Wind turbine generator running state identification method, device, equipment, medium and product

Abstract

The invention relates to the technical field of data processing, and discloses a wind turbine generator running state identification method, device and equipment, a medium and a product, and the method comprises the steps: obtaining initial wind turbine generator data; performing outlier abnormal data point identification on the initial wind turbine generator data by adopting a density clustering algorithm to obtain first wind turbine generator data after outlier abnormal data points are removed; dividing the first wind turbine generator data based on a two-dimensional nonparametric kernel density estimation method to obtain first normal wind turbine generator data and abnormal wind turbine generator data; re-cleaning the abnormal wind turbine generator data by adopting a quartile method to obtain re-cleaned second normal wind turbine generator data; reconstructing the first normal wind turbine generator data and the second normal wind turbine generator data by adopting an interpolation method to obtain target wind turbine generator data; and the operation state of the wind turbine generator is determined based on the target wind turbine generator data, so that the accuracy of operation state identification of the wind turbine generator can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a method, device, equipment, medium and product for identifying the operating status of a wind turbine generator set. Background Art

[0002] During operation, wind turbines often generate data records that are inconsistent with normal operating characteristics. These operating data include abnormal changes in various wind turbine performance parameters or data points that exceed the expected range, etc. However, they may also include some invalid abnormal data caused by interference factors such as environmental factors. These abnormal data will affect subsequent data processing.

[0003] The existing abnormal data cleaning and reconstruction methods for wind turbines mainly identify and remove abnormal data based on the statistical distribution characteristics of the data. This method usually sets a threshold manually and then performs statistical analysis based on the set threshold. However, this method often requires human experience to judge the set threshold, which is prone to problems such as normal data being deleted by mistake or abnormal data being omitted, thereby reducing the integrity and accuracy of the wind turbine operating data and resulting in poor accuracy in identifying the operating status of the wind turbine. Summary of the invention

[0004] In view of this, the present invention provides a method, device, equipment, medium and product for identifying the operating status of a wind turbine generator set to solve the problem of poor accuracy in identifying the operating status of a wind turbine generator set in the prior art.

[0005] In a first aspect, the present invention provides a method for identifying the operating status of a wind turbine generator set, the method comprising:

[0006] Obtain initial wind turbine data;

[0007] A density clustering algorithm is used to identify outlier data points of the initial wind turbine generator set data, and the first wind turbine generator set data after removing the outlier data points is obtained;

[0008] The data of the first wind turbine generator set is divided based on a two-dimensional nonparametric kernel density estimation method to obtain the data of the first normal wind turbine generator set and the data of the abnormal wind turbine generator set;

[0009] The quartile method is used to re-clean the abnormal wind turbine data to obtain the second normal wind turbine data after re-cleaning;

[0010] Reconstructing the first normal wind turbine set data and the second normal wind turbine set data by using an interpolation method to obtain the target wind turbine set data;

[0011] Based on the target wind turbine data, the operating status of the wind turbine is determined.

[0012] In this embodiment, the physical meaning represented by the wind turbine data itself is taken into consideration, and the data category in the transition area can be effectively identified, thereby avoiding problems such as accidental deletion of normal data and omission of abnormal data, and ensuring the validity and accuracy of the wind turbine data, which helps to improve the accuracy of subsequent data analysis, thereby improving the accuracy of identifying the operating status of the wind turbine.

[0013] In an optional implementation, a density clustering algorithm is used to identify outlier data points on the initial wind turbine generator set data, including:

[0014] Step a, determining the radius of a given area and the minimum number of data objects in the given area;

[0015] Step b, determining a first data object, where the first data is unclustered data in the initial wind turbine generator set data;

[0016] Step c, detecting a first number of all data objects within a given area radius of the first data object;

[0017] Step d, when the first number is not less than the minimum number, establishing a new cluster for all data objects within a given area radius of the first data object;

[0018] Step e, determining a second data object in the cluster, the second data object being a data object in the cluster other than the first data object;

[0019] Step f, detecting a second number of all data objects within a given area radius of the second data object;

[0020] Step g, when the second number is not less than the minimum number, adding all data objects within a given area radius of the second data object to the cluster;

[0021] Step h, repeating the above steps e to g until no new data objects are added to the cluster;

[0022] Step i, repeating the above steps b to h until all data in the initial wind turbine generator set data are detected;

[0023] In step j, the data objects that are not added to the cluster are regarded as outlier data points.

[0024] In this embodiment, a density clustering algorithm is used to identify and eliminate outlier data points. According to the density distribution of the data points, the data points that do not conform to the normal distribution can be marked as outliers and eliminated. This can retain the original distribution characteristics and key characteristics of the data points, providing a reliable basis for ultimately acquiring the target wind turbine data and identifying the operating status of the wind turbine.

