CN112682269A - Wind turbine generator state monitoring method based on OC-RKELM - Google Patents
Wind turbine generator state monitoring method based on OC-RKELM Download PDFInfo
- Publication number
- CN112682269A CN112682269A CN202011519151.XA CN202011519151A CN112682269A CN 112682269 A CN112682269 A CN 112682269A CN 202011519151 A CN202011519151 A CN 202011519151A CN 112682269 A CN112682269 A CN 112682269A
- Authority
- CN
- China
- Prior art keywords
- data
- wind turbine
- turbine generator
- rkelm
- health
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/72—Wind turbines with rotation axis in wind direction
Landscapes
- Wind Motors (AREA)
Abstract
A wind turbine state monitoring method based on OC-RKELM comprises the following steps: step 1: collecting operating data and environmental parameter data of a wind turbine generator; step 2: cleaning the collected wind turbine generator data; establishing a health data set of normal operation data of the wind turbine generator; and step 3: based on the health data set of the wind turbine generator, establishing mathematical description aiming at the health data set by utilizing an OC-RKELM model; and 4, step 4: determining a health threshold value based on analysis of historical normal data to judge whether the operating data belongs to a health data set of the wind turbine generator; and 5: and correspondingly processing the real-time operation data of the wind turbine generator, judging whether the real-time operation data is normal operation data or not based on the OC-RKELM model, and performing early warning and shutdown inspection on the wind turbine generator when early warning conditions are met. The method has less human intervention and simple implementation, can effectively detect the abnormity of the running state of the generator of the wind turbine generator, and can provide technical support for the operation and maintenance of the wind power plant.
Description
Technical Field
The invention relates to the technical field of wind power, in particular to a wind turbine generator state monitoring method based on OC-RKELM.
Background
In order to solve the problems of environmental pollution, energy crisis and the like, wind power industry is being vigorously developed in various countries in the world, and the accumulated installed capacity of wind generation sets is continuously increased. By the end of 2019, the installed capacity of global wind turbines reaches 651GW, and hundreds of thousands of wind turbines are running in a grid-connected mode. The design life of a wind turbine generator is usually 20 years, and in the service life cycle of the wind turbine generator, due to the adverse effects of various factors such as environment, load, materials and the like, the components of the wind turbine generator inevitably generate accumulated damage, resistance attenuation and function degradation, and even serious faults. The state monitoring technology can find abnormality before equipment fails, so that technical support is provided for optimizing operation and maintenance strategies and reducing operation and maintenance cost. Therefore, developing a condition monitoring technology of the wind turbine is one of the keys for improving the reliability and the economy of the wind power plant.
The existing state monitoring technology of a data acquisition and monitoring control (SCADA) system is realized only by setting a threshold value for a one-dimensional characteristic parameter, and only extreme abnormal data can be found; the existing wind power data visualization analysis technology needs professional analysts to analyze charts such as power curves of wind turbine generators or supervise wind power plants by experienced personnel, so that the labor cost is high and the real-time performance is poor; the existing state monitoring technology based on neural network modeling has the advantages that the parameters have strong randomness, and the generalization and stability of the model are difficult to guarantee.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a wind turbine state monitoring method based on a single-Class simplified Kernel Extreme Learning Machine (OC-RKELM).
The technical scheme adopted by the invention for solving the technical problems is as follows:
a wind turbine state monitoring method based on OC-RKELM comprises the following steps:
step 1: collecting operating data and environmental parameter data of a wind turbine generator, wherein the operating data of the wind turbine generator comprises a pitch angle, a generator rotating speed and active power, and the environmental parameters comprise a wind speed, a wind direction and an environmental temperature;
step 2: cleaning the collected wind turbine generator data, including processing missing values and abnormal values; clearing standby data, shutdown data and outlier data; establishing a health data set of normal operation data of the wind turbine generator;
and step 3: based on a health data set of the wind turbine generator, establishing mathematical description aiming at the health data set by utilizing an OC-RKELM model, wherein a one-dimensional index output by the OC-RKELM model represents the similarity degree of input real-time operation data and the health data set of the wind turbine generator;
and 4, step 4: determining a health threshold value based on analysis of historical normal data to judge whether the operating data belongs to a health data set of the wind turbine generator;
and 5: and correspondingly processing the real-time operation data of the wind turbine generator, judging whether the real-time operation data is normal operation data or not based on the OC-RKELM model, and performing early warning and shutdown inspection on the wind turbine generator when early warning conditions are met.
