is calculated by using a sliding window with a preset time length as a calculation period. The statistical characteristic value of the aerodynamic load of the impeller includes the load average value and the load standard deviation. In step S213, considering the influence of the transmission system characteristics on the load transmission effect, a load transmission correction factor is calculated based on the transmission system parameters, which is used to correct the actual transmission effect of the load fluctuation. In step S214, based on the load transmission correction factor and the statistical characteristic value of the aerodynamic load of the impeller, a load fluctuation coefficient is calculated. Specifically, the load fluctuation coefficient is proportional to the load standard deviation and inversely proportional to the load average value, and is corrected by the load transmission correction factor. The larger the coefficient value is, the greater the aerodynamic load fluctuation of the impeller is, and the stronger the impact on the gearbox is.

[0031] The response correlation degree is used to represent the correlation degree between the impeller state change and the dynamic response of the gearbox. The calculation steps are as follows by using the multi-dimensional feature fusion and correlation degree analysis method: In step S221, when performing multi-dimensional feature extraction on the multi-source operation data, the feature parameters are extracted from the pre-processed impeller operation data for the impeller side features, including but not limited to impeller strain distribution data, impeller vibration data, and impeller aerodynamic load calculation results. The extracted feature parameters include vibration amplitude, vibration frequency parameters, load parameters, and strain parameters. The vibration amplitude can be selected from parameters such as peak value, effective value, and peak-to-peak value that can reflect the vibration intensity. The vibration frequency parameters can be selected from parameters such as main frequency, multiple frequency, and side frequency that can reflect the vibration frequency characteristics. The load parameters can be selected from parameters such as load change rate, load peak value, and load average value that can reflect the load state. The strain parameters can be selected from parameters such as strain change rate, strain peak value, and strain average value that can reflect the strain characteristics. For the gearbox side features, the feature parameters are extracted from the pre-processed gearbox operation data, including but not limited to gearbox vibration data, gearbox speed data, gearbox temperature data, and gearbox oil parameters. The extracted feature parameters include vibration amplitude, vibration frequency parameters, speed parameters, and temperature parameters. The vibration amplitude can be selected from parameters such as peak value, effective value, and peak-to-peak value. The vibration frequency parameters can be selected from parameters such as main frequency, multiple frequency, side frequency, and frequency energy proportion. The speed parameters can be selected from parameters such as speed fluctuation value, speed change rate, and speed average value. The temperature parameters can be selected from parameters such as oil temperature change rate, bearing temperature peak value, and oil temperature average value.

[0032] In step S222, the multi-dimensional features of the impeller side and the gearbox side are standardized to a pre-set numerical interval using a standardization method to eliminate dimensional differences. The standardization method includes but is not limited to Min-Max standardization, Z-Score standardization, and decimal scaling standardization, which are suitable for numerical feature normalization. The pre-set numerical interval can be flexibly set according to the actual data distribution characteristics and correlation analysis requirements, such as the interval [0, 1] or the interval [-1, 1]. During the standardization process, different types of feature parameters can be selected for different standardization methods, or all feature parameters can be uniformly standardized using the same method, as long as the goal of eliminating dimensional differences and making different features comparable can be achieved.

[0033] Step S223, based on the completed standardized impeller side and gearbox side multi-dimensional features, a correlation matrix of the impeller side features and the gearbox side features is constructed, the row dimension of the matrix corresponds to each feature parameter of the impeller side, the column dimension corresponds to each feature parameter of the gearbox side, and the elements in the matrix are used to represent the correlation value between the corresponding impeller side features and the gearbox side features; the correlation between each pair of features in the matrix is calculated using a correlation analysis method, which includes but is not limited to mutual information method, Pearson correlation coefficient method, Spearman rank correlation coefficient method, Kendall concordance coefficient method and other analysis methods that can quantify the correlation between two variables; the greater the correlation value obtained by the correlation calculation, the higher the degree of mutual correlation between the corresponding impeller side features and the gearbox side features, and the stronger the consistency of the change trend or state response of the two.