[0025] In an optional implementation, the initial wind turbine data includes wind speed, power and pitch angle at the hub of the wind turbine, and the first wind turbine data is divided based on a two-dimensional non-parametric kernel density estimation method, including:

[0026] The joint probability density function of power and pitch angle is calculated by two-dimensional nonparametric kernel density estimation method, and the pitch angle boundary is determined by grid division method.

[0027] The first wind turbine generator set data is divided based on the pitch angle dividing line.

[0028] In this embodiment, the data of the first wind turbine generator set is divided based on the two-dimensional non-parametric kernel density estimation method, which can simultaneously consider the relationship between multiple physical variables such as power and pitch angle, and can comprehensively consider the physical meaning represented by the data itself, so as to more accurately divide the abnormal data, providing strong support for further operation status identification and analysis.

[0029] In an optional implementation, determining the operating state of the wind turbine generator set based on the target wind turbine generator set data includes:

[0030] According to the target wind turbine data, determine the key influencing factors of the wind turbine under all operating conditions;

[0031] Extract key impact factor feature vectors of key impact factors;

[0032] Inputting the characteristic vector of the key influencing factor into the pre-constructed normal behavior model of the wind turbine generator set to obtain the model output power; wherein the normal behavior model of the wind turbine generator set is constructed by taking the characteristic vector of the key influencing factor under normal operation as input and the wind turbine generator set power as output;

[0033] The operating status of the wind turbine is determined according to the model output power.

[0034] In this embodiment, through the pre-constructed normal behavior model of the wind turbine, the power output of the wind turbine under a given key influencing factor characteristic vector can be predicted with high accuracy, the impact of the key influencing factors on the performance of the wind turbine can be effectively captured, and reliable power output prediction can be provided, which provides important support and decision-making basis for determining the operating status of the wind turbine.

[0035] In an optional implementation, determining the key influencing factors of the wind turbine generator system under all operating conditions includes:

[0036] Adopting the correlation analysis method, based on the target wind turbine data, the correlation coefficient between different preset parameters and power is calculated;

[0037] According to the correlation coefficient, the potential key influencing factors of wind turbine power are extracted, and a multi-dimensional first potential key influencing factor set is constructed;

[0038] According to the last-place elimination method of correlation coefficient, irrelevant factors in the key influencing factor set are eliminated to form the second potential key influencing factor set;

[0039] Extracting features of potential key influencing factors of different dimensions in the second potential key influencing factor set to obtain feature vectors of potential key influencing factors of different dimensions;

[0040] The characteristic vectors of potential key influencing factors of different dimensions are input into a pre-constructed wind turbine power regression model, and based on the regression error evaluation index, the key influencing factors of the wind turbine in all operating conditions are determined; wherein, the wind turbine power regression model is constructed by taking the characteristic vectors of potential key influencing factors of different dimensions under normal operating conditions as input and the wind turbine power as output.

[0041] In this embodiment, based on the evaluation results of the wind turbine power regression model, the contribution of different orders of potential key influencing factors to the model performance can be analyzed, and then the key influencing factors can be determined by combining the regression error evaluation index with the feature importance analysis. By determining the key influencing factors under all operating conditions, the performance of the wind turbine under different working conditions can be fully understood and grasped, and the accuracy of the wind turbine operating state identification can be improved.

[0042] In an optional implementation, determining the operating state of the wind turbine generator set according to the model output power includes:

[0043] Get the rated wind speed of the wind turbine;

[0044] When the current wind speed of the wind turbine generator set is less than or equal to the rated wind speed, a first preset formula is used to determine the power residual rate index, wherein the first preset formula is a functional relationship between the actual power of the wind turbine generator set and the model output power;

[0045] When the current wind speed of the wind turbine generator set is greater than the rated wind speed, a second preset formula is used to determine the power residual rate index, wherein the second preset formula is a functional relationship between the actual power of the wind turbine generator set and the model output power;

[0046] Based on the power residual rate index, the operating status of the wind turbine is determined.

[0047] In this embodiment, determining the operating status of the wind turbine generator set based on the power residual rate index can effectively evaluate the operating status of the wind turbine generator set from the perspective of the entire machine, thereby improving the accuracy of the evaluation.

