Load parameter set construction method and system of wind turbine generator transmission chain test platform

Abstract

The invention discloses a load parameter set construction method and system of a wind turbine generator transmission chain test platform, and relates to the technical field of wind turbine generator system simulation test. The method comprises the following steps: dividing a wind environment area according to a wind power parameter index, and matching a corresponding wind environment area according to a simulation task; based on a wind environment area and power grid operation characteristics, considering extreme wind conditions, mechanical faults and electrical faults, and constructing a multi-working-condition operation scene; performing simulation according to the operation scene and the transmission chain parameters of the wind turbine generator to obtain simulation data; and according to actual requirements of different working conditions, key parameter extraction is carried out on simulation data by using a deep learning algorithm in combination with a key parameter extraction principle, and a final load parameter set is obtained. According to the method, the existing wind turbine generator load parameter extraction standard is improved based on the climate environment characteristics, multi-working-condition comprehensive joint simulation is carried out on the power grid side and the wind power machinery side, and key parameters are screened and extracted, so that the applicability and rationality of the wind turbine generator transmission chain test are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of wind turbine system simulation testing, and particularly relates to a load parameter set construction method and system of a wind turbine transmission chain test platform. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute the prior art.

[0003] With the rapid scale development of new energy power generation, a large number of power electronic devices are put into operation in the power grid, so that the power system has shown the characteristics of "double high", specifically, low inertia, weak power grid, wide frequency oscillation, etc. When the wind turbine is connected to the power system for operation, what influence does the "double high" characteristic have on the operation of the wind turbine? This needs to be verified through simulation technology.

[0004] However, the existing simulation technology is difficult to simulate the mechanical load on the wind turbine side and the electrical load applied to the wind turbine by the power grid at the same time, resulting in a lack of effective verification means for the performance of the unit, and the stability of the structure of the wind turbine and the risk of safe operation are highlighted. Among them, the transmission chain is the most important part of the wind turbine, and is also the most complex and highest failure rate structure. The transmission chain is subjected to wind load and multiple complex excitations in the power grid during operation, and its performance directly determines the safety and service life of the wind turbine, and is the core of affecting the safe and reliable operation of the whole machine. Therefore, it is urgent to fully utilize the test technology and digital technology to carry out simulation experiments, and to ensure the safe and reliable operation of the transmission chain of the wind turbine. The transmission chain test platform can simulate six degrees of freedom load and power grid, and can equivalent to the operating conditions of the unit, which is the most direct test method for the transmission chain at present.

[0005] However, in the aspect of load parameter set construction technology, the following problems are mainly faced: 1) the existing extraction standard of wind condition parameter set has poor applicability and cannot be applied to various climate environments, for example, the IEC 61400-1 standard wind condition model and wind parameters are calculated and simulated according to the specific climate environment characteristics of foreign countries, and errors and distortions will be generated in the wind turbine simulation test in the region of China. 2) The traditional wind turbine simulation software can only accurately simulate the load on the mechanical side of the wind turbine, and the model structure on the power grid side is very simple, and as the structure of the power system becomes more and more complex, it often cannot meet the simulation demand. 3) The simulation working condition data set is large, but the bearing capacity of the transmission chain test platform is limited, and all simulation working conditions cannot be loaded.

[0006] In summary, how to realize the efficient screening and extraction of the load parameters of the wind turbine transmission chain has become a technical problem to be solved in the prior art. SUMMARY

[0007] In view of the deficiencies of the prior art, the purpose of the present application is to provide a load parameter set construction method and system for a wind turbine transmission chain test platform, which improves the extraction standard of existing wind turbine load parameters based on climate environment characteristics, comprehensively simulates multiple working conditions on the power grid side and the wind power machinery side, and screens and extracts key parameters, so as to realize the applicability and rationality of the wind turbine transmission chain test.

[0008] In order to achieve the above-mentioned purpose, the present application is realized by the following technical solutions: The present application provides a load parameter set construction method for a wind turbine transmission chain test platform, comprising the following steps: Divide the wind environment area according to the wind parameter index, and match the corresponding wind environment area according to the simulation task; Based on the wind environment area and the power grid operation characteristics, considering extreme wind conditions, mechanical failures and electrical failures, the running scenarios of multiple working conditions are constructed; According to the running scenarios and the wind turbine transmission chain parameters, simulation is carried out to obtain simulation data; According to the actual requirements of different working conditions, the key parameters are extracted from the simulation data by using the deep learning algorithm combined with the key parameter extraction principle, and the final load parameter set is obtained.

