A method and system for rating the power curve of a wind turbine

Abstract

This invention discloses a calibration method and system for hierarchical denoising of wind turbine power curves. The method includes acquiring historical operating data sequences of each wind turbine in a wind farm, and dividing the historical operating data sequences into several subsets based on a preset environmental dynamic change cycle; constructing a progressive abnormal data identification system based on the design operating parameters and physical operating laws of the wind turbines; removing ineffective power generation operating data points of the wind turbines based on the identification results of the subset data to obtain an effective operating dataset; iteratively optimizing the effective operating dataset, obtaining the relative change rate of each round of iterative optimization, until the relative change rate reaches a preset convergence criterion, and determining the optimal power curve calibration result. This invention, through the above systematic hierarchical denoising and calibration process, significantly improves the physical consistency and state identification accuracy of data denoising, and is applicable to application scenarios such as wind turbine performance evaluation, wind power prediction, wind farm post-evaluation, and operation and maintenance optimization.

Description

Technical Field

[0001] This invention relates to the field of wind turbine technology, specifically to a calibration method and system for layered noise reduction of wind turbine power curves. Background Technology

[0002] The wind power industry primarily utilizes large-capacity, long-bladed wind turbines, which place higher demands on the control system design. In recent years, extreme weather events have led to numerous failures in key components such as wind turbine blades and pitch bearings, resulting in frequent instances of reduced power output. Coupled with grid curtailment affecting wind farms, this has resulted in some operating wind farms experiencing significantly lower annual on-grid power generation than designed, leading to substantial losses for power generation companies and hindering the development of wind power. Therefore, it is urgent to evaluate wind turbine performance, analyze power curves, identify the causes of insufficient generator performance, and improve the overall efficiency of wind farms.

[0003] During operation, the power curve of wind turbines is susceptible to dispersion due to various factors. Specifically, equipment failures, power curtailment, errors in wind measurement after the nacelle, and complex flow field disturbances (such as wakes) all introduce numerous outliers into the raw data. These outliers not only obscure the true power generation capacity of the turbines but also directly reduce the reliability of subsequent performance evaluations and power generation predictions. Currently, two main methods are used to remove outliers: one is the variance method and the median absolute deviation method; the other is machine learning models such as cluster analysis, isolated forests, and autoencoders. The core principle of these models is to learn the overall distribution characteristics of the dataset and identify statistical outliers that deviate from the main pattern.

[0004] However, the fundamental limitation of the above methods lies in the lack of physical interpretability, the incomplete elimination of outliers, and the difficulty in establishing a causal relationship between statistical anomalies and the specific operating status of the wind turbine. Summary of the Invention

[0005] Based on this, the purpose of this invention is to provide a calibration method and system for layered noise reduction of wind turbine power curves, which aims to solve the problems of current methods for removing outliers lacking physical interpretability, incomplete removal of outliers, and difficulty in establishing a causal relationship between statistical anomalies and the specific operating status of wind turbines.

[0006] To achieve the above objectives, this invention proposes a calibration method for layered noise reduction of wind turbine power curves, the method comprising: The historical operating data sequence of each wind turbine in the wind farm is obtained, and the historical operating data sequence is divided into several subsets based on a preset environmental dynamic change cycle; Based on the design and operation parameters and physical operation laws of wind turbines, a progressive abnormal data identification system is constructed. Based on the identification results of subset data, ineffective power generation operation data points of wind turbines are eliminated to obtain effective operation datasets. The effective running dataset is iteratively optimized to obtain the relative rate of change of each round of iterative optimization until the relative rate of change reaches the preset convergence criterion, and the optimal power curve calibration result is determined.

[0007] According to one aspect of the above technical solution, in the step of acquiring the historical operating data sequence of each wind turbine in the wind farm and dividing the historical operating data sequence into several subsets based on a preset environmental dynamic change cycle: Obtain historical operating data sequences of each wind turbine unit in the wind farm at interval sampling periods, and construct a numerical matrix using the historical operating data sequences. The historical operating data sequences include at least the average wind speed and the active power generated.

