CN110874665B - Control device and method for wind generating set - Google Patents
Control device and method for wind generating set Download PDFInfo
- Publication number
- CN110874665B CN110874665B CN201811012110.4A CN201811012110A CN110874665B CN 110874665 B CN110874665 B CN 110874665B CN 201811012110 A CN201811012110 A CN 201811012110A CN 110874665 B CN110874665 B CN 110874665B
- Authority
- CN
- China
- Prior art keywords
- sample
- data
- temperature
- value
- attribute
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000011217 control strategy Methods 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims description 42
- 230000008569 process Effects 0.000 claims description 26
- 238000012795 verification Methods 0.000 claims description 23
- 238000007781 pre-processing Methods 0.000 claims description 21
- 238000004140 cleaning Methods 0.000 claims description 17
- 239000011159 matrix material Substances 0.000 claims description 14
- 238000005457 optimization Methods 0.000 claims description 14
- 238000010200 validation analysis Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 7
- 238000005406 washing Methods 0.000 claims description 7
- 239000004519 grease Substances 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 5
- 238000000638 solvent extraction Methods 0.000 claims description 3
- 238000005192 partition Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 9
- 239000013598 vector Substances 0.000 description 9
- 230000000694 effects Effects 0.000 description 8
- 238000004422 calculation algorithm Methods 0.000 description 7
- 230000008901 benefit Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000012706 support-vector machine Methods 0.000 description 5
- 230000015654 memory Effects 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000010248 power generation Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000017525 heat dissipation Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010223 real-time analysis Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- 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
Abstract
A control apparatus and method for a wind turbine generator set are provided. The device comprises: a data obtaining unit configured to obtain historical operating data of a wind turbine generator set, wherein the historical operating data comprises a plurality of samples, and attributes of each sample comprise: the system comprises a temperature attribute indicating the temperature of a main shaft bearing of the wind generating set, an output power attribute indicating the output power of the wind generating set, a temperature attribute indicating the ambient temperature of the wind generating set and a state attribute indicating the running state of the wind generating set; a model creation unit configured to create a prediction model using the historical operation data; a prediction unit configured to predict a state of the wind park using the collected current operating data of the wind park based on the created prediction model and to determine a corresponding control strategy based on the prediction result.
Description
Technical Field
The present application relates generally to the field of wind power, and more particularly, to a control apparatus and method for a wind turbine generator set.
Background
In the operation process of the wind generating set, the wind generating set needs to monitor the temperature of the main shaft bearing and the generator in real time so as to ensure that the main shaft bearing and the generator are within the designed temperature range, and thus, the wind generating set can be ensured to run healthily on the premise of meeting the service life of the set.
In the prior art, the temperature of a generator is usually monitored, and the temperature rise of the generator are adopted to control whether a wind generating set operates, so that a main shaft bearing and the generator are always within a design temperature range, the temperature monitoring of the main shaft bearing with worse heat dissipation condition caused by being in a closed environment is ignored, the temperature of the main shaft bearing is often higher, the service life of the wind generating set cannot reach a design expected target, the later potential maintenance cost of the wind generating set is indirectly higher, and the income of a wind power plant is influenced.
In addition, even if some wind driven generator sets monitor the temperature of the main shaft bearing, the shutdown operation is only carried out based on the data acquired by the temperature sensor singly, the influence of the region, the environment temperature and the output power of the wind driven generator set on the shutdown is not comprehensively considered, and the influence of the temperatures of different parts of the main shaft bearing on the shutdown is also not considered.
On the other hand, the conventional shutdown scheme is to perform shutdown operation when an excessive temperature is detected. However, the wind power plant is in a severe environment, which is often affected by extreme weather, and is located in a remote position, so that the shutdown is usually continued for a long time, which directly affects the economic benefit of the owner.
Disclosure of Invention
In order to solve at least the above problems in the prior art, the present application provides a control apparatus and method for a wind turbine generator system, which can predict the state of the wind turbine generator system and make a corresponding control strategy based on the current operating state of the wind turbine generator system by analyzing and learning the past historical operating data of the wind turbine generator system, thereby reducing the downtime and improving the economic benefit of the wind turbine plant.
According to an exemplary embodiment of the invention, a control device for a wind park is provided, the device comprising: a data obtaining unit configured to obtain historical operating data of the wind turbine generator set, wherein the historical operating data comprises a plurality of samples, and an attribute of each sample comprises: the system comprises a temperature attribute indicating the temperature of a main shaft bearing of the wind generating set, an output power attribute indicating the output power of the wind generating set, a temperature attribute indicating the ambient temperature of the wind generating set and a state attribute indicating the running state of the wind generating set; a model creation unit configured to create a prediction model using the historical operation data; and a prediction unit configured to predict a state of the wind park using the collected current operating data of the wind park based on the created prediction model, and to determine a corresponding control strategy based on the prediction result.