Further, the process of step 2 is as follows:
step 21: clearing missing data in the acquired data; clearing data beyond the given range of the parameters;
step 22: in order to obtain pure normal operation data of the wind turbine generator, firstly, data of the generator in standby and shutdown states are removed; secondly, removing outlier data which are seriously deviated from the healthy data cluster through a local abnormal factor outlier detection algorithm;
step 23: and correspondingly normalizing the data after data cleaning in order to eliminate dimension influence.
In the step 2, corresponding normalization is performed on the wind turbine generator data in order to eliminate dimensional influence. Normalization is a linear transformation of the original data, so that the result of the transformation falls in the [0,1] interval, and the conversion function is as follows:
the outlier data is cleaned by adopting a local abnormal factor algorithm, wherein the local abnormal factor algorithm is defined by the following parameters:
k-distance: for a certain object p in the dataset D, defining a distance k-dist (p) between p and the object o e D (p, o), the object o satisfies the following two conditions: at least k objects o 'belong to D \ p }, and D (p, o') is less than or equal to D (p, o); at most k-1 objects o 'belong to D \ p }, and D (p, o') < D (p, o) is satisfied;
k-distance neighborhood: k-distance neighborhood N of object pk(p), defined as all objects in D that are not more than k-distance from p:
Nk(p)＝{q∈D\{p}|d(p,q)≤k-dist(p)} (2)
the reachable distance is: object o to object p reachable distance reach distk(p, o) is defined as:
reach-distk(p,o)＝max{k-dist(o),d(p,o)} (3)
i.e. the larger value in the k-distance of object p and object o;
local accessible density: local reachable density lrd of object pk(p) is defined as:
representing the ratio of objects within the k-distance neighborhood of object p to the reachable distance of object p;
local abnormality factor: the local anomaly factor for object p is defined as:
represents the average of the ratio of the local achievable density of objects within the k-distance neighborhood of p to the local achievable density of p. The local anomaly factor can describe the outlier degree of the object p, and when the local anomaly factor is less than or equal to 1, the object p is positioned in the data dense area; when the local anomaly factor is greater than 1, it indicates that p may be far away from the data-dense region, with the larger the value the higher the degree of outlier of p.
Still further, the process of step 3 is: selecting a support vector: in order to select representative wind turbine generator operation data, a system sampling method is adopted for sampling, and the operation is as follows: parameters such as power, time and the like are segmented, corresponding sub-sample sets are divided, and samples are extracted from each sub-sample set to serve as support vectors.
In the step 3, an OC-RKELM state monitoring model needs to be constructed. Extracting a number of health samples, X ═ X1,x2,…,xN]TThe target output vector is R ═ 1,1, …,1]T. The output is in the form of:
in the formula (I), the compound is shown in the specification,
for the model output parameter, betajFor the output weights, K (-) is the kernel function,
the method is a set of support vectors, selects a radial basis function as a kernel mapping function, and has the following form:
K(x,xj)＝exp(-||x-xj||2/σ) (7)
where σ is a hyperparameter, xjIs a support vector;
the single class problem solved by the OC-RKELM model is described as:
in the formula, KN×L＝K(X,XL) To simplify the kernel matrix, the training errors and output weights are balanced by introducing L2 regularization, and equation (8) can be rewritten as:
where ξ is the training error and λ is the regularization coefficient. And introducing a Lagrangian function to solve the formula (9):
wherein A is a Lagrange multiplier vector. According to the Karush-Kuhn-Tucker (KKT) condition, the solution of the output weight is obtained as follows:
in the formula, I is an identity matrix.
The distance of a sample from the normal class can be defined as:
the greater the distance of the sample from the normal class, the more the sample deviates from the normal class.
Further, the process of step 4 is: determining a health threshold by a kernel density estimation method, comprising: calculating a one-dimensional output index corresponding to normal operation data based on an OC-RKELM model, estimating a density function of the one-dimensional output index by using a nuclear density estimation method, and determining a health threshold value according to a certain confidence coefficient.
In step 4, a health threshold is determined by using a kernel density estimation method, and a decision function is defined as:
where Sign (. cndot.) is a Sign function and thres is a threshold. When d (x) is less than or equal to the threshold value, the monitoring data is considered to belong to normal data; when the value is larger than the threshold value, the monitoring data is considered as abnormal data.