[0009] Further, the wind parameter index includes wind speed, temperature, humidity, terrain, altitude, turbulence, air density, wind shear coefficient and inflow angle.

[0010] Further, the specific steps of matching the corresponding wind environment area according to the simulation task are as follows: Perform white transformation operation on the wind parameter index; According to the simulation task, the wind parameter index is matched by using K-nearest neighbor algorithm to obtain the corresponding wind environment area.

[0011] Further, the specific steps of constructing the running scenarios of multiple working conditions are as follows: According to the extreme wind conditions, mechanical failures and electrical failures, the existing working condition types are combined; The correlation test is performed on various combination modes to optimize the number of scenarios.

[0012] Further, when simulating according to the running scenarios and the wind turbine transmission chain parameters, a data transmission interface is established between the electromagnetic simulation platform and the wind turbine simulation software for joint simulation, and mechanical load and electrical load are simultaneously loaded on the wind turbine transmission chain.

[0013] Further, the principle of extracting key parameters from simulation data by using deep learning algorithm combined with key parameter extraction principle includes: Under normal working conditions, the key parameters are extracted according to the reliability principle, the representativeness principle and the mandatory principle. The key parameters are extracted according to the reliability principle, the representative principle, the mandatory principle and the safety principle under extreme working conditions.

[0014] The second aspect of the application provides a load parameter set construction system of a wind turbine transmission chain test platform, comprising: The region division module is configured to divide wind environment regions according to wind parameter indexes, and match corresponding wind environment regions according to simulation tasks; The scene construction module is configured to construct multi-working-condition operation scenes based on wind environment regions and power grid operation characteristics, considering extreme wind conditions, mechanical faults and electrical faults; The simulation test module is configured to perform simulation according to operation scenes and wind turbine transmission chain parameters to obtain simulation data; The parameter screening module is configured to extract key parameters from simulation data according to actual requirements of different working conditions and by using a deep learning algorithm combined with key parameter extraction principles to obtain a final load parameter set.

[0015] The third aspect of the application provides a computer readable storage medium, which stores a computer program, and the computer program is suitable for being loaded and executed by a processor to perform the steps of the load parameter set construction method of the wind turbine transmission chain test platform according to the first aspect of the application.

[0016] The fourth aspect of the application provides a computer device, which comprises: A processor suitable for executing a computer program; A computer readable storage medium, which stores a computer program, and the computer program is executed by the processor to implement the load parameter set construction method of the wind turbine transmission chain test platform according to the first aspect of the application.

[0017] The fifth aspect of the application provides a computer program product or a computer program, which comprises computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device perform the steps of the load parameter set construction method of the wind turbine transmission chain test platform according to the first aspect of the application.

[0018] The above one or more technical solutions have the following beneficial effects: The application provides a load parameter set construction method and system of a wind turbine transmission chain test platform.

[0019] The wind turbine working condition setting method provided by the application is based on the wind condition parameters measured in different regions, contains highlands, complex mountainous areas, plains, hilly areas, sandy and barren areas and various terrains, covers typhoons, cold waves, earthquakes, ice loads and various extreme conditions, and includes various electrical and mechanical fault working conditions such as power grid power failure, frequency modulation, inertia response and overspeed, yaw error, thereby providing powerful data support for stable and safe operation of wind turbines in different regions and time periods.

[0020] The wind turbine simulation is usually performed by single software simulation, for example, bladed, OpenFAST and the like, even if the joint simulation technology is used, the control strategy of the wind turbine itself is usually focused on, the influence of the power grid on the operation of the wind turbine is rarely considered, and there is no precedent of establishing a high-proportion new energy power system on the power grid side. The application intends to use the iES-GTS full electromagnetic simulation mechanism to build a large power grid electromagnetic simulation platform and perform joint simulation with the GH Bladed wind turbine simulation software, the speed and torque of the wind turbine in the GH Bladed are transmitted to the drag motor model in the iES-GTS, and the voltage, current and other electrical parameters in the iES-GTS are fed back to the controller model of the wind turbine in the GH Bladed.