Optionally, the temperature property indicative of the temperature of the main shaft bearing of the wind park may comprise at least one of the following temperature properties: the temperature control system comprises a first temperature attribute indicating the temperature of a front end bearing inner ring of the wind generating set, a second temperature attribute indicating the temperature of a front end bearing outer ring of the wind generating set, a third temperature attribute indicating the temperature of a rear end bearing inner ring of the wind generating set, a fourth temperature attribute indicating the temperature of a rear end bearing outer ring of the wind generating set and a fifth temperature attribute indicating the temperature of grease on a main shaft bearing.
Optionally, the apparatus may further comprise: a data preprocessing unit configured to preprocess the historical operating data acquired by the data acquisition unit, and the model creation unit may create the prediction model using the preprocessed historical operating data. The data preprocessing unit may include: the data cleaning module is configured to perform data cleaning processing on the acquired historical operating data to remove singular items in the samples of the historical operating data; the data serialization module is configured to serialize the historical operation data subjected to the data cleaning processing to obtain a corresponding matrix of M multiplied by N dimensions, wherein M represents the total number of samples of the historical operation data, and N represents the number of attributes of each sample of the historical operation data; a data normalization module configured to normalize the data values associated with each attribute in the M N dimensional matrix.
Optionally, the data cleansing module may cleanse the historical operating data by: determining, for a particular temperature attribute among temperature attributes of samples of historical operating data, whether a difference between a sample value of the particular temperature attribute and an expected value of a t-th sample in the historical operating data is greater than a first predetermined threshold; determining the sample value as a singular item and replacing the sample value with the expected value as the data value of the specific temperature attribute of the t-th sample if the difference is greater than a first predetermined threshold, determining the sample value as the data value of the specific temperature attribute of the t-th sample if the difference is not greater than the first predetermined threshold, wherein the expected value of the specific temperature attribute of the t-th sample is determined based on the data value of the specific temperature attribute of the (t-1) th sample and the data value of the specific temperature attribute of the (t-2) th sample, and 3 ≦ t ≦ M.
Optionally, if the number of the plurality of samples successively determined to have singular terms is greater than a first predetermined number for the particular temperature attribute, the data cleansing module may be further configured to: determining whether a difference between a sample value and an expected value of a particular temperature attribute of a particular sample subsequent to a plurality of samples successively determined as singular items is greater than a first predetermined threshold and less than a second predetermined threshold, if the difference between the sample value and the expected value of the particular temperature attribute of the particular sample is greater than the first predetermined threshold and less than the second predetermined threshold, determining the sample value of the particular temperature attribute of the particular sample as the data value of the particular temperature attribute of the particular sample, and if the difference between the sample value and the expected value of the particular temperature attribute of the particular sample is greater than or equal to the second predetermined threshold, determining the expected value of the particular temperature attribute of the particular sample as the data value of the particular temperature attribute of the particular sample.
Optionally, if the number of the plurality of samples successively determined to have singular terms is greater than a second predetermined number for the particular temperature attribute, the data cleansing module may be further configured to: selecting a new sample among the samples of the historical operating data as a starting sample of the data cleansing process, wherein the second predetermined number is greater than the first predetermined number.
Optionally, if another sample satisfying a predetermined condition is found after the particular sample having a difference between the sample value of the particular temperature attribute and the expected value greater than or equal to a second predetermined threshold, the data cleansing module may be further configured to: selecting a new sample from the samples of the historical operating data as a starting sample of the data cleaning process, wherein the predetermined condition is that: a difference between a sample value of the particular temperature attribute of the other sample and a desired value is greater than a first predetermined threshold and less than a second predetermined threshold.
Alternatively, the model creation unit may include: a data set partitioning module configured to partition the historical operating data into a training data set and a validation data set; a model training module configured to create and train a predictive model using a training dataset and obtain corresponding predictive model parameters; and the model optimization module is configured to optimize the corresponding prediction model parameters by using the verification data set based on the created prediction model, and determine the optimal prediction model parameters as the prediction model parameters of the created prediction model to obtain the optimal prediction model.