[0048] In a second aspect, the present invention provides a device for identifying the operating status of a wind turbine generator set, the device comprising:

[0049] An acquisition module, used for acquiring initial wind turbine data;

[0050] A clustering module, used to identify outlier and abnormal data points of the initial wind turbine generator set data by using a density clustering algorithm, and obtain the first wind turbine generator set data after removing the outlier and abnormal data points;

[0051] A partitioning module, used for partitioning the first wind turbine group data based on a two-dimensional non-parametric kernel density estimation method to obtain first normal wind turbine group data and abnormal wind turbine group data;

[0052] A re-cleaning module, used to re-clean the abnormal wind turbine data by using the quartile method to obtain the second normal wind turbine data after re-cleaning;

[0053] An interpolation module, used to reconstruct the first normal wind turbine group data and the second normal wind turbine group data by using an interpolation method to obtain target wind turbine group data;

[0054] The identification module is used to determine the operating status of the wind turbine generator set based on the target wind turbine generator set data.

[0055] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the wind turbine operating status identification method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0056] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the method for identifying the operating status of a wind turbine set according to the first aspect or any corresponding embodiment thereof.

[0057] In a fifth aspect, the present invention provides a computer program product, comprising computer instructions for causing a computer to execute the method for identifying the operating status of a wind turbine generator set according to the first aspect or any corresponding embodiment thereof.

[0058] It should be noted that, since the wind turbine operating state identification device, computer equipment and computer-readable storage medium provided by the present invention correspond to the above-mentioned wind turbine operating state identification method, therefore, for the beneficial effects of the wind turbine operating state identification device, computer equipment and computer-readable storage medium, please refer to the description of the corresponding beneficial effects of the wind turbine operating state identification method above, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] Figure 1 is a schematic flow chart of a method for identifying the operating status of a wind turbine generator system according to an embodiment of the present invention;

[0061] Figure 2 is a structural block diagram of a wind turbine operating status identification device according to an embodiment of the present invention;

[0062] Figure 3 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0064] The existing wind turbine abnormal data cleaning and reconstruction methods mainly identify and remove abnormal data based on the statistical distribution characteristics of the data. However, this method does not fully consider the physical meaning represented by the data itself, so it is impossible to effectively identify the data category in the transition area between normal data and abnormal data. Therefore, it is urgent to study the wind turbine abnormal data cleaning algorithm based on physical mechanism and statistical characteristics, and to study the wind turbine abnormal data reconstruction and missing data interpolation algorithm.

[0065] Furthermore, at present, the operating status identification of key components of wind turbines is mainly based on SCADA (Supervisory Control and Data Acquisition) data. There are relatively few studies on the identification of the operating status of the entire wind turbine. In addition, the SCADA data of wind turbines has many parameters and the nonlinear relationship between the parameters is complex. Therefore, it is urgent to deeply explore the key influencing parameters of the operating status of wind turbines and study the refined identification method of the entire wind turbine. The refined identification of the operating status of wind turbines can provide a reliable data basis for the evaluation of the operation and maintenance efficiency of wind farms.

[0066] In view of this, according to an embodiment of the present invention, an embodiment of a method for identifying the operating status of a wind turbine is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0067] In this embodiment, a method for identifying the operating status of a wind turbine generator set is provided, which can be executed by a server, a terminal, a mobile terminal and other devices. Figure 1 is a flow chart of a method for identifying the operating status of a wind turbine according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0068] Step S101, obtaining initial wind turbine data, wherein the wind turbine data may include data such as wind speed, power, pitch angle, etc. at the hub height of the wind turbine. Further, the collected data may be normalized and preprocessed to form a sample data set, and the sample data set is used as the initial wind turbine data.

[0069] Step S102, using a density clustering algorithm to identify outlier abnormal data points on the initial wind turbine data, and obtaining the first wind turbine data after removing the outlier abnormal data points. The density clustering algorithm can use DBSCAN (Density-Based Spatial Clustering of Applications with Noise, density-based spatial clustering algorithm) to identify outliers. DBSCAN is a relatively representative density-based clustering algorithm, which can define a cluster as the largest set of density-connected points, can divide areas with sufficiently high density into clusters, and can find clusters of any shape in a noisy spatial database. Using the DBSCAN density clustering algorithm to identify outlier abnormal data points on the initial wind turbine data can effectively identify and remove abnormal points in the initial wind turbine data.

[0070] Step S103 : dividing the first wind turbine group data based on a two-dimensional non-parametric kernel density estimation method to obtain first normal wind turbine group data and abnormal wind turbine group data.

[0071] Based on the control principle of wind turbines, a two-dimensional nonparametric kernel density estimation method can be used to calculate the joint probability density function of power and pitch angle, and the position of the pitch angle dividing line can be determined by the grid division method. Then, based on the pitch angle dividing line, the data of the first wind turbine can be divided, so that normal wind turbine data and abnormal wind turbine data can be obtained.

[0072] Step S104, using the quartile method to re-clean the abnormal wind turbine data to obtain re-cleaned second normal wind turbine data.

[0073] For abnormal data missed after cleaning, the quartile method can be used to further clean the data.