[0021] Advantages of the additional aspects of the application will be partially given in the following description, partially become obvious from the following description, or be learned by the practice of the application. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description only represent some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0023] Figure 1 The load parameter set construction method flow chart of the wind turbine transmission chain test platform in the embodiment of the present application is provided. Figure 2 The interface schematic diagram of setting the GH Bladed environment parameters in the embodiment of the present application is provided. Figure 3 The other interface schematic diagram of setting the GH Bladed environment parameters in the embodiment of the present application is provided. Figure 4 The joint simulation schematic diagram in the embodiment of the present application is provided. DETAILED DESCRIPTION It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present application. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0024] It should be noted that the terms used herein are only for the purpose of describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and it should also be understood that when the terms "comprise" and / or "include" are used in the specification, there is a presence of a feature, step, operation, device, component and / or combination thereof; The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0025] Embodiment I: The embodiment I of the present application provides a load parameter set construction method of a wind turbine transmission chain test platform. Taking China as an example, as shown in the following figure, the method comprises the following steps: Figure 1 Step 1: According to the wind parameter index, the wind environment area is divided, and the corresponding wind environment area is matched according to the simulation task.

[0026] Step 1.1: According to the wind parameter index, the wind environment area is divided.​

[0027] wherein the wind parameter indicators include wind speed, temperature, humidity, terrain, altitude, turbulence, air density, wind shear coefficient and inflow angle.

[0028] In a specific embodiment, China is divided into different wind environment zones according to the wind parameters of different regions in China, and for offshore wind turbines, the corresponding sea state parameters are additionally measured. Sea state involves the interaction of wind conditions and waves, and the wave speed V is determined by the joint distribution parameters of the wind speed V hub , wave height H s and wave period T p , as shown in Table 1.

[0029] Table 1. Sea state parameter table

[0030] The traditional division method divides China's wind environment zones into 9 categories according to the annual average wind speed and wind energy density: the Qinghai-Tibet Plateau, the northern ventilation corridor, the southern Xinjiang and Jiuquan, the Loess Plateau, the Northeast Plain, the North China Plain, the lower reaches of the Qinghai-Tibet Plateau, the Hengduan Mountains and the Yunnan-Guizhou Plateau, and the Lingnan. However, this method is based on the total wind energy resource development potential at the wind energy utilization height and is usually used for wind farm site selection. In the process of constructing the load parameter set of the transmission chain test platform, more emphasis is placed on the safety of the entire operation of the wind turbine, so the above zoning strategy needs to be improved. In this embodiment, more detailed wind parameter indicators are used, including wind speed, temperature, humidity, terrain, altitude, turbulence, air density, wind shear coefficient and inflow angle, and the threshold values of the wind parameter indicators can be adjusted according to actual conditions. Taking China as an example, 7 typical wind environment zones for wind power are given according to the above wind parameter indicators: Wind Environment Zone 1: Altitude 2000-3000m, turbulence 0.15, air density 0.9, wind shear coefficient 0.2, inflow angle 5°, mainly including eastern Yunnan, southwestern Guizhou, northeastern Shanxi, northwestern Hubei, etc. Wind Environment Zone 2: Altitude 3000-4000m, turbulence 0.16, air density 0.85, wind shear coefficient 0.2, inflow angle 5°, mainly including northwestern Yunnan, Qinghai Province, western Sichuan, etc.

[0031] Wind Environment Zone 3: Low temperature and cold wave region, turbulence 0.12, air density 1.45, wind shear coefficient 0.25, inflow angle 3°, mainly including the three northeastern provinces and the Inner Mongolia grassland region.

[0032] Wind environment zone 4: complex terrain area, turbulence 0.18, air density 1.1, wind shear coefficient 0.25, inflow angle 8°, mainly including Shanxi, Guangxi and other regions.

[0033] Wind environment zone 5: plain high shear zone, turbulence 0.12, air density 1.225, wind shear coefficient 0.35, inflow angle 3°, mainly including North China, East China inland plain, coastal plain and other regions.

[0034] Wind environment zone 6: above 4000 meters above sea level, turbulence 0.18, air density 0.8, wind shear coefficient 0.2, inflow angle 5°, mainly for Qinghai-Tibet Plateau region, at mountain pass and plateau edge.

[0035] Wind environment zone 7: conventional area, turbulence 0.12, air density 1.225, wind shear coefficient 0.2, inflow angle 3°, mainly including the middle and lower reaches of the Yangtze River plain, Guangdong and Fujian non-coastal areas, these areas have stable wind speed, small turbulence, and wind shear coefficient and inflow angle tend to be normal values.

[0036] Step 1.2: match the corresponding wind environment zone according to the simulation task.

[0037] Step 1.2.1: white transformation operation is performed on the wind force parameter index.

[0038] In a specific embodiment, according to the input of the vector [wind speed temperature humidity terrain altitude turbulence air density wind shear coefficient inflow angle], the corresponding wind environment zone can be determined. Taking the above 7 wind environment zones as an example, the attribute values are first whitened to eliminate the correlation and scale difference between the attributes, so that the processed data has unit variance and zero covariance.