According to another exemplary embodiment of the invention, a control method for a wind park is provided, the method comprising: obtaining historical operating data of a wind generating set, wherein the historical operating data comprises a plurality of samples, and attributes of each sample comprise: the system comprises a temperature attribute indicating the temperature of a main shaft bearing of the wind generating set, an output power attribute indicating the output power of the wind generating set, a temperature attribute indicating the ambient temperature of the wind generating set and a state attribute indicating the running state of the wind generating set; creating a predictive model using historical operating data; and predicting the state of the wind generating set by using the collected current operation data of the wind generating set based on the created prediction model, and determining a corresponding control strategy based on the prediction result.
Optionally, the temperature attribute indicative of the temperature of the main shaft bearing of the wind park may comprise: the temperature control system comprises a first temperature attribute indicating the temperature of a front end bearing inner ring of the wind generating set, a second temperature attribute indicating the temperature of a front end bearing outer ring of the wind generating set, a third temperature attribute indicating the temperature of a rear end bearing inner ring of the wind generating set, a fourth temperature attribute indicating the temperature of a rear end bearing outer ring of the wind generating set and a fifth temperature attribute indicating the temperature of grease on a main shaft bearing.
Optionally, the method may further comprise: the acquired historical operating data is preprocessed prior to creating the predictive model, and the step of creating the predictive model may use the preprocessed historical operating data to create the predictive model. The step of preprocessing the acquired historical operating data samples may include: performing data cleaning processing on the acquired historical operating data to remove singular items in a sample of the historical operating data; serializing the historical operating data subjected to the data cleaning processing to obtain a corresponding matrix of M multiplied by N dimensions, wherein M represents the total number of samples of the historical operating data, and N represents the number of attributes of each sample of the historical operating data; normalizing the data values associated with each attribute in the M N dimensional matrix.
Optionally, the step of performing data cleaning processing on the acquired historical operating data to remove singular items in the sample of the historical operating data may include: determining, for a particular temperature attribute among temperature attributes of samples of historical operating data, whether a difference between a sample value of the particular temperature attribute and an expected value of a t-th sample in the historical operating data is greater than a first predetermined threshold; determining the sample value as a singular item and replacing the sample value with the expected value as the data value of the specific temperature attribute of the t-th sample if the difference is greater than a first predetermined threshold, determining the sample value as the data value of the specific temperature attribute of the t-th sample if the difference is not greater than the first predetermined threshold, wherein the expected value of the specific temperature attribute of the t-th sample is determined based on the data value of the specific temperature attribute of the (t-1) th sample and the data value of the specific temperature attribute of the (t-2) th sample, and 3 ≦ t ≦ M.
Optionally, the method may further comprise: if the number of the plurality of samples successively determined to have singular terms is greater than a first predetermined number for the specific temperature attribute, determining whether a difference between a sample value of the particular temperature property and an expected value of a particular sample subsequent to the plurality of samples successively determined as singular terms is greater than a first predetermined threshold and less than a second predetermined threshold, if the difference between the sample value of the particular temperature attribute of the particular sample and the expected value is greater than a first predetermined threshold and less than a second predetermined threshold, determining a sample value of the particular temperature property of the particular sample as a data value of the particular temperature property of the particular sample, if the difference between the sample value of the particular temperature attribute of the particular sample and the expected value is greater than or equal to a second predetermined threshold, a desired value of the particular temperature attribute for the particular sample is determined as a data value of the particular temperature attribute for the particular sample.
Optionally, the method may further comprise: selecting a new sample among the samples of the historical operating data as a starting sample of the data washing process if the number of the plurality of samples continuously determined to have the singular term is greater than a second predetermined number for the specific temperature attribute, wherein the second predetermined number is greater than the first predetermined number.
Optionally, the method may further comprise: selecting a new sample among the samples of the historical operating data as a starting sample of the data washing process if another sample satisfying a predetermined condition is found after the specific sample having a difference between the sample value of the specific temperature property and the expected value greater than or equal to a second predetermined threshold, wherein the predetermined condition is: a difference between a sample value of the particular temperature attribute of the other sample and a desired value is greater than a first predetermined threshold and less than a second predetermined threshold.
Optionally, the step of creating a predictive model using historical operating data may include: dividing historical operating data into a training data set and a verification data set; establishing and training a prediction model by using a training data set, and acquiring corresponding prediction model parameters; and optimizing the corresponding prediction model parameters by using a verification data set based on the created prediction model, and determining the optimal prediction model parameters as the prediction model parameters of the created prediction model to obtain the optimal prediction model.
According to another exemplary embodiment of the present invention, a computer-readable storage medium is provided, in which a computer program is stored, characterized in that the computer program comprises instructions for executing the aforementioned control method.
According to another exemplary embodiment of the present invention, a computer device is provided, comprising a readable medium having a computer program stored thereon, characterized in that the computer program comprises instructions for executing the aforementioned control method.