[0034] Step S224, according to the characteristics of the impeller and gearbox transmission chain, the correlation values in the correlation matrix are given weights, and the response correlation is obtained through composite operation; the greater the correlation value, the higher the dynamic interaction strength of the impeller and the gearbox; the transmission chain characteristics include but are not limited to load transmission path, energy transmission efficiency, component connection relationship, transmission system structure design characteristics, etc., the weight assignment can be based on the importance of the interaction influence of each link of the transmission chain, the sensitivity of the feature parameters to the component state change, engineering practice experience or algorithm optimization result determination, for example, the correlation values corresponding to the features of the components in the main load transmission path are given higher weights, and the correlation values corresponding to the features with lower sensitivity are given lower weights; then the weighted correlation values are processed through composite operation to obtain the final response correlation, the composite operation includes but is not limited to weighted summation, weighted average, weighted product and normalization, etc. The operation mode that can integrate the correlation information of multiple features; the greater the final response correlation value, the higher the dynamic interaction strength between the impeller state change and the gearbox dynamic response, and the more significant the mutual influence of the two in the running process.

[0035] In this embodiment, the load fluctuation coefficient is combined with the impeller aerodynamic load and the transmission system parameter correction, which not only quantifies the degree of impeller load fluctuation, but also considers the influence of load transmission efficiency; the response correlation is calculated through multi-dimensional feature fusion and weighted correlation calculation, which reflects the dynamic interaction nature of the impeller and the gearbox, and improves the accuracy of working condition judgment; by establishing the spatial mapping relationship between the impeller part and the gearbox part, and giving the correlation value weight combined with the transmission chain characteristics, the response correlation calculation is no longer an abstract numerical operation, but a quantitative analysis based on the physical interaction mechanism of the components, which not only conforms to the component degradation correlation research logic in the disclosure material, but also gives the dynamic interaction influence index a clear physical meaning, avoiding the defects of the existing index deviating from the actual interaction of the components.

[0036] More specifically, in order to accurately calculate the load transmission correction factor, its constituent parts, i.e. the gearbox speed correction coefficient, the transmission ratio correction coefficient and the main shaft stiffness correction coefficient, need to be determined respectively; the load transmission correction factor is calculated by the product of the above three correction coefficients.

[0037] The calculation method of the gearbox speed correction coefficient is as follows: first, a reference value is set, which is determined based on the load transmission characteristic experimental data of the wind turbine when it is running at the rated speed; second, the ratio of the real-time speed to the rated speed is calculated according to the real-time running state of the gearbox; then, the ratio is multiplied by a preset weight coefficient, which is used to adjust the sensitivity of the speed change to the load transmission efficiency; the determination method of the weight coefficient is: based on the historical running data, the measured values of the load transmission efficiency under multiple different speed conditions are selected, and the optimal value of the weight coefficient is solved by regression analysis or optimization algorithm, with the objective of minimizing the mean square error between the measured transmission efficiency and the predicted transmission efficiency calculated by the weight coefficient and the speed ratio; finally, the product result is added to the reference value to obtain the gearbox speed correction coefficient; the real-time speed is derived from the gearbox speed data collected and preprocessed in step S1, and the rated speed is a design parameter of the wind turbine.

[0038] The calculation method of the transmission ratio correction coefficient is as follows: first, a reference value is set, which is determined based on the load transmission characteristic experimental data of the wind turbine when it is running at the rated transmission ratio; second, the real-time transmission ratio of the gearbox is calculated, which is the ratio of the output shaft speed to the input shaft speed of the gearbox; then, the ratio of the real-time transmission ratio to the rated transmission ratio is calculated, and the ratio is multiplied by a preset weight coefficient, which is used to adjust the sensitivity of the transmission ratio change to the load transmission path; the determination method of the weight coefficient is: based on the dynamic simulation model or bench test data of the wind turbine transmission system, the stiffness or damping variation law of the load transmission path under different transmission ratios is analyzed, and the influence degree of the transmission ratio change on the torque fluctuation of the input shaft of the gearbox is taken as the quantitative index, and the value of the weight coefficient is determined by fitting analysis; finally, the product result is added to the reference value to obtain the transmission ratio correction coefficient; the rated transmission ratio is a design parameter of the wind turbine.