The process of the step 5 is as follows:
step 51: inputting real-time operation data into the constructed OC-RKELM model, and judging the real-time operation data as abnormal operation data when the output index exceeds a health threshold value;
step 52: and carrying out statistical analysis on the output result of the OC-RKELM model, setting a proper early warning condition, and prompting the abnormality of the wind turbine generator when the early warning condition is met.
In the invention, the OC-RKELM method is an improved single classification method on the basis of a single classification Extreme Learning Machine (OC-ELM), and not only is the generalization capability of the model improved, but also the lower computation time complexity is kept by combining a kernel method and randomly extracting a support vector. On the basis, the method combines the characteristics of the wind turbine generator data, adopts a system sampling and hierarchical sampling method to extract a more representative support vector, and further improves the generalization capability of the model.
The invention has the following beneficial effects: according to the method, the health state of the wind turbine generator can be automatically monitored in real time only by analyzing the health data set of the wind turbine generator, and a nuclear method is introduced to effectively improve the generalization and stability of the state monitoring model.
Drawings
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a schematic diagram of a shutdown data, standby data cleaning, wherein (a) indicates before cleaning and (b) indicates after cleaning;
FIG. 3 is a schematic diagram of outlier data cleaning, wherein (a) indicates before cleaning and (b) indicates after cleaning;
fig. 4 is a schematic diagram of the visualization analysis of the generator fault, in which (a) shows a distribution diagram after the data is reduced to the 8 days before the fault, (b) shows a distribution diagram after the data is reduced to the 7 days before the fault, (c) shows a distribution diagram after the data is reduced to the 6 days before the fault, and (d) shows a distribution diagram after the data is reduced to the day of the fault.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Referring to fig. 1 to 4, a wind turbine generator state monitoring method based on OC-RKELM includes the following steps:
step 1: collecting wind turbine generator operating data and environmental parameter data, wherein the wind turbine generator operating data comprises a pitch angle, a generator rotating speed, active power and the like, and the environmental parameters comprise a wind speed, a wind direction, an environmental temperature and the like;
step 2: cleaning the collected wind turbine generator data, including processing missing values and abnormal values; clearing standby data, shutdown data, outlier data and the like, wherein the main purpose of the clearing is to establish a healthy data set of normal operation data of the wind turbine;
and step 3: based on a health data set of the wind turbine generator, establishing mathematical description aiming at the health data set by utilizing an OC-RKELM model, wherein a one-dimensional index output by the OC-RKELM model represents the similarity degree of input real-time operation data and the health data set of the wind turbine generator;
and 4, step 4: determining a health threshold value based on analysis of historical normal data to judge whether the operating data belongs to a health data set of the wind turbine generator;
and 5: and correspondingly processing the real-time operation data of the wind turbine generator, and judging whether the real-time operation data is normal operation data or not based on the OC-RKELM model. And when the early warning condition is met, the wind turbine generator is subjected to early warning and shutdown inspection.
In the step 2, corresponding normalization is performed on the wind turbine generator data in order to eliminate dimensional influence. Normalization is a linear transformation of the original data, so that the result of the transformation falls in the [0,1] interval, and the conversion function is as follows:
the outlier data is cleaned by adopting a local abnormal factor algorithm, wherein the local abnormal factor algorithm is defined by the following parameters:
k-distance: for a certain object p in the dataset D, defining a distance k-dist (p) between p and the object o e D (p, o), the object o satisfies the following two conditions: at least k objects o 'belong to D \ p }, and D (p, o') is less than or equal to D (p, o); at most k-1 objects o 'belong to D \ p }, and D (p, o') < D (p, o) is satisfied;
k-distance neighborhood: k-distance neighborhood N of object pk(p), defined as all objects in D that are not more than k-distance from p:
Nk(p)＝{q∈D\{p}|d(p,q)≤k-dist(p)} (2)
the reachable distance is: object o to object p reachable distance reach distk(p, o) is defined as:
reach-distk(p,o)＝max{k-dist(o),d(p,o)} (3)
i.e. the larger value in the k-distance of object p and object o;
local accessible density: local reachable density lrd of object pk(p) is defined as:
representing the ratio of objects within the k-distance neighborhood of object p to the reachable distance of object p;
local abnormality factor: the local anomaly factor for object p is defined as:
represents the average of the ratio of the local achievable density of objects within the k-distance neighborhood of p to the local achievable density of p. The local anomaly factor can describe the outlier degree of the object p, and when the local anomaly factor is less than or equal to 1, the object p is positioned in the data dense area; when the local anomaly factor is greater than 1, it indicates that p may be far away from the data-dense region, with the larger the value the higher the degree of outlier of p.