[0008] in, It is a numerical matrix. Representing a time series, Indicates the number of wind turbine units. Represents wind speed. Represents power, Indicates the first No. 1 fan, No. Wind speed values ​​for a given time period; Indicates the first No. 1 fan, No. Power values ​​over a time period =1, ..., ; =1, ..., .

[0009] According to one aspect of the above technical solution, after constructing the numerical matrix, the historical operational data sequence is divided into several subsets based on a preset environmental dynamic change cycle: , , ... The number of subsets is ,and ≥ ≥1.

[0010] According to one aspect of the above technical solution, the step of constructing a progressive anomaly data identification system and removing ineffective power generation operation data points of wind turbine units based on the identification results of subset data includes: Based on the design and operation parameters of the wind turbine, the cut-in wind speed threshold and the rated power of the wind turbine are obtained. Data points with wind speeds lower than the cut-in wind speed threshold and data points with wind speeds not lower than the cut-in wind speed threshold and active power generation less than or equal to zero are removed to obtain the initial screening dataset. The power generation capacity between full power generation and shutdown is divided into several power boxes with equal spacing. Based on the average value and variance of each power box, the relative stability index of the power box is calculated, and the wind speed fluctuation index is calculated using the wind speed difference in each power box. The power stability and wind speed fluctuation of the power curtailment characteristics are judged by using the relative stability index and the wind speed fluctuation index respectively. Once the power bins achieve the desired power stability and wind speed fluctuation, the linear correlation between wind speed and generated power in the power bins is obtained. When the linear correlation reaches the desired level, the current power bin is determined to be curtailed data and is removed to obtain a second-screen dataset.

[0011] According to one aspect of the above technical solution, after obtaining the second-screen dataset, the wind speeds within the range of wind speeds entering and exiting the wind turbine are divided into bins to obtain several wind speed bins of equal width. The average power within each wind speed compartment Based on this, calculate the first... No. 1 fan, No. Power deviation of each data point within the wind speed sub-compartment :

[0012]

[0013] in, For the first No. 1 fan, No. Within the wind speed distribution box, the first The active power value of each data point. For the first No. 1 fan, the The number of data points within each wind speed sub-box For the first Power within each wind speed compartment The arithmetic mean of the power deviation At that time, the wind speed distribution box showed a severe dispersion point and was therefore rejected.

[0014] According to one aspect of the above technical solution, after the abnormal power deviation detection of each wind speed sub-box is completed, the corresponding power deviation sequence is extracted and the median deviation is calculated. Based on the median deviation, calculate the absolute value sequence and median absolute deviation between each deviation value and the median deviation. Data points with an absolute value sequence greater than a preset multiple of the median absolute deviation are identified as outliers and removed to obtain a valid running dataset.

[0015] According to one aspect of the above technical solution, in the step of iteratively optimizing the effective running dataset, obtaining the relative rate of change of each round of iterative optimization, until the relative rate of change reaches a preset convergence criterion, and determining the optimal power curve calibration result: Using the median absolute deviation of a preset multiple as a threshold, the median absolute deviation is tightened round by round by decreasing the multiple until the relative change rate of the median power in each wind speed compartment in the current round reaches the preset convergence standard. The current power curve is then determined to be converged, and the power curve of this round is determined to be the optimal power curve calibration result.

[0016] This invention also provides a calibration system for layered noise reduction of wind turbine power curves. The system is used to implement the aforementioned calibration method for layered noise reduction of wind turbine power curves. The system includes: The sequence segmentation module is used to acquire the historical operating data sequence of each wind turbine in the wind farm, and to segment the historical operating data sequence into several subsets based on a preset environmental dynamic change cycle; The anomaly identification module is used to build a progressive anomaly data identification system based on the design operating parameters and physical operating laws of wind turbines. Based on the identification results of subset data, it removes ineffective power generation operation data points of wind turbines and obtains effective operation datasets. The curve calibration module is used to iteratively optimize the effective running dataset, obtain the relative rate of change of each round of iterative optimization, until the relative rate of change reaches the preset convergence criterion, and determine the optimal power curve calibration result.