[0039] The calculation method of the spindle stiffness correction coefficient is as follows: the coefficient reflects the influence of the spindle stiffness state on load transmission, and the value is the ratio of the real-time equivalent stiffness of the spindle to the rated equivalent stiffness of the spindle; the real-time equivalent stiffness of the spindle is obtained by multiplying the real-time elastic modulus of the spindle by the real-time sectional inertia moment of the spindle; the real-time elastic modulus of the spindle is obtained by the following method: arranging a temperature sensor on the surface of the spindle, and collecting the temperature of the spindle in real time; according to the known thermal elastic characteristic curve of the spindle material, the real-time elastic modulus at the temperature is obtained by backstepping from the real-time temperature; the real-time sectional inertia moment of the spindle is obtained by one of the following methods: (1) backstepping by modal analysis of the spindle vibration data; (2) according to the real-time outer diameter and / or inner diameter data of the spindle obtained by detection during regular maintenance, and combining the inertia moment formula of the spindle sectional shape such as solid circular shaft or hollow circular shaft to calculate; the rated equivalent stiffness of the spindle is obtained by multiplying the rated elastic modulus of the spindle by the rated sectional inertia moment of the spindle, wherein the rated elastic modulus is the standard value of the spindle material at normal temperature, and the rated sectional inertia moment is calculated according to the design size of the spindle in the design drawing of the wind turbine.

[0040] In the embodiment, the real-time rotating speed, the real-time transmission ratio and the real-time stiffness of the spindle obtained by the inversion algorithm are introduced, so that the correction factor is no longer a fixed design value, but a variable that can be dynamically adjusted with the running state of the wind turbine, which helps to improve the applicability of the coefficient in different working conditions and different life cycle stages of the equipment; the real-time elastic modulus is backstepped from the surface temperature of the spindle, and the real-time sectional inertia moment is obtained from the vibration data or regular detection, so that the correction factor can adapt to the performance changes of the unit in different environmental temperatures.

[0041] In some embodiments of the application, the dynamic interaction strength between the impeller and the gear box changes in real time with the aging of the components and the fluctuation of the wind field environment. If a fixed threshold is used to determine the running condition of the wind turbine, it is easy to appear the problems of misjudgment of the aging unit and misalignment of the high turbulence environment.

[0042] The threshold difference of the dynamic interaction influence index can directly reflect the degree of deviation of the index from the safety range, and the greater the difference, the more the interaction influence exceeds the safety boundary. The wind field turbulence intensity determines the actual influence degree of the index difference on the working condition risk. In a high turbulence environment, the same difference corresponds to a higher risk; in a low turbulence environment, the same difference corresponds to a lower risk. If only the index difference is calculated, it is impossible to distinguish between the high risk of small difference in a high turbulence environment and the low risk of large difference in a low turbulence environment; if only the turbulence intensity is used as the weight, it is also impossible to quantify the deviation degree of the internal interaction characteristics. Therefore, the dynamic threshold difference calculation and the turbulence intensity weight distribution need to be combined to accurately quantify the internal deviation degree and the external risk weight, and to solve the technical problems of poor adaptability and high misjudgment rate caused by fixed threshold and single-dimensional judgment.