In the step 3, an OC-RKELM state monitoring model needs to be constructed. Extracting a number of health samples, X ═ X1,x2,…,xN]TThe target output vector is R ═ 1,1, …,1]T. The output is in the form of:
in the formula (I), the compound is shown in the specification,
for the model output parameter, betajFor the output weights, K (-) is the kernel function,
is a set of support vectors. The radial basis function is selected as a kernel mapping function, and the form of the kernel mapping function is as follows:
K(x,xj)＝exp(-||x-xj||2/σ) (7)
where σ is a hyperparameter, xjIs a support vector.
The single classification problem solved by the OC-RKELM model can be described as:
in the formula, KN×L＝K(X,XL) To simplify the kernel matrix. By balancing the training errors and output weights by introducing L2 regularization, equation (8) can be rewritten as:
where ξ is the training error and λ is the regularization coefficient. And introducing a Lagrangian function to solve the formula (9):
wherein A is a Lagrange multiplier vector. According to the Karush-Kuhn-Tucker (KKT) condition, the solution of the output weight is obtained as follows:
in the formula, I is an identity matrix.
The distance of a sample from the normal class can be defined as:
the greater the distance of the sample from the normal class, the more the sample deviates from the normal class.
In step 4, a health threshold is determined by using a kernel density estimation method, and a decision function is defined as:
where Sign (. cndot.) is a Sign function and thres is a threshold. When d (x) is less than or equal to the threshold value, the monitoring data is considered to belong to normal data; when the value is larger than the threshold value, the monitoring data is considered as abnormal data.
In order to further improve the generalization capability of the model, the characteristics of the wind turbine data are combined, and the wind turbine data with more representativeness is extracted as the support vector by adopting a system sampling and hierarchical sampling method. According to the idea of system sampling, wind turbine generator data are arranged according to a time sequence and segmented according to a certain interval: determining a segmented interval k, segmenting data, and when N/N (N is the number of the sub data sets) is an integer, taking k as N/N to obtain N sub data sets; according to the idea of hierarchical sampling, the data in the segmented sub data set is layered: dividing the data into l levels according to a certain power interval, and randomly extracting samples from the l levels. After sampling, L-n × L samples are obtained, and the samples have good representativeness and are used as support vectors in the OC-RKELM model.
In this embodiment, the condition monitoring of the generator of the wind turbine is explained, and the condition monitoring method of the wind turbine based on OC-RKELM is verified by using SCADA data, and the process is as follows:
step 1: collecting data such as operating data, environmental parameters and the like of the wind turbine generator; extracting characteristic parameters which are closely related to the health state of the generator, such as: wind speed, generator speed, winding temperature, etc.
Step 2: processing the missing value and the abnormal value; as shown in fig. 2, the standby and shutdown data are cleared; as shown in fig. 3, the outlier data is cleaned; and establishing a health data set of the normal operation data of the generator of the wind turbine generator.
And step 3: and establishing mathematical description aiming at the generator health data set by utilizing an OC-RKELM model based on the health data set of the generator of the wind turbine generator.
And 4, step 4: and determining a health threshold value based on the analysis of the historical normal data to judge whether the operation data belongs to the health data set of the wind turbine generator.
And 5: and correspondingly processing the real-time operation data of the wind turbine generator, and judging whether the real-time operation data is normal operation data or not based on the OC-RKELM model. The method successfully carries out early warning on the faults of the embodiment and carries out visual analysis on abnormal data as shown in figure 4.
The embodiments described in this specification are merely illustrative of implementations of the inventive concepts, which are intended for purposes of illustration only. The scope of the present invention should not be construed as being limited to the particular forms set forth in the examples, but rather as being defined by the claims and the equivalents thereof which can occur to those skilled in the art upon consideration of the present inventive concept.