[0017] The present invention also proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the calibration method for layered noise reduction of wind turbine power curves as described above.

[0018] The present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the calibration method for layered noise reduction of wind turbine power curves as described above.

[0019] In summary, the hierarchical denoising calibration method for wind turbine power curves proposed in this invention overcomes the limitations of traditional annual power curve calibration methods that ignore dynamic changes in environmental parameters and operational state transitions, as well as the shortcomings of purely data-driven machine learning methods that lack physical interpretability and robustness, by deeply integrating the physical mechanisms of wind turbine operation with multi-level statistical diagnostic logic. Through this systematic hierarchical denoising and calibration process, the physical consistency and state identification accuracy of the data are significantly improved, making it suitable for applications such as wind turbine performance evaluation, wind power prediction, wind farm post-evaluation, and operation and maintenance optimization.

[0020] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0021] Figure 1 This is a flowchart of the calibration method for layered noise reduction of wind turbine power curves in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the calibration system for layered noise reduction of wind turbine power curves in Embodiment 2 of the present invention; Figure 3 This is a structural block diagram of the electronic device in Embodiment 4 of the present invention. Detailed Implementation

[0022] To make the objectives, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Several embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of the present invention will be more thorough and complete.

[0023] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there may be an intervening element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be an intervening element. The terms "vertical," "horizontal," "left," "right," "upper," "lower," and similar expressions used herein are for illustrative purposes only and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as limiting the invention.

[0024] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances. The term "and / or" as used herein includes any and all combinations of one or more of the related listed items.

[0025] Example 1 like Figure 1 The diagram shows a flowchart of a method for calibrating a wind turbine power curve with layered noise reduction according to Embodiment 1 of the present invention. The method includes the following steps S01-S03, wherein: S01. Obtain the historical operating data sequence of each wind turbine in the wind farm, and divide the historical operating data sequence into several subsets based on a preset environmental dynamic change cycle; S02. Based on the design operating parameters and physical operating laws of wind turbine units, a progressive abnormal data identification system is constructed. Based on the identification results of subset data, ineffective power generation operation data points of wind turbine units are eliminated to obtain effective operation datasets. S03. Iteratively optimize the effective running dataset, obtain the relative rate of change of each round of iterative optimization, until the relative rate of change reaches the preset convergence criterion, and determine the optimal power curve calibration result.

[0026] Historical operating data sequences of each wind turbine in the wind farm are obtained during the sampling period of their operating intervals. A numerical matrix is ​​constructed using these historical operating data sequences. The historical operating data sequences include at least the average wind speed and the active power generated. In this embodiment, a sampling period of 10 minutes is used as an example.

[0027] in, It is a numerical matrix. Representing a time series, Indicates the number of wind turbine units. Represents wind speed. Represents power, Indicates the first No. 1 fan, No. Wind speed values ​​for a given time period; Indicates the first No. 1 fan, No. Power values ​​over a time period =1, ..., ; =1, ..., .

[0028] After constructing the numerical matrix, the historical operational data sequence is segmented based on a preset environmental dynamic change period. Subset: , , ... The number of subsets is ,and ≥ ≥1.

[0029] After being divided into several subsets, abnormal data is identified in each subset in turn. Based on the identification results of the subset data, non-effective power generation operation data points of the wind turbine are removed.

[0030] First, an initial screening is conducted to remove data related to shutdowns or outages: Based on the design and operating parameters of wind turbines, define The cut-in wind speed threshold for wind turbine units is typically set to 2.5 m / s or 3 m / s. The rated power of the wind turbine is derived from the numerical matrix. , ,......, Remove all data points that meet any of the following conditions: wind speed Below the unit cut-in wind speed threshold The data points were removed because the wind turbine had not yet started generating electricity; secondly, the wind speed... The cut-in wind speed threshold has been reached or exceeded. However, the active power of power generation Data points less than or equal to zero are discarded to obtain the initial screening dataset.