[0039] Suppose there are matrix, y samples, b features, the original data matrix is Q, and the whitening transformation steps are as follows: 1) Decentering , .

[0040] is the global average value of each feature, is the feature matrix of the vth sample, is the centering matrix, which has the physical meaning of translating the data to the origin, and is used to eliminate the overall deviation of the data and make the mean value become a zero vector. That is: .

[0041] 2) Calculate the covariance matrix .

[0042] Σ is the covariance matrix, which is a symmetric positive definite matrix with all eigenvalues ​​being positive real numbers, used to represent the linear correlation between features.

[0043] 3) Eigenvalue decomposition .

[0044] Where U is the eigenvector matrix (orthogonal matrix), which physically represents the principal direction of the data. Λ is the diagonal eigenvalue matrix, which physically represents the variance of the corresponding principal components.

[0045] 4) Whitening transformation .

[0046] Its function is to project the data onto the principal component space. Its function is to scale the variance to 1 along each principal component direction.

[0047] Step 1.2.2: Use the K-nearest neighbor algorithm to match wind force parameters based on the simulation task to obtain the corresponding wind environment zone.

[0048] Specifically, after the whitening transformation is completed, the K-nearest neighbor algorithm is used to normalize the attributes first, and then the Euclidean distance is calculated to match the input parameters with the corresponding scene.

[0049] Step 2: Based on the characteristics of the wind environment zone and power grid operation, and considering extreme wind conditions, mechanical failures and electrical failures, construct multi-condition operation scenarios.

[0050] Step 2.1: Combine the existing operating conditions based on extreme wind conditions, mechanical failures, and electrical failures.

[0051] In one specific implementation, this embodiment constructs normal and extreme fault operation scenarios based on wind conditions and power grid operation characteristics in different regions, and studies the combination of extreme wind conditions, mechanical and electrical fault conditions. Considering the characteristics of power grid operation refers to constructing a high-proportion renewable energy power system within a large-scale power system electromagnetic simulation platform.

[0052] The working condition types include normal power generation, normal power generation with fault, starting, stopping, emergency stopping, idling, idling with fault, and maintenance. The extreme wind conditions include normal turbulence model (NTM), normal wind profile model (NWP), extreme turbulence model (ETM), extreme direction change (EDC), extreme gust model (ECD), extreme wind shear model (EWS), extreme operating gust model (EOG), and extreme wind speed model (EWM), etc. The faults include electrical faults and mechanical faults. The electrical faults include short-circuit fault, sensor failure fault, yaw fault, variable pitch fault, and radar fault. The mechanical faults include NA overspeed, single blade feathering runaway, and single blade pitch runaway.

[0053] Step 2.2: Perform correlation test on various combination modes to optimize the number of scenarios.

[0054] The RF model features are optimized to extract fault influence features.

[0055] In a specific embodiment, the original data in the fault data includes [fault type, fault duration, wind speed, wind direction, temperature, humidity, terrain, altitude, turbulence, air density, wind shear coefficient, inflow angle, wind model], an RF model is established for the variables, and variables with higher contribution to the model are obtained.

[0056] RF is an integrated machine learning algorithm that uses Bootstrap and node random splitting technology to build multiple decision trees, and obtains the final classification result by voting. The feature selection technology MeanDecreaseGini in the node random splitting technology is the basis for selecting feature variables. MeanDecreaseGini calculates the heterogeneity of each variable on each node of the classification tree through the Gini index, so as to compare the importance of variables. That is, the contribution of each feature to each tree in RF is judged, and then the average value is taken. The larger the value, the greater the importance of the variable.

[0057] The Gini coefficient represents the probability that a randomly selected sample is misclassified. Gini coefficient = probability of sample being selected The probability that a sample is misclassified. Assuming that there are S classes, the probability that a sample point belongs to the s-th class is p s Therefore, the Gini coefficient formula is: .

[0058] The Gini coefficient gain represents the gain of Gini coefficient of a certain division. The original Gini coefficient is subtracted from the Gini coefficient of each branch weighted by the sample proportion to calculate it. The larger the Gini coefficient gain, the better the division. Specifically as follows: For the data set D, assume the sample size is d, and for the variable Z, assume there are r partitions, i.e. Z = [z1, z2,..., z r ], and the corresponding sample size of the set is d = [d1, d2,..., dr] r ], then The Gini coefficient of D is: .