[0043] Specifically, the working condition determination method is as follows: Step S31, based on the historical operation data stored in the real-time database in the foregoing embodiments, and combined with the real-time operation state of the wind turbine, the dynamic threshold corresponding to the load fluctuation coefficient and the dynamic threshold corresponding to the response correlation degree are generated, to ensure that the threshold can be adjusted in real time with the change of the working condition. The specific determination steps of the dynamic threshold are as follows: Taking the turbulence intensity and the component state as variables, a dynamic calculation model is constructed. First, all the operation data similar to the current turbulence intensity are retrieved from the historical data, and the safe operation data in which the gear box vibration peak value and the impeller strain are in the safety range are selected. The specific quantile of the load fluctuation coefficient value in this set of safe operation data is calculated, which is taken as the turbulence adaptation threshold suitable for the current turbulence environment. Second, based on the data such as the gear box oil pollution degree and the main shaft temperature collected in the foregoing embodiments, which can reflect the component state, the component aging coefficient is calculated, and then the state correction coefficient is obtained according to the component aging coefficient. Finally, the real-time load fluctuation coefficient dynamic threshold is obtained by multiplying the turbulence adaptation threshold by the state correction coefficient. If there is no similar historical data for the current turbulence intensity, the linear interpolation method is used to supplement the calculation of the turbulence adaptation threshold. Taking the turbulence intensity and the transmission efficiency as variables, a dynamic calculation model is constructed. The safe operation data similar to the current turbulence intensity are retrieved from the historical data, and the specific quantile of the response correlation degree value in this set of data is calculated as the turbulence adaptation threshold suitable for the current turbulence environment. Based on the gear box speed data collected in the foregoing embodiments, the real-time transmission efficiency is calculated, and then the transmission efficiency correction coefficient is obtained according to the real-time transmission efficiency. The real-time response correlation degree dynamic threshold is obtained by multiplying the turbulence adaptation threshold by the transmission efficiency correction coefficient. If the transmission efficiency data is missing, the transmission efficiency correction coefficient is calculated based on the rated transmission efficiency of the gear box.

[0044] Step S32, based on the load fluctuation coefficient and the response correlation degree calculated in real time, and the respective dynamic threshold values, the difference between the two is calculated to quantify the degree to which the indicators deviate from the safety boundary. The specific calculation steps are as follows: Subtract the corresponding dynamic threshold value from the real-time load fluctuation coefficient to obtain the load fluctuation coefficient difference. If the load fluctuation coefficient difference is less than or equal to zero, it means that the current load fluctuation coefficient has not exceeded the dynamic threshold value, and the internal load fluctuation risk is low. If the difference is greater than zero, it means that the load fluctuation coefficient has exceeded the dynamic threshold value, and the greater the difference, the higher the impact risk of load fluctuation on the gearbox. Subtract the corresponding dynamic threshold value from the real-time response correlation degree to obtain the response correlation degree difference. If the response correlation degree difference is less than or equal to zero, it means that the current response correlation degree has not exceeded the dynamic threshold value, and the interaction between the impeller and the gearbox is within a safe range. If the difference is greater than zero, it means that the response correlation degree has exceeded the dynamic threshold value, and the greater the difference, the higher the degree to which the internal interaction coupling strength deviates from the safe range.

[0045] Step S33, based on the real-time wind field turbulence intensity, combined with the turbulence intensity value verified by the auxiliary meteorological station, the load fluctuation coefficient difference and the response correlation degree difference are assigned weights to reflect the influence of the external environment on the working condition risk. First, the real-time turbulence intensity value is standardized to a fixed interval. Second, set the turbulence weight distribution rule. The higher the turbulence intensity, the greater the influence of the external environment on the load fluctuation. At this time, the weight of the load fluctuation coefficient difference should be correspondingly increased. At the same time, high turbulence environment can easily intensify the vibration feedback of the transmission chain, so the weight of the response correlation degree difference also needs to be adjusted synchronously. If the turbulence intensity exceeds the limit turbulence intensity designed by the wind turbine, the weight of the load fluctuation coefficient difference is preferentially increased.

[0046] Step S34, calculate the working condition judgment comprehensive value by weighted summation, and combine the preset judgment rule to divide the low interaction working condition or high interaction working condition of the wind turbine: The working condition judgment comprehensive value is calculated by weighted summation. Only the positive deviation part of the indicators exceeding the threshold value is considered in the calculation process, i.e. the part with positive value. The negative value is not included in the calculation to ensure that the comprehensive value only quantifies the actual risk after the deviation degree of the indicators exceeding the threshold value and the external risk weight are superimposed. Based on the corresponding relationship between the comprehensive value in the historical data and the working condition risk, a comprehensive determination threshold is set. If the working condition determination comprehensive value is less than or equal to the comprehensive determination threshold, it is determined as a low interaction influence working condition. At this time, the internal index does not exceed the dynamic threshold or the over-threshold degree is low, and the external turbulent flow risk weight is small, so it is not necessary to start the complex prediction control. If the working condition determination comprehensive value is greater than the comprehensive determination threshold, it is determined as a high interaction influence working condition. At this time, the internal index over-threshold degree is high or the external turbulent flow risk weight is large, so it is necessary to trigger the prediction model control to avoid the component damage risk. If the comprehensive value is in the set critical interval, the trend of the comprehensive value in the previous fixed time interval is corrected. If the comprehensive value shows an upward trend, it is determined as a high interaction influence working condition. If the comprehensive value shows a downward trend, it is determined as a low interaction influence working condition.