Claims (8)
1. A wind turbine generator state monitoring method based on OC-RKELM is characterized by comprising the following steps:
step 1: collecting operating data and environmental parameter data of a wind turbine generator, wherein the operating data of the wind turbine generator comprises a pitch angle, a generator rotating speed and active power, and the environmental parameters comprise a wind speed, a wind direction and an environmental temperature;
step 2: cleaning the collected wind turbine generator data, including processing missing values and abnormal values; clearing standby data, shutdown data and outlier data; establishing a health data set of normal operation data of the wind turbine generator;
and step 3: based on a health data set of the wind turbine generator, establishing mathematical description aiming at the health data set by utilizing an OC-RKELM model, wherein a one-dimensional index output by the OC-RKELM model represents the similarity degree of input real-time operation data and the health data set of the wind turbine generator;
and 4, step 4: determining a health threshold value based on analysis of historical normal data to judge whether the operating data belongs to a health data set of the wind turbine generator;
and 5: and correspondingly processing the real-time operation data of the wind turbine generator, judging whether the real-time operation data is normal operation data or not based on the OC-RKELM model, and performing early warning and shutdown inspection on the wind turbine generator when early warning conditions are met.
2. The method for monitoring the state of the wind turbine generator based on the OC-RKELM as claimed in claim 1, wherein the process of the step 2 is as follows:
step 21: clearing missing data in the acquired data; clearing data beyond the given range of the parameters;
step 22: in order to obtain pure normal operation data of the wind turbine generator, firstly, data of the generator in standby and shutdown states are removed; secondly, removing outlier data which are seriously deviated from the healthy data cluster through a local abnormal factor outlier detection algorithm;
step 23: and correspondingly normalizing the data after data cleaning in order to eliminate dimension influence.
3. The method for monitoring the condition of the wind turbine generator based on the OC-RKELM according to claim 2, wherein in the step 2, in order to eliminate the dimensional influence, the wind turbine generator data is normalized accordingly, the normalization is a linear transformation of the original data, so that the result of the transformation falls into the [0,1] interval, and the transformation function is as follows:
the outlier data is cleaned by adopting a local abnormal factor algorithm, wherein the local abnormal factor algorithm is defined by the following parameters:
k-distance: for a certain object p in the dataset D, defining a distance k-dist (p) between p and the object o e D (p, o), the object o satisfies the following two conditions: at least k objects o 'belong to D \ p }, and D (p, o') is less than or equal to D (p, o); at most k-1 objects o 'belong to D \ p }, and D (p, o') < D (p, o) is satisfied;
k-distance neighborhood: k-distance neighborhood N of object pk(p), defined as all objects in D that are not more than k-distance from p:
Nk(p)＝{q∈D\{p}|d(p,q)≤k-dist(p)} (2)
the reachable distance is: object o to object p reachable distance reach distk(p, o) is defined as:
reach-distk(p,o)＝max{k-dist(o),d(p,o)} (3)
i.e. the larger value in the k-distance of object p and object o;
local accessible density: local reachable density lrd of object pk(p) is defined as:
representing the ratio of objects within the k-distance neighborhood of object p to the reachable distance of object p;
local abnormality factor: the local anomaly factor for object p is defined as:
representing an average of a ratio of a local achievable density of objects within a k-distance neighborhood of p to a local achievable density of p, a local anomaly factor may describe an outlier of object p, when the local anomaly factor is less than or equal to 1, it is indicated that p is located within a data-dense region; when the local anomaly factor is greater than 1, it indicates that p may be far away from the data-dense region, with the larger the value the higher the degree of outlier of p.
4. The method for monitoring the state of the wind turbine generator based on the OC-RKELM as claimed in any one of claims 1 to 3, wherein the process of the step 3 is as follows: selecting a support vector: in order to select representative wind turbine generator operation data, a system sampling method is adopted for sampling, and the operation is as follows: parameters such as power, time and the like are segmented, corresponding sub-sample sets are divided, and samples are extracted from each sub-sample set to serve as support vectors.