[0059] The Gini coefficient gain Gain(D, ) is: Gain(D, ) = Gini(D) - Gini(D, ).

[0060] In summary, the effective feature variables are identified, and the key variables [fault type, wind speed, wind direction, terrain, altitude, and wind model] are obtained. According to the original data, the key variables are divided into bins, and the working condition combinations in the following table are obtained.

[0061] Table 1. Selected working condition combinations

[0062] Step 3: Perform simulation according to the operating scenario and wind turbine transmission chain parameters to obtain simulation data.

[0063] In a specific embodiment, when performing simulation according to the operating scenario and wind turbine transmission chain parameters, a data transmission interface is established between a large-scale power system electromagnetic simulation platform and wind turbine simulation software for joint simulation, and mechanical load and electrical load are simultaneously loaded on the wind turbine transmission chain. In the wind turbine simulation software interface, load the unit, tower, and controller model, input the wind and sea conditions parameters, form the normal and extreme working condition load characteristics and the limit load spectrum. Input the corresponding unit model, tower model, controller model, and environmental parameters into GH Bladed, as shown in Figure 2 and Figure 3 , and mark the input positions of the sea condition parameters wave speed V , wave height H s , and wave period T p .

[0064] Step 3.1: Build a high-proportion new energy power system. Build a high proportion of new energy power system, in the power system large full electromagnetic simulation platform for the joint simulation of subsequent preparation work. The application uses Shandong Jicheng Electronics iES-GTS full electromagnetic simulation mechanism to build a large power grid electromagnetic simulation platform. Because each regional power grid company in China uses PSASP software as their own simulation analysis tool, therefore using iES-GTS system not only can realize localization, the biggest advantage is that the software can import domestic PSASP (synchro-stable) model, provide a convenient channel for further research.

[0065] Step 3.2: Establish a simulation interface.

[0066] Bladed Hardware Test module is a bridge to realize the joint simulation of GH Bladed and other software. Bladed Hardware Test allows to test a real control system in the laboratory, factory and field, which runs on an embedded platform. The data exchanged between Bladed and the hardware test module is the same as the data exchanged between Bladed and the external controller in Bladed DLL format. The signal names are visible in Bladed, and the hardware test module realizes data read and write with Bladed software through these data interfaces, and at the same time, new variable names are specified to be associated with these data interfaces, and the new variable interfaces are used to realize data interaction with the hardware controller, realizing the joint simulation mode, as shown in Figure 4 .

[0067] Step 4: According to the actual requirements of different working conditions, the key parameter extraction principle is used to extract the key parameters of the simulation data, and the final load parameter set is obtained.

[0068] Step 4.1: Set the key parameter extraction principle.

[0069] Step 4.1.1: Under normal working conditions, the key parameters are extracted according to the reliability principle, the representative principle and the mandatory principle.

[0070] 1. Normal-reliability principle, including bearing reliability. The sliding bearing has the advantages of small radial size, strong bearing capacity and low cost, and can be applied to replace the rolling bearing in the wind power transmission system. In the test stage, it is necessary to detect the service life under non-steady state conditions such as frequent start-stop of wind turbine and impact load. This refining principle includes start-up conditions DLC31, DLC32, DLC33 and shutdown conditions DLC41, DLC42.

[0071] 2、Normal-representative principle, including the demand of wind power development in high altitude area and low temperature environment. The wind energy resources are abundant in plateau area and the three provinces in the northeast of China, but the climate environment is also poor. Whether the wind turbine can run stably in the corresponding environment needs to test the corresponding working condition. This refining principle includes normal power generation working conditions DLC121, DLC122, DLC123 and DLC126.

[0072] 3、Normal-mandatory principle, including mechanical load measurement procedures of the unit, grid adaptability test procedures and recommended load working conditions.

[0073] Step 4.1.2: Extract key parameters according to reliability principle, representative principle, mandatory principle and safety principle under extreme working conditions.