[0047] In the embodiment, the dynamic threshold can be adjusted in real time with the change of the turbulent flow intensity, the component aging degree and the transmission efficiency. It can adapt to the turbulent flow difference of different wind fields and is compatible with the whole life cycle operation state of the wind turbine from the new unit to the aging period. The problem of too loose threshold for the new unit and too strict threshold for the aging unit is avoided. In the complex scene of component aging and turbulent flow intensity fluctuation, the misjudgment rate of working condition determination is reduced. Through the calculation mode of combining the internal deviation degree and the external risk weight, the internal interaction characteristics and the external environment are deeply integrated. The amplification effect of the external risk on the internal index deviation can be accurately reflected.

[0048] As some embodiments of the present application, when it is determined as a low interaction influence working condition, a first control strategy based on the historical optimal parameters is used to control the wind turbine. At this time, the dynamic coupling of the impeller and the gearbox is weak, the system runs relatively stably, and complex real-time prediction calculation is not needed. By calling the optimal parameter combination, the rapid application of the control parameters is realized, which reduces the calculation load and guarantees the control effect. Before applying the first control strategy, it is necessary to determine the composition, source and storage mode of the historical optimal control parameter combination to ensure the retrievability of the parameters. The historical optimal control parameter combination includes key control variables such as variable pitch angle, generator torque reference value and transmission chain damping coefficient. These parameters are the optimization results considering the power generation efficiency and mechanical safety. Specifically, they include: Variable pitch angle parameter: angle value optimized based on historical data of impeller aerodynamic load, used to adjust the blade wind angle to balance the load and power; Generator torque reference value: torque setting value optimized according to historical data of gearbox speed and transmission efficiency, to ensure the stable operation of the transmission chain; Transmission chain damping coefficient: damping parameter extracted from historical vibration data, used to suppress residual vibration under low-intensity interaction; The historical optimal parameter combination is derived from the actual operation data of the previous operation cycle. The specific generation logic is as follows: in the historical operation, when the system is in a stable working condition, the control unit records the control parameters in this period; the parameters are evaluated by a multi-objective optimization algorithm, and the objective function includes the power generation efficiency and the mechanical safety index; after optimization, the Pareto optimal solution is selected as the historical optimal parameter combination to ensure the balance between efficiency and safety; A historical database is constructed in the main control system of the wind turbine, and the parameter combinations are stored in a structured manner according to the time stamp and the working condition label; each parameter combination is associated with a working condition fingerprint, which is composed of the average value of the turbulence intensity, the average value of the load fluctuation coefficient and the average value of the response correlation in this period, so as to facilitate quick matching; at the same time, the historical database adopts a ring buffer design, and the data of the last N operation cycles are retained, and the old data is automatically overwritten, so as to ensure the timeliness of the parameters.

[0049] More specifically, the retrieval algorithm of the optimal parameter combination adopts a nearest neighbor search algorithm based on similarity, which is as follows: Based on the current real-time data, a working condition feature vector is generated; in the historical database, all parameter combinations with the label of low interaction influence working condition are retrieved, the Euclidean distance between the working condition fingerprint and the working condition feature vector is calculated; the first control strategy is selected from the parameter combination with the highest power generation efficiency among the first k records with the smallest distance; the first control strategy is directly input into the wind turbine variable pitch system and the generator controller, and a smooth transition mechanism is added when the first control strategy is applied, so that the parameter combination corresponding to the first control strategy is gradually applied through a first-order low-pass filter, so as to avoid the impact on the transmission chain caused by parameter mutation.