[0031] After denoising the data using the above method, exclude fully denoised data. ≥ 0.95 ) and shutdown ( = 0) state, power Greater than 0 and less than 0.95 times the rated power The interval was divided into 190 equal-width intervals for binning:

[0032] in, For the first No. 1 fan, the The power collection within each power distribution box The recommended width for each box is 0.5%. , The value range of is [1, 190], with the th being the . Taking the power distribution of typhoon generators as an example, the distribution result is: [0, ), [ ,2 ),[2 3 ), ..., [189 190 ).

[0033] After obtaining the power distribution boxes, the power stability of the power distribution boxes is judged based on the power curtailment characteristics. When power curtailment occurs, the power fluctuates with wind speed and is basically stable. The power fluctuation characteristics that meet the power curtailment criteria can be identified by calculating the relative stability index of the power in each power distribution box.

[0034] First, calculate the average value within each power sub-box. and :

[0035]

[0036] in, For the first The average value of a single box within each sub-box For the first The sample standard deviation of all power values ​​within each power bin For where Number of points in the sub-box For the first Inside the power distribution box, the first The active power value of each data point. Representing the No. 1 wind turbine.

[0037] Further calculation of power stability index :

[0038] Among them, power stability index , is a dimensionless constant. This index represents the first... The dispersion of data within each power bin relative to its average power, when the power stability index When the value is below 0.1, the power output within the sub-box is considered to be in a stable state, which is consistent with the characteristics of power fluctuations caused by power rationing.

[0039] Complete power stability assessment and determine wind speed fluctuations to reflect power curtailment characteristics: In the power sub-bins judged to be power stable, the wind speed fluctuation index is calculated. :

[0040] Among them, wind speed fluctuation index For the first Maximum wind speed in each power distribution box Subtract the minimum wind speed .in For the first No. 1 fan Collective wind speed within the compartment For the first The maximum wind speed corresponding to all data points in each power sub-box is obtained by traversing all wind speed records in the sub-box, reflecting the upper limit of the wind speed in that sub-box. For the first The minimum wind speed corresponding to all data points within a power sub-box reflects the lower limit of the wind speed within that sub-box. For the first The range of wind speed variation in each power distribution box, when When the speed exceeds a certain threshold (generally 3 m / s), it meets the characteristics of wind speed fluctuations during power rationing.

[0041] After a power distribution box simultaneously meets both the power stability and wind speed fluctuation criteria, further verification of the correlation between power and wind speed is required. Under normal operating conditions, power and wind speed should exhibit a strong positive correlation; however, under power curtailment conditions, changes in wind speed do not cause synchronous changes in power, and the correlation between the two is significantly reduced. Therefore, this embodiment uses the Pearson correlation coefficient. Perform quantitative discrimination:

[0042]

[0043]

[0044] in, For the first No. 1 fan, No. Inside the power distribution box, the first The active power value of each data point. For the first Power in each power compartment The arithmetic mean; The corresponding wind speed value. Indicates the first Wind speed in each power compartment The arithmetic mean, For the first Typhoon turbine, the The Pearson correlation coefficient between power and wind speed within a power sub-division is dimensionless and ranges from -1 to 1. This coefficient measures the strength of the linear correlation between the two variables. When the absolute value is less than the threshold of 0.2, it is considered that there is a weak correlation between power and wind speed, and it is judged as power rationing data.

[0045] The above three levels of power stability indicators are used to determine power stability. <0.1), wind speed fluctuation index ( >3) and indicators with weak correlation between power and wind speed ( <0.2) Identify and discard data points that are in a power-limited operation state to obtain a second-screen dataset.