[0074] The core of the quartile method is to calculate the quartiles and interquartile range of ordered data samples. Quartiles are the values ​​of the three split points where an ordered data sample is evenly divided into four parts, namely quartiles, which are recorded as Q1, Q2, and Q3 respectively. Suppose an ascending sample X = {x1, x2, ... xn}, the quartile calculation method is as follows:

[0075] The second quartile, i.e. the median Q2 is:

[0076]

[0077] The first and third quartiles are Q1 and Q3. When n = 2k (k = 0, 1, 2, ...), the data sample is divided into two parts X1 and X2 from Q2. The medians of X1 and X2 are calculated according to the above formula, which are Q1 and Q3 (Q1 <Q3)。

[0078] When n=4k+1, we have:

[0079]

[0080] When n=4k+3, we have:

[0081]

[0082] The interquartile range is calculated by quartiles, that is, I QR For: I QR =Q3-Q1.

[0083] According to the interquartile range, the inner limit of the outlier in the data sample X can be determined as:

[0084] [F l ,F u ]=[Q1-1.5I QR ,Q3+1.5I QR ];

[0085] In the interval [F l ,F u ] are regarded as outliers, thus achieving the purpose of data cleaning.

[0086] Step S105 : reconstructing the first normal wind turbine set data and the second normal wind turbine set data by using an interpolation method to obtain target wind turbine set data.

[0087] Specifically, the piecewise cubic Hermite interpolation method can be used to reconstruct the missing data, thereby forming a complete normal sample data set. Hermite, or Hermite interpolation, is an interpolation method that uses Hermite polynomials.

[0088] Among them, according to the Hermite interpolation theory and its application requirements, the typical two-point cubic Hermite interpolation method is selected, that is, the known mutually different nodes x k and x k+1 and its corresponding function value g i and the first-order derivative value g i '(i=k,k+1), then there exists a unique Hermite interpolation polynomial H3(x) at node x k and x k+1 It satisfies: H3(x)=g(x), H3'(x)=g'(x). Using the basis function method, we can get:

[0089]

[0090] Among them, g i '=(g i -g m ) / (x i -x m )i=k, m=k-1 or i=k+1, m=k+2.

[0091] Step S106: determining the operating status of the wind turbine generator set based on the target wind turbine generator set data.

[0092] In this embodiment, a density clustering algorithm is used to identify outliers, and outlier abnormal data points are eliminated. Then, based on the control principle of the wind turbine, a two-dimensional non-parametric kernel density estimation method is used to calculate the joint probability density function of power and pitch angle, and data division is performed. For the abnormal data missed after cleaning, the quartile method is continued to be used for data re-cleaning, and finally the interpolation method is used to reconstruct the missing data to form a complete normal sample data set, that is, the target wind turbine data. In this embodiment, the physical meaning represented by the wind turbine data itself is taken into account, and the data category of the transition area can be effectively identified to avoid problems such as the accidental deletion of normal data and the omission of abnormal data, and to ensure the validity and accuracy of the wind turbine data, which helps to improve the accuracy of subsequent data analysis, thereby improving the accuracy of the wind turbine operating status identification.

[0093] In some optional implementations, the above step S102, i.e., using a density clustering algorithm to identify outlier data points on the initial wind turbine generator set data, includes:

[0094] Step a, determine the radius of a given region and the minimum number of data objects in the given region. The DBSCAN clustering algorithm can control the generation of clusters by two parameters, Eps and MinPts. Eps is the size of the radius of the given region, and MinPts is the minimum number of data objects contained in the neighborhood of the specified point.

[0095] Step b: determining a first data object, where the first data is the unclustered data in the initial wind turbine generator set data, that is, any unclustered first data object Q can be found in the sample data set.

[0096] Step c: detecting a first number of all data objects within a given area radius of the first data object Q.

[0097] Step d: when the first number is not less than the minimum number, a new cluster is created for all data objects within a given area radius of the first data object.

[0098] That is, check the neighborhood of the given area radius Eps. If the data objects contained in the neighborhood are not less than the neighborhood density threshold Minpts, a new cluster C is established and all other samples in the neighborhood are added to cluster C.

[0099] Step e: determining a second data object in the cluster, where the second data object is a data object in the cluster other than the first data object.

[0100] Step f: detecting a second number of all data objects within a given area radius of the second data object.

[0101] Step g: when the second number is not less than the minimum number, add all data objects within the given area radius of the second data object into the cluster.

[0102] That is, the unprocessed samples in cluster C are checked again to see if their Eps neighborhood is not less than Minpts. If so, the data objects in their neighborhood that are not included in any cluster set are also added to cluster C.

[0103] Step h: repeat the above steps e to g until no new data objects are added to the cluster, that is, until no new objects can be added to cluster C.