[0050] As some embodiments of the present application, when it is determined in step S3 that the wind turbine is in a high interaction influence working condition, a second control strategy based on a prediction model is enabled. The second control strategy dynamically generates control parameters considering safety and efficiency by predicting the future state of the gearbox, avoiding response lag or component overload under sudden wind changes.

[0051] Specifically, the gearbox response prediction model includes a vibration prediction sub-model and a stress prediction sub-model, wherein the input of the vibration prediction sub-model is the fluctuation characteristics of the impeller aerodynamic load and the real-time vibration frequency spectrum of the gearbox, and the output is the predicted value of the vibration acceleration of the gearbox bearing seat in a future set time. The input of the stress prediction sub-model is the fluctuation of the input shaft speed of the gearbox, the damping state of the transmission chain and the vibration prediction result, and the output is the predicted value of the gear tooth surface contact stress in a future set time. The gearbox response prediction model is trained based on historical high interaction working condition data, and the actual vibration / stress value is used as the label, and the prediction accuracy is optimized through time series alignment. The precise prediction of the state of the gearbox is realized through the cooperation of the double models. The vibration sub-model captures the load transmission effect, and the stress sub-model quantifies the mechanical response.

[0052] If the vibration prediction value exceeds the safety threshold or the stress prediction value exceeds the fatigue limit, it is marked as a high-risk state; otherwise, it is marked as a controllable state; for the high-risk state, a dynamic pitch rate is calculated based on the turbulence intensity gradient to quickly reduce the impeller load; the upper limit of the generator torque is limited to force the transmission chain stress to be reduced; reverse damping is injected to suppress vibration energy accumulation; for the controllable state, the pitch angle and torque reference values are dynamically adjusted based on the difference between the predicted vibration / stress value and the safety threshold to maximize the power generation within the safety boundary; and the generated parameter combination is real-time issued to the pitch system, generator controller and transmission chain damper.

[0053] As shown in Figure 2 The application also provides a wind turbine control system based on component interaction influence analysis, which specifically comprises the following modules: A multi-source data acquisition module is configured to acquire real-time operation data of a wind turbine impeller system, operation data of a gearbox system and wind farm turbulence intensity data; A dynamic interaction calculation module is configured to receive the multi-source operation data and calculate a dynamic interaction influence index between the impeller and the gearbox based on the multi-source operation data; A working condition determination module is configured to receive the dynamic interaction influence index and the wind farm turbulence intensity data, and determine whether the wind turbine is currently in a low interaction influence working condition or a high interaction influence working condition according to the dynamic interaction influence index and the wind farm turbulence intensity; A first execution module is configured to call a first control strategy based on historical optimal parameters and control the wind turbine according to the first control strategy when the working condition determination module determines that the wind turbine is currently in the low interaction influence working condition; A second execution module is configured to call a second control strategy based on a prediction model and control the wind turbine according to the second control strategy when the working condition determination module determines that the wind turbine is currently in the high interaction influence working condition.

[0054] In this embodiment, the operation data of the impeller and the gearbox and the wind farm turbulence intensity are acquired in real time by the multi-source data acquisition module; the dynamic interaction relationship between the two is quantified by the dynamic interaction calculation module; the working condition is divided by the working condition determination module in combination with the interaction index and the turbulence intensity to avoid the limitations of single parameter determination; finally, the first and second execution modules are activated for different working conditions, stable control is realized in the low interaction influence working condition relying on the historical optimal parameters, and adaptive and accurate control is realized in the high interaction influence working condition through the prediction model, which effectively weakens the adverse effects of the fluctuation of aerodynamic load under wind speed fluctuation or high turbulence environment on the dynamic response of the gearbox, reduces the vibration and stress level of the gearbox, and improves the service life of the transmission system and the stability of the whole machine power generation.

[0055] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.