[0046] After obtaining the second-screen dataset, the wind speed at which the wind turbine cuts in is... and fan output speed The wind speed within the interval is divided into boxes, with a box width of 0.5 m / s. , , ... represent wind speed ranges of [1.75, 2.25), [2.25, 2.75), [2.75, 3.25), ... respectively; The average power within each wind speed compartment Based on this, calculate the first... No. 1 fan, No. Power deviation of each data point within the wind speed sub-compartment :

[0047]

[0048] in, For the first No. 1 fan, No. Within the wind speed distribution box, the first The active power value of each data point. For the first No. 1 fan, the The number of data points within each wind speed sub-box For the first Power within each wind speed compartment The arithmetic mean.

[0049] Within each wind speed sub-box, a deviation threshold of 0.2 times the power arithmetic mean is set. When power deviation At that time, the wind speed distribution box showed a severe dispersion point and was therefore rejected.

[0050] After the power deviation anomaly detection of each wind speed compartment is completed, the corresponding power deviation sequence is extracted. Calculate the median deviation ;

[0051] in, Used to find the median of a score bin sequence. For the first No. 1 fan, No. The median deviation of each data point within each wind speed sub-box during the first iteration.

[0052] Based on the median deviation, calculate the absolute value sequence between each deviation value and the median deviation. and median absolute deviation :

[0053] set up times As a threshold, the initial multiple Q=8, the absolute value sequence Exceed Data points that meet the threshold are considered outliers. This point is removed to obtain a valid running dataset.

[0054] Referring to the steps above for calculating the median absolute deviation, set... times As a threshold, the median absolute deviation is progressively tightened by decreasing multiples, from 8 times to 7 times, 6 times, and so on. After each round of denoising, the median power in each wind speed sub-box is recalculated until the relative change rate of the median power in each wind speed sub-box from the previous round is less than 0.5%. At this point, the power curve is considered converged, and the power curve for this round is the recommended calibration result. Specifically: First, in the above steps, multiply by 8. After removing discrete data using a threshold, the data is used as the initial data for the next iteration. The median absolute deviation is tightened in successive rounds by decreasing the multiple, reducing the median threshold to 7 times, 6 times, 5 times, 4 times...; Recalculate the absolute value of the power deviation, and identify data points whose absolute power deviation exceeds the threshold as outliers and remove them. When the relative change rate of the median power in each wind speed compartment is less than 0.5% compared to the previous round, the power curve is considered to have converged. The power curve for this round is the recommended calibration result, and the calculation formula is as follows: Let the iteration after the (r-1)th round be... =median( ); Let the iteration after the r-th round be... =median( ); when Determine if the power curve converges.

[0055] in, and These represent the median of the power sample sets after removing invalid data in the r-th and (r-1)-th wind speed sub-boxes, respectively.

[0056] In summary, the hierarchical denoising calibration method for wind turbine power curves proposed in this invention overcomes the limitations of traditional annual power curve calibration methods that ignore dynamic changes in environmental parameters and operational state transitions, as well as the shortcomings of purely data-driven machine learning methods that lack physical interpretability and robustness, by deeply integrating the physical mechanisms of wind turbine operation with multi-level statistical diagnostic logic. Through this systematic hierarchical denoising and calibration process, the physical consistency and state identification accuracy of the data are significantly improved, making it suitable for applications such as wind turbine performance evaluation, wind power prediction, wind farm post-evaluation, and operation and maintenance optimization.

[0057] Example 2 Another aspect of this invention provides a calibration system for layered noise reduction of wind turbine power curves; please refer to [link / reference]. Figure 2 The diagram shows a schematic of the calibration system for layered noise reduction of wind turbine power curves in Embodiment 2 of the present invention. The calibration system for layered noise reduction of wind turbine power curves includes: The sequence segmentation module 11 is used to acquire the historical operating data sequence of each wind turbine in the wind farm, and to segment the historical operating data sequence into several subsets based on a preset environmental dynamic change cycle. Anomaly identification module 12 is used to construct a progressive anomaly data identification system based on the design operating parameters and physical operating laws of wind turbine units. Based on the identification results of subset data, it removes ineffective power generation operation data points of wind turbine units and obtains effective operation datasets. The curve calibration module 13 is used to iteratively optimize the effective running dataset, obtain the relative rate of change of each round of iterative optimization, until the relative rate of change reaches the preset convergence criterion, and determine the optimal power curve calibration result.