[0074] 1、Extreme-reliability principle, including elastic bearing life reliability requirement, planetary pin shaft error reliability requirement, and wind and wave coupling long-term effect reliability requirement of offshore wind turbine. The load at the wind wheel and the load at the output end of the gearbox can be transmitted to the elastic support. Due to the existence of wind shear, the aerodynamic force received by the wind turbine during operation is different at different azimuth angles, which causes the overall shift of the time domain vibration signal of the elastic support. In actual turbulent wind conditions, when an elastic support is damaged and cannot provide sufficient stiffness, the vibration will increase rapidly and pose a serious threat to the normal operation of the wind turbine. The elastic bearing life reliability requirement includes extreme working conditions DLC13, DLC14 and DLC15. The planetary gear train in the wind turbine gearbox has typical compound motion, and its fatigue life analysis is very complex. Due to the inevitable manufacturing and installation errors of the pin shaft supporting the planetary gear, the load sharing performance of the planetary gear train is affected, leading to reduced bearing life. The planetary pin shaft error reliability requirement includes extreme working conditions DLC24. Under the action of random wind and wave coupling, the transmission chain of offshore wind turbine will be affected by random aerodynamic load and inertial load due to large-scale movement of the supporting platform. Compared with the transmission chain of onshore wind turbine, the vibration characteristics of offshore wind turbine are more complex. The wind and wave coupling long-term effect reliability requirement of offshore wind turbine includes the corresponding offshore working conditions of DLC13, DLC14 and DLC15.

[0075] 2. The principle of extreme representativeness includes the requirements for typhoon resistance and seismic resistance of wind turbines. Typhoons are usually accompanied by strong winds exceeding 50 meters per second. Onshore wind turbines are designed to withstand high wind speeds, but when the wind speed approaches or exceeds the turbine's limit, it can damage the turbine structure and even lead to failure or breakage. High wind speeds not only increase the load on the blades but may also cause blade deformation or breakage, especially when the turbine is not shut down in time or the blades are subjected to unbearable wind pressure due to excessive wind speed. Offshore wind turbines are more severely affected by typhoons than onshore wind turbines. When a typhoon strikes, wind speed increases sharply and wind direction changes drastically. The turbine must cope with the sudden and enormous wind force, which places extremely high demands on the turbine blades, rotor, and generator system. Typhoons not only bring strong winds but also large waves and ocean waves. Offshore wind turbines are typically installed on the seabed. The giant waves of typhoons can damage the supporting structures (such as foundations and towers), and may even cause the turbines to collapse or suffer severe damage. Therefore, to assess the impact of typhoons on the structural safety of wind turbines, it is necessary to set up typhoon-enveloping operating conditions to simulate extreme wind conditions and extreme loads, providing technical support for the operation and maintenance of wind turbines in the southeastern coastal areas. The typhoon resistance requirements for wind turbines include extreme operating conditions such as idling (DLC61, DLC62, DLC63, DLC64) and idling + fault (DLC71). Wind turbines are devices that are highly dependent on structural stability and precision mechanical systems. Issues such as vibrations, foundation changes, and mechanical damage caused by earthquakes may have a significant impact on the safety, performance, and service life of wind turbines. Precision equipment such as gearboxes, generators, and electrical control systems of wind turbines may be damaged by vibrations during earthquakes. Strong vibrations can cause wear and tear on mechanical parts, loss of precision, or failure of internal electronic components, affecting the normal operation of the unit. In order to mitigate the impact of earthquakes, wind turbines must be designed to simulate earthquake conditions and assess the seismic resistance of the unit. The seismic resistance requirements of wind turbines include earthquake conditions DLC95, DLC96, and DLC97.

[0076] 3. Extreme-mandatory principle, including load measurement operating condition procedures, wind turbine fault voltage ride-through capability test procedures, and wind turbine grid adaptability test procedures. 4. The principle of extreme-safety, including the ultimate load spectrum. The working conditions used for testing on the test bench need to be within the ultimate load range of the entire working condition. Therefore, the test bench needs to be expanded according to the ultimate load spectrum, and the corresponding working conditions when a certain degree of freedom load reaches the limit need to be selected.

[0077] Step 4.2: Use deep learning algorithms combined with key parameter extraction principles to extract key parameters from simulation data.

[0078] The embodiment utilizes a random forest algorithm to screen the load parameter set envelope, automatically determines the number of load modes through mutual reachable distance and stability pruning, greatly reduces the complexity of working condition processing based on feature compression and hierarchical pruning, can also explicitly separate extreme working conditions, and finally extracts representative load working conditions for transmission chain test platform testing. The specific steps are as follows: Step 4.2.1: Construct a multi-channel feature space according to the working condition combination.

[0079] Input working condition set G={W j}, N is the total number of working conditions, wherein W j ∈R T×6 is the six-degree-of-freedom time sequence of the jth working condition, and R is a real number set. Feature extraction is performed on the working condition set, and for each channel c (Fx, Fy, Fz, Mx, My, Mz), the following is calculated: .