Claims

1. A wind turbine control method based on component interaction effect analysis, characterized in that, The method comprises: real-time acquisition of multi-source operation data of a wind turbine impeller system and a gearbox system and wind farm turbulence intensity; based on the multi-source operation data, calculating a dynamic interaction influence index between the impeller and the gearbox; determining whether the wind turbine is currently in a low interaction influence working condition or a high interaction influence working condition according to the dynamic interaction influence index and the wind farm turbulence intensity; when it is determined that the wind turbine is in the low interaction influence working condition, using a first control strategy based on historical optimal parameters to control the wind turbine; when it is determined that the wind turbine is in the high interaction influence working condition, using a second control strategy based on a prediction model to control the wind turbine.

2. The wind turbine control method based on component interaction effect analysis according to claim 1, characterized in that, The multi-source operation data at least includes impeller strain distribution data, impeller vibration data, gearbox vibration data and gearbox rotational speed data.

3. The wind turbine control method based on component interaction analysis according to claim 1 or 2, characterized in that, The dynamic interaction influence index includes a load fluctuation coefficient and a response correlation degree.

4. The wind turbine control method based on component interaction influence analysis according to claim 3, characterized in that, The calculation of the load fluctuation coefficient comprises: calculating real-time aerodynamic load of the impeller based on the impeller strain distribution data; combining transmission system parameters to calculate a load transmission correction factor; generating the load fluctuation coefficient according to statistical characteristic values of the impeller aerodynamic load and the load transmission correction factor.

5. The wind turbine control method based on component interaction effect analysis according to claim 4, characterized in that, The load transmission correction factor is calculated by the product of a gearbox rotational speed correction coefficient, a transmission ratio correction coefficient and a main shaft stiffness correction coefficient; wherein the main shaft stiffness correction coefficient is dynamically updated based on real-time elastic modulus and real-time cross-sectional moment of inertia of the main shaft.

6. The wind turbine control method based on component interaction effect analysis according to claim 5, wherein, Determining whether the wind turbine is currently in a low interaction influence working condition or a high interaction influence working condition according to the dynamic interaction influence index and the wind farm turbulence intensity comprises: determining dynamic threshold values corresponding to the load fluctuation coefficient and the response correlation degree respectively based on historical operation data and real-time operation state; calculating the difference between the load fluctuation coefficient and the response correlation degree and their respective dynamic threshold values; weighting and fusing the difference values by taking the wind farm turbulence intensity as a weight factor to obtain a working condition determination value; determining whether the wind turbine is in a low interaction influence working condition or a high interaction influence working condition according to the working condition determination value.

7. The wind turbine control method based on component interaction effect analysis according to claim 6, characterized in that, The first control strategy is a historical optimal control parameter combination most similar to the current working condition characteristics retrieved from a historical database.

8. The wind turbine control method based on component interaction influence analysis according to claim 7, characterized in that, The second control strategy predicts the vibration and stress state of the gearbox in a future time period through a gearbox response prediction model, and generates a control parameter combination based on the prediction result.

9. The wind turbine control method based on component interaction influence analysis according to claim 8, wherein, The control parameter combination includes a variable pitch angle, a generator torque reference value and a transmission chain damping coefficient.

10. A wind turbine control system based on component interaction analysis, characterized by, The method comprises: a multi-source data acquisition module for real-time acquisition of operation data of a wind turbine impeller system, operation data of a gearbox system and wind farm turbulence intensity data; a dynamic interaction calculation module for receiving multi-source operation data and calculating a dynamic interaction influence index between the impeller and the gearbox based on the multi-source operation data; a working condition determination module for receiving the dynamic interaction influence index and wind farm turbulence intensity data, and determining whether the wind turbine is currently in a low interaction influence working condition or a high interaction influence working condition according to the dynamic interaction influence index and the wind farm turbulence intensity; a first execution module for calling a first control strategy based on historical optimal parameters and controlling the wind turbine according to the first control strategy when the working condition determination module determines that the wind turbine is currently in a low interaction influence working condition; The second execution module is configured to call a second control strategy based on a prediction model and control the wind turbine according to the second control strategy when the working condition determination module determines that the wind turbine is currently in the high interaction working condition.

Images