[0058] Example 3 In another aspect, the present invention also proposes a computer-readable storage medium having stored thereon one or more computer programs that, when executed by a processor, implement the above-described calibration method for layered noise reduction of wind turbine power curves.

[0059] Those skilled in the art will understand that the logic or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable storage medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable storage medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0060] More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable storage media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0061] Example 4 Figure 3 This is a structural block diagram of an electronic device provided in Embodiment 4. The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the calibration method for layered noise reduction of the wind turbine power curve in the above embodiments. Figure 3 The electronic device 30 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of the present invention.

[0062] like Figure 3 As shown, the electronic device 30 can be manifested as a general-purpose computing device, such as a server device. The components of the electronic device 30 may include, but are not limited to: at least one processor 31, at least one memory 32, and a bus 33 connecting different system components (including memory 32 and processor 31).

[0063] Bus 33 includes a data bus, an address bus, and a control bus.

[0064] The memory 32 may include volatile memory, such as RAM 321 (random access memory), and / or cache memory 322, and may further include ROM 323 (read-only memory).

[0065] The memory 32 may also include a program tool 325 having a set (at least one) of program modules 324, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0066] The processor 31 executes various functional applications and data processing by running computer programs stored in the memory 32, such as the wind turbine power curve layering noise reduction calibration method described above.

[0067] Electronic device 30 can also communicate with one or more external devices 34 (e.g., keyboard, pointing device, etc.). This communication can be performed via I / O interface 35 (input / output interface). Furthermore, electronic device 30 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public network, such as the Internet) via network adapter 36. Figure 3 As shown, network adapter 36 communicates with other modules of the model-generated electronic device 30 via bus 33. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the model-generated electronic device 30, including but not limited to: microcode, device drivers, redundant processors, disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0068] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0069] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0070] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this patent should be determined by the appended claims.

Claims

1. A method for rating the power curve de-noising of a wind turbine generator, characterized in that, The calibration method for layered noise reduction of the wind turbine power curve includes: The historical operating data sequence of each wind turbine in the wind farm is obtained, and the historical operating data sequence is divided into several subsets based on a preset environmental dynamic change cycle; Based on the design and operation parameters and physical operation laws of wind turbines, a progressive abnormal data identification system is constructed. Based on the identification results of subset data, ineffective power generation operation data points of wind turbines are eliminated to obtain effective operation datasets. The effective running dataset is iteratively optimized to obtain the relative rate of change of each round of iterative optimization until the relative rate of change reaches the preset convergence criterion, and the optimal power curve calibration result is determined.

2. The rating method for wind turbine power curve de-noising in layers according to claim 1, characterized in that, In the step of acquiring the historical operating data sequence of each wind turbine in the wind farm and dividing the historical operating data sequence into several subsets based on a preset environmental dynamic change cycle: Obtain historical operating data sequences of each wind turbine unit in the wind farm at interval sampling periods, and construct a numerical matrix using the historical operating data sequences. The historical operating data sequences include at least the average wind speed and the active power generated. wherein, is a numerical matrix, denotes a time series, denotes the number of wind turbines, represents the wind speed, represents the power, denotes the wind speed value of the th wind turbine in the th time period; denotes the power value of the th wind turbine in the th time period, = 1,..., ; = 1,..., .

3. The calibration method for layered noise reduction of wind turbine power curves according to claim 2, characterized in that, After constructing the numerical matrix, the numerical matrix is used to divide the historical operation data sequence into several subsets based on a preset environment dynamic change period: , , , wherein the number of subsets is , and ≥ ≥1.