[0080] Among them, denotes the feature of the jth working condition in channel c, Fx, Fy, and Fz are the forces along the x, y, and z axes at the fan hub, respectively, and Mx, My, and Mz are the moments along the x, y, and z axes at the fan hub, respectively. c μc is the mean value of channel c, σ c is the standard deviation of channel c, skewnessc is the skewness of channel c, kurtosisc is the kurtosis of channel c, RMs c is the root mean square value of channel c, is the absolute maximum value of channel c.

[0081] Then the multivariate joint feature is represented as: .

[0082] Among them, the dimension feature d=6×8=48.

[0083] This step compresses the long time sequence into a fixed dimension feature, retaining the statistical characteristics and spectral energy distribution.

[0084] Step 4.2.2: Define distance metrics to meet the similarity of fan loads.

[0085] The formula is:

[0086] Among them, is the mutual reachable distance of working conditions i and j, k is the local density parameter, is the distance from working condition i to its kth nearest neighbor, is the dynamic time warping distance, and the calculation formula is as follows:

[0087] wherein, is the set of time-corresponding paths, is the feature vector is the pth element of the feature vector, is the aligned path by dynamic time warping, i.e. . Through the calculation of the above distance metric, this embodiment can adaptively handle the phase shift and local density variation of the fan load.

[0088] Step 4.2.3: Weighted core density estimation for each operating condition.

[0089] Calculate the normalized density weight: .

[0090] wherein is the normalized density weight of operating condition i, is the set of k nearest neighbors index of operating condition i, is the distance median. This embodiment quantifies the density significance of each operating condition in the feature space through the above formula, and identifies the high load concentration area.

[0091] Step 4.2.4: Construct a hierarchical density clustering tree.

[0092] Construct a minimum spanning tree (MST) according to the mutual reachable distance, and cut the MST in descending order of distance threshold , each point being a cluster, when , merge cluster and cluster , generate a tree graph:

[0093] wherein, is the set of clustering trees, is the clustering result under threshold , k m is the number of clusters of the mth layer, is the k m th cluster of the mth layer, is the distance threshold of the mth cut. According to the above formula, a multi-scale clustering structure is established to adapt to different load mode granularities.

[0094] Step 4.2.5: Define a stability function and extract the optimal stable clustering by pruning.

[0095] Define the stability function and solve the optimal pruning: , , .

[0096] Optimal cluster set: .

[0097] Noise identification: .

[0098] wherein, is the stability measure of cluster C, defined as the sum of all points' residence time within the cluster, is the distance threshold for a point x to join the cluster for the first time, is the distance threshold for a point to leave the cluster, is the stability threshold. The embodiment automatically determines the number of load mode categories with clear physical meaning through the above extraction process, without presetting the K value.

[0099] Step 4.2.6: Extract representative working conditions based on the optimal stable cluster.

[0100] For each optimal cluster extract the typical working condition, the density center: .

[0101] Representative working condition: .

[0102] Output .

[0103] Finally, is the density center working condition of cluster , the sum of DTW distances within the cluster is minimum, is the extreme working condition of cluster , the maximum feature vector is taken, is the infinite norm of the feature vector, and the embodiment extracts a total of 10 representative working conditions for the normal working condition through the above method for the transmission chain test bed test, and a total of 80 representative working conditions for the extreme working condition for the transmission chain test bed test.

[0104] Embodiment two: The embodiment two of the application provides a load parameter set construction system of a wind turbine transmission chain test platform, which comprises: A region division module is configured to divide wind environment regions according to wind parameter indexes, and match corresponding wind environment regions according to simulation tasks; A scene construction module is configured to construct a multi-working condition operation scene based on wind environment regions and power grid operation characteristics, considering extreme wind conditions, mechanical faults and electrical faults; A simulation test module is configured to simulate according to operation scenes and wind turbine transmission chain parameters to obtain simulation data; The parameter screening module is configured to extract key parameters from simulation data according to actual requirements of different working conditions by using a deep learning algorithm in combination with a key parameter extraction principle, so as to obtain a final load parameter set.

[0105] Embodiment three: The embodiment three of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is suitable for being loaded and executed by a processor to implement the steps in the load parameter set construction method of the wind turbine transmission chain test platform according to the embodiment one of the present application.

[0106] Embodiment four: The embodiment four of the present application provides a computer device, which comprises: a processor suitable for executing a computer program; a computer readable storage medium, which stores a computer program, and the computer program is executed by the processor to implement the steps in the load parameter set construction method of the wind turbine transmission chain test platform according to the embodiment one of the present application.