5. The method for monitoring the state of the wind turbine generator based on the OC-RKELM as claimed in claim 4, wherein in the step 3, a state monitoring model of the OC-RKELM is required to be constructed. Extracting a number of health samples, X ═ X1,x2,…,xN]TThe target output vector is R ═ 1,1, …,1]T. The output is in the form of:
in the formula (I), the compound is shown in the specification,
for the model output parameter, betajFor the output weights, K (-) is the kernel function,
the method is a set of support vectors, selects a radial basis function as a kernel mapping function, and has the following form:
K(x,xj)＝exp(-||x-xj||2/σ) (7)
where σ is a hyperparameter, xjIs a support vector;
the single class problem solved by the OC-RKELM model is described as:
in the formula, KN×L＝K(X,XL) For simplified kernel matrices, training errors and output weights are balanced by introducing L2 regularization, equation (8) is rewritten as:
in the formula, ξ is a training error, λ is a regularization coefficient, and a lagrange function is introduced to solve the formula (9):
in the formula, a is a lagrange multiplier vector, and according to the carlo-kuen-tak condition, the solution of the output weight is obtained as follows:
in the formula, I is an identity matrix;
the distance of the sample from the normal class is defined as:
the greater the distance of the sample from the normal class, the more the sample deviates from the normal class.
6. The method for monitoring the state of the wind turbine generator based on the OC-RKELM as claimed in any one of claims 1 to 3, wherein the process of the step 4 is as follows: determining a health threshold by a kernel density estimation method, comprising: calculating a one-dimensional output index corresponding to normal operation data based on an OC-RKELM model, estimating a density function of the one-dimensional output index by using a nuclear density estimation method, and determining a health threshold value according to a certain confidence coefficient.
7. The method for monitoring the state of the wind turbine generator based on the OC-RKELM as claimed in claim 6, wherein in the step 4, the health threshold is determined by using a kernel density estimation method, and the decision function is defined as:
wherein Sign (·) is a Sign function, thres is a threshold, and when d (x) is less than or equal to the threshold, the monitoring data is considered to belong to normal data; when the value is larger than the threshold value, the monitoring data is considered as abnormal data.
8. The method for monitoring the state of the wind turbine generator based on the OC-RKELM as claimed in any one of claims 1 to 3, wherein the process of the step 5 is as follows:
step 51: inputting real-time operation data into the constructed OC-RKELM model, and judging the real-time operation data as abnormal operation data when the output index exceeds a health threshold value;
step 52: and carrying out statistical analysis on the output result of the OC-RKELM model, setting a proper early warning condition, and prompting the abnormality of the wind turbine generator when the early warning condition is met.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011519151.XA CN112682269A (en) | 2020-12-21 | 2020-12-21 | Wind turbine generator state monitoring method based on OC-RKELM |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011519151.XA CN112682269A (en) | 2020-12-21 | 2020-12-21 | Wind turbine generator state monitoring method based on OC-RKELM |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112682269A true CN112682269A (en) | 2021-04-20 |
Family
ID=75449946
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011519151.XA Pending CN112682269A (en) | 2020-12-21 | 2020-12-21 | Wind turbine generator state monitoring method based on OC-RKELM |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112682269A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114046228A (en) * | 2021-10-26 | 2022-02-15 | 华能(浙江)能源开发有限公司清洁能源分公司 | Wind turbine generator abnormity diagnosis method and system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105046374A (en) * | 2015-08-25 | 2015-11-11 | 华北电力大学 | Power interval predication method based on nucleus limit learning machine model |
| KR101599210B1 (en) * | 2015-07-08 | 2016-03-03 | 한국기계연구원 | A method of setting alarm levels for condition monitoring and fault diagnosis of wind turbine generator |
| CN110410282A (en) * | 2019-07-24 | 2019-11-05 | 河北工业大学 | Wind turbines health status on-line monitoring and method for diagnosing faults based on SOM-MQE and SFCM |
| CN110795690A (en) * | 2019-10-24 | 2020-02-14 | 大唐（赤峰）新能源有限公司 | Wind power plant operation abnormal data detection method |
| CN110930057A (en) * | 2019-12-06 | 2020-03-27 | 国网湖北省电力有限公司电力科学研究院 | Quantitative evaluation method for reliability of distribution transformer test result based on LOF algorithm |