[0104] Step i: repeat the above steps b to h until all data in the initial wind turbine generator set data are detected, that is, until all samples are processed.

[0105] In step j, the data objects that are not added to the cluster are regarded as outlier data points.

[0106] In this embodiment, a density clustering algorithm is used to identify and eliminate outlier data points. According to the density distribution of the data points, the data points that do not conform to the normal distribution can be marked as outliers and eliminated. This can retain the original distribution characteristics and key characteristics of the data points, providing a reliable basis for ultimately acquiring the target wind turbine data and identifying the operating status of the wind turbine.

[0107] In some optional implementations, the initial wind turbine data includes wind speed, power and pitch angle at the hub of the wind turbine. The above step S103, i.e., dividing the first wind turbine data based on a two-dimensional non-parametric kernel density estimation method, includes:

[0108] The joint probability density function of power and pitch angle is calculated by two-dimensional nonparametric kernel density estimation method, and the pitch angle boundary is determined by grid division method.

[0109] Based on the pitch angle dividing line, the data of the first wind turbine generator set is divided, the power and pitch angle data are divided into normal data and abnormal data, and the normal wind speed and power data are obtained by using the data time calibration method.

[0110] Among them, the two-dimensional probability density nonparametric estimation algorithm includes:

[0111] Assume that the two-dimensional data sample set is (X,Y)={(x i ,y i ),i=1,2,,…N}, then its two-dimensional probability density function can be expressed as the following formula:

[0112]

[0113] in, is the two-dimensional probability density function based on nonparametric kernel density estimation, h x and h y is the window width, N is the total number of data samples, K x (·) and K y (·) is the kernel function. The selection of window width and kernel function is a key factor affecting the accuracy of probability density function estimation. In this embodiment, the window width value is determined by cross-identification method, and Gaussian kernel function is selected as the kernel function of two-dimensional probability density function estimation, as shown in the following formula:

[0114]

[0115] Combining the above two formulas, the two-dimensional probability density function can be rewritten as:

[0116]

[0117] Based on the pitch angle dividing line, the power and pitch angle data are divided into normal data and abnormal data, and the normal wind speed and power data are obtained by using the data time calibration method.

[0118] In this embodiment, the data of the first wind turbine generator set is divided based on the two-dimensional non-parametric kernel density estimation method, which can simultaneously consider the relationship between multiple physical variables such as power and pitch angle, and can comprehensively consider the physical meaning represented by the data itself, so as to more accurately divide the abnormal data, providing strong support for further operation status identification and analysis.

[0119] In some optional implementations, the above step S106, i.e. determining the operating state of the wind turbine generator set based on the target wind turbine generator set data, includes:

[0120] According to the target wind turbine data, determine the key influencing factors of the wind turbine in all operating states, where the operating state can include normal operating state and abnormal operating state. In normal operation, the wind turbine can be stable, that is, the power output is in line with expectations, and the pitch angle and wind speed are usually within the normal range; abnormal operating state may include: abnormal wind speed, abnormal pitch angle, etc. The key influencing factors may include: wind speed, power output, pitch angle, temperature, vibration frequency, equipment parameters, etc.

[0121] Extract key impact factor feature vectors of key impact factors;

[0122] Inputting the characteristic vector of the key influencing factor into the pre-constructed normal behavior model of the wind turbine generator set to obtain the model output power; wherein the normal behavior model of the wind turbine generator set is constructed by taking the characteristic vector of the key influencing factor under normal operation as input and the wind turbine generator set power as output;

[0123] The operating status of the wind turbine is determined according to the model output power.

[0124] In this embodiment, through the pre-constructed normal behavior model of the wind turbine, the power output of the wind turbine under a given key influencing factor characteristic vector can be predicted with high accuracy, the impact of the key influencing factors on the performance of the wind turbine can be effectively captured, and reliable power output prediction can be provided, which provides important support and decision-making basis for determining the operating status of the wind turbine.

[0125] In some optional implementations, determining the key influencing factors of the wind turbine generator system under all operating conditions includes:

[0126] Adopting the correlation analysis method, based on the target wind turbine data, the correlation coefficient between different preset parameters and power is calculated;

[0127] According to the correlation coefficient, the potential key influencing factors of wind turbine power are extracted, and a multi-dimensional first potential key influencing factor set is constructed;

[0128] According to the last-place elimination method of correlation coefficient, irrelevant factors in the key influencing factor set are eliminated to form the second potential key influencing factor set;

[0129] Extracting features of potential key influencing factors of different dimensions in the second potential key influencing factor set to obtain feature vectors of potential key influencing factors of different dimensions;

[0130] The characteristic vectors of potential key influencing factors of different dimensions are input into a pre-constructed wind turbine power regression model, and based on the regression error evaluation index, the key influencing factors of the wind turbine in all operating conditions are determined; wherein, the wind turbine power regression model is constructed by taking the characteristic vectors of potential key influencing factors of different dimensions under normal operating conditions as input and the wind turbine power as output.