[0107] Embodiment five: The embodiment five of the present application provides a computer program product or a computer program, which comprises computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to make the computer device execute the steps in the load parameter set construction method of the wind turbine transmission chain test platform according to the embodiment one of the present application.

[0108] The steps and method embodiments one corresponding to the above embodiments two, three, four and five are described in the specific implementation manner, and the specific implementation manner can be referred to the related description part of the embodiment one.

[0109] Those skilled in the art can be aware that units and algorithm steps of each example described in combination with the embodiments disclosed in the present application can be realized in electronic hardware or computer software and combination of electronic hardware and computer software. Whether the functions are realized in hardware or software mode depends on specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application. In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to the embodiments of the present application are generated. The computer can be a general purpose computer, a special purpose computer, a computer network or other programmable device. The computer instructions can be stored in or transmitted by a computer readable storage medium. The computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through a wired (for example, coaxial cable, optical fiber, digital line (DSL)) or wireless (for example, infrared, wireless, microwave, etc.) manner. The computer readable storage medium can be any available medium that can be accessed by a computer or a data processing device such as a server, data center, etc. integrated with one or more available media. The available media can be a magnetic medium (for example, floppy disk, hard disk, magnetic tape), an optical medium (for example, DVD) or a semiconductor medium (for example, solid state disk (SSD)) and the like. The above description is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for constructing a load parameter set of a wind turbine generator transmission chain test platform, characterized in that, The method comprises the following steps: According to the wind parameter index, the wind environment area is divided, and the corresponding wind environment area is matched according to the simulation task; Based on the wind environment area and the power grid operation characteristics, considering extreme wind conditions, mechanical failures and electrical failures, a multi-condition operation scene is constructed; According to the operation scene and the wind turbine transmission chain parameters, simulation is carried out to obtain simulation data; According to the actual requirements of different working conditions, the simulation data is extracted by using the deep learning algorithm combined with the key parameter extraction principle to obtain the final load parameter set.

2. The method for constructing the load parameter set of the wind turbine drive train test platform as described in claim 1, characterized in that, The wind parameter index includes wind speed, temperature, humidity, terrain, altitude, turbulence, air density, wind shear coefficient and inflow angle.

3. The method of claim 2, wherein the load parameter set is constructed by: The specific steps of matching the corresponding wind environment area according to the simulation task are: Performing white transformation operation on the wind parameter index; According to the simulation task, the wind parameter index is matched by using K nearest neighbor algorithm to obtain the corresponding wind environment area.

4. The method of claim 1, wherein, The specific steps of constructing the multi-condition operation scene are: According to the extreme wind conditions, mechanical failures and electrical failures, the existing working condition types are combined; The correlation test is carried out on various combination modes to optimize the scene quantity.

5. The method of claim 1, wherein, When simulating according to the operation scene and the wind turbine transmission chain parameters, a data transmission interface is established between the electromagnetic simulation platform and the wind turbine simulation software for joint simulation, and mechanical load and electrical load are simultaneously loaded on the wind turbine transmission chain.

6. The method of claim 1, wherein, The principles of extracting key parameters from simulation data by using deep learning algorithm combined with key parameter extraction principle include: Under normal working conditions, the key parameters are extracted according to the reliability principle, the representative principle and the mandatory principle; Under extreme working conditions, the key parameters are extracted according to the reliability principle, the representative principle, the mandatory principle and the safety principle.

7. A load parameter set construction system of a wind turbine generator set drive chain test platform, characterized in that, It comprises: The area division module is configured to divide the wind environment area according to the wind parameter index, and to match the corresponding wind environment area according to the simulation task; The scene construction module is configured to construct a multi-condition operation scene based on the wind environment area and the power grid operation characteristics, considering extreme wind conditions, mechanical failures and electrical failures; The simulation test module is configured to simulate according to the operation scene and the wind turbine transmission chain parameters to obtain simulation data; The parameter screening module is configured to extract key parameters from simulation data by using deep learning algorithm combined with key parameter extraction principle according to the actual requirements of different working conditions to obtain the final load parameter set.

8. A computer program product, characterised in that, The computer program product comprises a computer program, which is executed by a processor to realize the load parameter set construction method of the wind turbine transmission chain test platform according to any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, which is suitable for being loaded and executed by a processor to realize the load parameter set construction method of the wind turbine transmission chain test platform according to any one of claims 1-6.

10. A computer device, comprising: It comprises: A processor suitable for executing a computer program; A computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the load parameter set construction method of the wind turbine transmission chain test platform according to any one of claims 1-6.

Images