| CN111160457A (en) * | 2019-12-27 | 2020-05-15 | 南京航空航天大学 | Turboshaft engine fault detection method based on soft class extreme learning machine |
| CN111539553A (en) * | 2020-03-31 | 2020-08-14 | 华北电力大学 | Wind turbine generator fault early warning method based on SVR algorithm and skewness |
-
2020
- 2020-12-21 CN CN202011519151.XA patent/CN112682269A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101599210B1 (en) * | 2015-07-08 | 2016-03-03 | 한국기계연구원 | A method of setting alarm levels for condition monitoring and fault diagnosis of wind turbine generator |
| CN105046374A (en) * | 2015-08-25 | 2015-11-11 | 华北电力大学 | Power interval predication method based on nucleus limit learning machine model |
| CN110410282A (en) * | 2019-07-24 | 2019-11-05 | 河北工业大学 | Wind turbines health status on-line monitoring and method for diagnosing faults based on SOM-MQE and SFCM |
| CN110795690A (en) * | 2019-10-24 | 2020-02-14 | 大唐（赤峰）新能源有限公司 | Wind power plant operation abnormal data detection method |
| CN110930057A (en) * | 2019-12-06 | 2020-03-27 | 国网湖北省电力有限公司电力科学研究院 | Quantitative evaluation method for reliability of distribution transformer test result based on LOF algorithm |
| CN111160457A (en) * | 2019-12-27 | 2020-05-15 | 南京航空航天大学 | Turboshaft engine fault detection method based on soft class extreme learning machine |
| CN111539553A (en) * | 2020-03-31 | 2020-08-14 | 华北电力大学 | Wind turbine generator fault early warning method based on SVR algorithm and skewness |
Non-Patent Citations (1)
| Title |
|---|
| QIAN LENG等: "One-Class Classification with Extreme Learning Machine", 《MATHEMATICAL PROBLEMS IN ENGINEERING》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114046228A (en) * | 2021-10-26 | 2022-02-15 | 华能(浙江)能源开发有限公司清洁能源分公司 | Wind turbine generator abnormity diagnosis method and system |
| CN114046228B (en) * | 2021-10-26 | 2023-11-07 | 华能(浙江)能源开发有限公司清洁能源分公司 | Wind turbine generator abnormality diagnosis method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110410282B (en) | SOM-MQE and SFCM (Small form-factor pluggable) based wind turbine generator health state online monitoring and fault diagnosis method | |
| CN111709490B (en) | Fan health state assessment method based on GRU neural network | |
| Wang et al. | A fast abnormal data cleaning algorithm for performance evaluation of wind turbine | |
| CN110362045B (en) | Marine doubly-fed wind turbine generator fault discrimination method considering marine meteorological factors | |
| Yang et al. | Fault early warning of wind turbine gearbox based on multi‐input support vector regression and improved ant lion optimization | |
| CN114370380A (en) | Wind turbine generator fault diagnosis method considering meteorological factors | |
| CN108734359A (en) | A kind of wind power prediction data preprocessing method | |
| CN115313930A (en) | Method for predicting excitation current of large-scale steam turbine generator in real time | |
| CN112682269A (en) | Wind turbine generator state monitoring method based on OC-RKELM | |
| CN108074045B (en) | Wind turbine generator complete machine vulnerability analysis and fault sequencing method and electronic terminal | |
| CN114463143A (en) | Method for enhancing SCADA fault data of offshore doubly-fed wind generator | |
| Dong et al. | Fault diagnosis and classification in photovoltaic systems using scada data | |
| CN111623905B (en) | Wind turbine generator bearing temperature early warning method and device | |
| CN117332353A (en) | Method for diagnosing abnormal vibration faults of rotor of steam turbine generator | |
| CN112508278A (en) | Multi-connected system load prediction method based on evidence regression multi-model | |
| CN116821811A (en) | Coal mill unit fault diagnosis method and system based on multi-layer graph convolution neural network | |
| Ma et al. | Anomaly Detection of Mountain Photovoltaic Power Plant Based on Spectral Clustering | |
| CN114154567A (en) | Wind power plant station operation data anomaly identification method based on machine learning | |
| Jing et al. | Two-layer PSDG based fault diagnosis for wind turbines | |
| CN111458143A (en) | Temperature fault diagnosis method for main bearing of wind turbine generator | |
| Zhang et al. | Fault Prediction Method for Distribution Network Outage Based on Feature Selection and Ensemble Learning | |
| Xie et al. | Data Cleaning and Modeling of Wind Power Curves | |
| Lu et al. | Data driven induction motor condition identification and fault prediction | |
| Léonard et al. | Hydro-turbine monitoring: from self-learned equipment behavior to a single global deviation indicator | |
| Xuefeng et al. | Study on Multi-Sensor Parameter Warning Based on Gaussian Mixed Model and Nonlinear State Estimation Technique |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210420 |