[0131] Specifically, a complete normal sample data set after abnormal operation data identification and reconstruction of wind turbines and missing data interpolation is extracted, namely, target wind turbine data; then, the correlation analysis method is used to calculate the correlation coefficients between different parameters and power, and the potential key influencing factors of wind turbine power are extracted to form an n-order potential key influencing factor set; then, according to the last-place elimination method of the correlation coefficient, different orders (n-order, n-1-order, n-2-order...3-order) of potential key influencing factors are formed, and the data dimensionality reduction method is used to extract features of different orders of potential key influencing factors to obtain feature vectors of potential key influencing factors of different dimensions;

[0132] Among them, it is necessary to first take the characteristic vectors of different orders of potential key influencing factors under normal operating conditions as input and the wind turbine power as output, and build a wind turbine power regression model based on machine learning. Then, the characteristic vectors of potential key influencing factors of different dimensions are input into the wind turbine power regression model, and based on the regression error evaluation index, the key influencing factors of the wind turbine in all operating conditions are determined.

[0133] In this embodiment, based on the evaluation results of the wind turbine power regression model, the contribution of different orders of potential key influencing factors to the model performance can be analyzed, and then the key influencing factors can be determined by combining the regression error evaluation index with the feature importance analysis. By determining the key influencing factors under all operating conditions, the performance of the wind turbine under different working conditions can be fully understood and grasped, and the accuracy of the wind turbine operating state identification can be improved.

[0134] In some optional implementations, determining the operating state of the wind turbine generator system according to the model output power includes:

[0135] Get the rated wind speed of the wind turbine.

[0136] When the current wind speed of the wind turbine generator set is less than or equal to the rated wind speed, a first preset formula is used to determine the power residual rate index, wherein the first preset formula is a functional relationship between the actual power of the wind turbine generator set and the model output power.

[0137] When the current wind speed of the wind turbine generator set is greater than the rated wind speed, a second preset formula is used to determine the power residual rate index, wherein the second preset formula is a functional relationship between the actual power of the wind turbine generator set and the model output power.

[0138] Based on the power residual rate index, the operating status of the wind turbine is determined.

[0139] In this embodiment, the operating state of the wind turbine can be determined based on the model output power prediction residual evaluation index. Specifically, the current wind speed V and the rated wind speed V rated As the identification basis of wind turbine operation control stage.

[0140] When V≤V rated When the power residual rate index ΔP' is used to determine the operating state of the wind turbine generator set, the power residual rate index ΔP' is calculated by a first preset formula, and the first preset formula is:

[0141] When V>V rated When the power residual index ΔP is used to determine the operating state of the wind turbine, the power residual rate index ΔP' is calculated by the second preset formula, and the second preset formula is: ΔP = P a -P f .

[0142] Among them, P a is the actual power of the wind turbine, P f , P r All are the output power of wind turbine model.

[0143] In this embodiment, determining the operating status of the wind turbine generator set based on the power residual rate index can effectively evaluate the operating status of the wind turbine generator set from the perspective of the entire machine, thereby improving the accuracy of the evaluation.

[0144] In this embodiment, based on the cleaned target wind turbine data, a method combining statistical analysis and feature extraction is adopted to deeply explore the key parameters that affect the operating status of the wind turbine. A qualitative grading method of key influencing parameters based on the influencing mode is used to reveal the influencing mechanism and coupling relationship of key influencing parameters of different levels on the operating status of the wind turbine. The operating status of the wind turbine is determined by using key influencing factors, which effectively improves the accuracy of determining the operating status of the wind turbine.

[0145] In this embodiment, a wind turbine operating status identification device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0146] This embodiment provides a device for identifying the operating status of a wind turbine group. Figure 2 As shown, including:

[0147] An acquisition module 201 is used to acquire initial wind turbine generator set data;

[0148] The clustering module 202 is used to identify outlier data points on the initial wind turbine generator set data using a density clustering algorithm to obtain first wind turbine generator set data after removing the outlier data points;

[0149] A division module 203 is used to divide the first wind turbine group data based on a two-dimensional non-parametric kernel density estimation method to obtain first normal wind turbine group data and abnormal wind turbine group data;

[0150] A re-cleaning module 204 is used to re-clean the abnormal wind turbine data by using a quartile method to obtain second normal wind turbine data after re-cleaning;

[0151] An interpolation module 205 is used to reconstruct the first normal wind turbine group data and the second normal wind turbine group data by using an interpolation method to obtain target wind turbine group data;

[0152] The identification module 206 is used to identify the operating status of the wind turbine generator set based on the target wind turbine generator set data.