4. The rating method for wind turbine power curve de-noising in layers according to claim 1, wherein, The step of constructing a progressive anomaly data identification system, which involves eliminating ineffective power generation operation data points of wind turbine units based on the identification results of subset data, includes: Based on the design and operation parameters of the wind turbine, the cut-in wind speed threshold and the rated power of the wind turbine are obtained. Data points with wind speeds lower than the cut-in wind speed threshold and data points with wind speeds not lower than the cut-in wind speed threshold and active power generation less than or equal to zero are removed to obtain the initial screening dataset. The power generation capacity between full power generation and shutdown is divided into several power boxes with equal spacing. Based on the average value and variance of each power box, the relative stability index of the power box is calculated, and the wind speed fluctuation index is calculated using the wind speed difference in each power box. The power stability and wind speed fluctuation of the power curtailment characteristics are judged by using the relative stability index and the wind speed fluctuation index respectively. Once the power bins achieve the desired power stability and wind speed fluctuation, the linear correlation between wind speed and generated power in the power bins is obtained. When the linear correlation reaches the desired level, the current power bin is determined to be curtailed data and is removed to obtain a second-screen dataset.

5. The rating method for wind turbine power curve de-noising by stratification according to claim 4, characterized in that, After obtaining the second-screen dataset, the wind speeds within the range of wind turbine cut-in wind speed and wind turbine cut-out wind speed are divided into bins, and several wind speed bins with equal widths are obtained. The power average value in each wind speed bin The power deviation of each data point in the first :​​ in, For the first No. 1 fan, No. Within the wind speed distribution box, the first The active power value of each data point. For the first No. 1 fan, the The number of data points within each wind speed sub-box For the first Power within each wind speed compartment The arithmetic mean of the power deviation At that time, the wind speed distribution box showed a severe dispersion point and was therefore rejected.

6. The calibration method for layered noise reduction of wind turbine power curves according to claim 5, characterized in that, After the power deviation anomaly detection of each wind speed sub-box is completed, the corresponding power deviation sequence is extracted and the median deviation is calculated. Based on the median deviation, calculate the absolute value sequence and median absolute deviation between each deviation value and the median deviation. Data points with an absolute value sequence greater than a preset multiple of the median absolute deviation are identified as outliers and removed to obtain a valid running dataset.

7. The calibration method for layered noise reduction of wind turbine power curves according to claim 6, characterized in that, In the step of iteratively optimizing the effective running dataset, obtaining the relative rate of change for each round of iterative optimization, until the relative rate of change reaches a preset convergence criterion, and determining the optimal power curve calibration result: Using the median absolute deviation of a preset multiple as a threshold, the median absolute deviation is tightened round by round by decreasing the multiple until the relative change rate of the median power in each wind speed compartment in the current round reaches the preset convergence standard. The current power curve is then determined to be converged, and the power curve of this round is determined to be the optimal power curve calibration result.

8. A calibration system for layered noise reduction of wind turbine power curves, characterized in that, The wind turbine power curve layered noise reduction calibration system is used to implement the wind turbine power curve layered noise reduction calibration method according to any one of claims 1-7, the system comprising: The sequence segmentation module is used to acquire the historical operating data sequence of each wind turbine in the wind farm, and to segment the historical operating data sequence into several subsets based on a preset environmental dynamic change cycle; The anomaly identification module is used to build a progressive anomaly data identification system based on the design operating parameters and physical operating laws of wind turbines. Based on the identification results of subset data, it removes ineffective power generation operation data points of wind turbines and obtains effective operation datasets. The curve calibration module is used to iteratively optimize the effective running dataset, obtain the relative rate of change of each round of iterative optimization, until the relative rate of change reaches the preset convergence criterion, and determine the optimal power curve calibration result.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the calibration method for layered noise reduction of wind turbine power curves as described in any one of claims 1-7.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the calibration method for layered noise reduction of the power curve of the wind turbine as described in any one of claims 1-7.

Images