[0153] In some optional implementations, the clustering module 202 includes:

[0154] A first determination unit is used to determine a given area radius and a minimum number of data objects within the given area; and to determine a first data object, where the first data is data that is not clustered in the initial wind turbine group data;

[0155] A first detection unit is used to detect a first number of all data objects within a given area radius of the first data object; when the first number is not less than a minimum number, a new cluster is established for all data objects within the given area radius of the first data object;

[0156] A second determining unit, configured to determine a second data object in the cluster, the second data object being a data object in the cluster other than the first data object;

[0157] A second detection unit is used to detect a second number of all data objects within a given area radius of the second data object; if the second number is not less than a minimum number, all data objects within the given area radius of the second data object are added to the cluster; until no new data objects are added to the cluster; until all data in the initial wind turbine group data are detected;

[0158] The elimination unit is used to treat the data objects that are not added to the cluster as outlier abnormal data points.

[0159] In some optional implementations, the division module 203 includes:

[0160] The partitioning unit is used to calculate the joint probability density function of power and pitch angle by using a two-dimensional non-parametric kernel density estimation method, and determine the pitch angle dividing line by a grid partitioning method; based on the pitch angle dividing line, the data of the first wind turbine generator set is partitioned.

[0161] In some optional implementations, the identification module 206 includes:

[0162] The key influencing factor determination unit is used to calculate the correlation coefficient between different preset parameters and power based on the target wind turbine data by using the correlation analysis method; extract the potential key influencing factors of the wind turbine power according to the correlation coefficient, and construct a multi-dimensional first potential key influencing factor set; eliminate irrelevant factors in the key influencing factor set according to the last-place elimination method of the correlation coefficient to form a second potential key influencing factor set; extract features of the potential key influencing factors of different dimensions in the second potential key influencing factor set to obtain feature vectors of potential key influencing factors of different dimensions; input the feature vectors of potential key influencing factors of different dimensions into a pre-constructed wind turbine power regression model, and determine the key influencing factors of the wind turbine under all operating conditions based on the regression error evaluation index; wherein the wind turbine power regression model is constructed by taking the feature vectors of potential key influencing factors of different dimensions under normal operating conditions as input and the wind turbine power as output.

[0163] The identification unit is used to determine the key influencing factors of the wind turbine under all operating conditions according to the target wind turbine data; extract the key influencing factor feature vector of the key influencing factor; input the key influencing factor feature vector into the pre-constructed normal behavior model of the wind turbine to obtain the model output power; wherein the normal behavior model of the wind turbine is constructed by taking the key influencing factor feature vector under normal operating conditions as input and the wind turbine power as output; and determine the operating state of the wind turbine according to the model output power. It is also used to obtain the rated wind speed of the wind turbine; when the current wind speed of the wind turbine is less than or equal to the rated wind speed, the power residual rate index is determined by using the first preset formula, wherein the first preset formula is the functional relationship between the actual power of the wind turbine and the model output power; when the current wind speed of the wind turbine is greater than the rated wind speed, the power residual rate index is determined by using the second preset formula, wherein the second preset formula is the functional relationship between the actual power of the wind turbine and the model output power; based on the power residual rate index, the operating state of the wind turbine is determined.

[0164] The wind turbine operating status identification device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0165] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0166] The embodiment of the present invention also provides a computer device having the above Figure 2 The wind turbine operating status identification device shown.

[0167] See also Figure 3 , Figure 3 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 3 As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 3A processor 10 is taken as an example.

[0168] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0169] The memory 20 stores instructions executable by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0170] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0171] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0172] The computer device further comprises a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0173] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0174] A part of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the existence of the computer program instruction in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc., and accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium accessible to the computer.

[0175] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for identifying the operating status of a wind turbine generator set, characterized in that: The method comprises: Obtain initial wind turbine data; Using a density clustering algorithm to identify outlier data points on the initial wind turbine generator set data, and obtaining first wind turbine generator set data after removing the outlier data points; Dividing the first wind turbine group data based on a two-dimensional nonparametric kernel density estimation method to obtain first normal wind turbine group data and abnormal wind turbine group data; Re-cleaning the abnormal wind turbine data by using the quartile method to obtain re-cleaned second normal wind turbine data; Reconstructing the first normal wind turbine set data and the second normal wind turbine set data by using an interpolation method to obtain target wind turbine set data; Based on the target wind turbine group data, an operating state of the wind turbine group is determined.

2. The method according to claim 1, characterized in that The method of using a density clustering algorithm to identify outlier and abnormal data points on the initial wind turbine generator set data includes: Step a, determining the radius of a given area and the minimum number of data objects in the given area; Step b, determining a first data object, wherein the first data is unclustered data in the initial wind turbine generator set data; Step c, detecting a first number of all data objects within the given area radius of the first data object; Step d: when the first number is not less than the minimum number, a new cluster is established for all data objects within the given area radius of the first data object; Step e, determining a second data object in the cluster, where the second data object is a data object in the cluster other than the first data object; Step f, detecting a second number of all data objects within the given area radius of the second data object; Step g, when the second number is not less than the minimum number, adding all data objects within the given area radius of the second data object to the cluster; Step h, repeating the above steps e to g until no new data object is added to the cluster; Step i, repeating the above steps b to h until all data in the initial wind turbine generator set data are detected; In step j, the data objects that are not added to the cluster are regarded as outlier data points.

3. The method according to claim 1, characterized in that The initial wind turbine data includes wind speed, power and pitch angle at the hub of the wind turbine. The dividing of the first wind turbine data based on the two-dimensional non-parametric kernel density estimation method includes: A two-dimensional nonparametric kernel density estimation method is used to calculate the joint probability density function of the power and the pitch angle, and a pitch angle dividing line is determined by a grid division method; The first wind turbine generator set data is divided based on the pitch angle dividing line.

4. The method according to claim 3, characterized in that The step of determining the operating state of the wind turbine generator set based on the target wind turbine generator set data includes: According to the target wind turbine data, determine the key influencing factors of the wind turbine under all operating conditions; Extracting a key influencing factor feature vector of the key influencing factor; Input the key influencing factor characteristic vector into a pre-constructed normal behavior model of a wind turbine generator set to obtain the model output power; wherein the normal behavior model of a wind turbine generator set is constructed by taking the key influencing factor characteristic vector under normal operation as input and the wind turbine generator set power as output; The operating state of the wind turbine generator set is determined according to the output power of the model.

5. The method according to claim 4, characterized in that The key influencing factors of the wind turbine generator system under all operating conditions are determined as follows: Using a correlation analysis method, based on the target wind turbine data, the correlation coefficient between different preset parameters and power is calculated; According to the correlation coefficient, extracting potential key influencing factors of wind turbine power, and constructing a multi-dimensional first potential key influencing factor set; According to the last-place elimination method of correlation coefficient, irrelevant factors in the key influencing factor set are eliminated to form a second potential key influencing factor set; Performing feature extraction on the potential key influencing factors of different dimensions in the second potential key influencing factor set to obtain feature vectors of the potential key influencing factors of different dimensions; The characteristic vectors of the potential key influencing factors of different dimensions are input into a pre-constructed wind turbine power regression model, and based on the regression error evaluation index, the key influencing factors of the wind turbine in all operating conditions are determined; wherein the wind turbine power regression model is constructed by taking the characteristic vectors of the potential key influencing factors of different dimensions under normal operating conditions as input and the wind turbine power as output.

6. The method according to claim 4, characterized in that Determining the operating state of the wind turbine generator set according to the model output power includes: Get the rated wind speed of the wind turbine; When the current wind speed of the wind turbine generator set is less than or equal to the rated wind speed, a first preset formula is used to determine the power residual rate index, wherein the first preset formula is a functional relationship between the actual power of the wind turbine generator set and the output power of the model; When the current wind speed of the wind turbine generator set is greater than the rated wind speed, a second preset formula is used to determine the power residual rate index, wherein the second preset formula is a functional relationship between the actual power of the wind turbine generator set and the output power of the model; Based on the power residual rate indicator, the operating state of the wind turbine generator set is determined.

7. A device for identifying the operating status of a wind turbine generator set, characterized in that: The device comprises: An acquisition module, used for acquiring initial wind turbine data; A clustering module, used for identifying outlier and abnormal data points on the initial wind turbine generator set data by using a density clustering algorithm, and obtaining first wind turbine generator set data after removing the outlier and abnormal data points; A division module, used for dividing the first wind turbine group data based on a two-dimensional non-parametric kernel density estimation method to obtain first normal wind turbine group data and abnormal wind turbine group data; A re-cleaning module, used for re-cleaning the abnormal wind turbine data by using the quartile method to obtain the second normal wind turbine data after re-cleaning; An interpolation module, used for reconstructing the first normal wind turbine group data and the second normal wind turbine group data by using an interpolation method to obtain target wind turbine group data; The identification module is used to determine the operating status of the wind turbine generator set based on the target wind turbine generator set data.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for identifying the operating status of a wind turbine set according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the wind turbine operating status identification method according to any one of claims 1-6.

10. A computer program product, characterized in that The method comprises computer instructions, wherein the computer instructions are used to enable a computer to execute the method for identifying the operating status of a wind turbine set according to any one of claims 1 to 6.

Images