Advantageous effects
By applying the control device and the method for the wind generating set, the operating condition of the wind generating set can be monitored more accurately by acquiring the temperatures of a plurality of positions of the main shaft bearing of the wind generating set, the operating state of the wind generating set can be predicted more accurately, and a corresponding control strategy can be formulated, so that the temperature of the wind generating set can be controlled by power reduction operation and other modes before the shutdown temperature of the wind generating set is reached, various faults which may occur can be predicted effectively, the shutdown time is reduced effectively, the loss of the wind generating set is reduced, and the wind power generation benefit is improved.
Additional aspects and/or advantages of the present general inventive concept will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the general inventive concept.
Drawings
The above and/or other aspects of the exemplary embodiments will be described below with reference to the accompanying drawings, in which:
fig. 1 is a block diagram illustrating a control apparatus for a wind turbine generator set according to an exemplary embodiment of the present invention.
Fig. 2 is a diagram illustrating a data preprocessing unit according to an exemplary embodiment of the present invention.
Fig. 3A is a block diagram illustrating a model creation unit of a control apparatus for a wind park according to an exemplary embodiment of the present invention.
Fig. 3B is a schematic diagram illustrating the operation of the model creation unit of the control apparatus for a wind park according to an exemplary embodiment of the present invention.
Fig. 4 is a flowchart illustrating a control method for a wind park according to an exemplary embodiment of the present invention.
The present invention will hereinafter be described in detail with reference to the drawings, wherein like or similar elements are designated by like or similar reference numerals throughout.
Detailed Description
The following description is provided with reference to the accompanying drawings to assist in a comprehensive understanding of exemplary embodiments of the invention as defined by the claims and their equivalents. The description includes various specific details to aid understanding, but these details are to be regarded as illustrative only. Thus, one of ordinary skill in the art will recognize that: various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Moreover, descriptions of well-known functions and constructions may be omitted for clarity and conciseness.
Fig. 1 is a block diagram illustrating a control apparatus 100 for a wind turbine generator set according to an exemplary embodiment of the present invention.
Referring to fig. 1, a control apparatus 100 according to an exemplary embodiment of the present invention may include a data acquisition unit 110, a model creation unit 120, and a prediction unit 130.
The data acquisition unit 110 may be used to acquire historical operating data of the wind turbine generator set. In an exemplary embodiment of the present invention, the historical operating data may include a plurality of samples, and each sample may include a plurality of attributes.
For example, in a wind generating set, a main shaft bearing is a rotating component for connecting a generator and a base, and there may be one, two or more different wind generating sets, which bear large load, and when the working environment of the wind generating set is severe, it is the weakest link in the operation of the wind generating set, and in addition, the environmental temperature has a great influence on the performance of electrical components. Thus, in an exemplary embodiment of the present invention, the attributes of the sample may include: a temperature attribute indicative of a temperature of a main shaft bearing of the wind turbine generator set, an output power attribute indicative of an output power of the wind turbine generator set, a temperature attribute indicative of an ambient temperature of the wind turbine generator set, and a status attribute indicative of an operational status of the wind turbine generator set. Here, the operating state of the wind park may include, for example, a normal operating state, a fault state (the control strategy for which may be shutdown), and a critical state (i.e., the temperature is approaching the critical temperature for the fault, the strategy for which may be derated). However, it should be understood that the present invention is not so limited and more or fewer states may be set depending on the desired control strategy of the wind park.
In an exemplary embodiment of the invention, said temperature profile indicative of the temperature of the main shaft bearing of the wind park may comprise one or more of the following temperature profiles: for example, a first temperature attribute indicative of a front end bearing inner ring temperature of the wind turbine generator set, a second temperature attribute indicative of a front end bearing outer ring temperature of the wind turbine generator set, a third temperature attribute indicative of a rear end bearing inner ring temperature of the wind turbine generator set, a fourth temperature attribute indicative of a rear end bearing outer ring temperature of the wind turbine generator set, and a fifth temperature attribute indicative of a grease temperature on the main shaft bearing.
In addition, it should be understood that more attributes can be set for each sample according to actual conditions.
The model creation unit 120 may create a predictive model using historical operating data.
Here, if the historical operating data acquired by the data acquisition unit 110 is in a form that can be directly used by the model creation unit 120, the model creation unit 120 may directly use the historical operating data acquired by the data acquisition unit 110 to create a prediction model.
Preferably, if the historical operating data acquired by the data acquisition unit 110 is difficult to be directly used by the model creation unit 120, or in order to make the prediction effect of the prediction model created by the model creation unit 120 using the historical operating data better, some data preprocessing operation may also be performed on the historical operating data before the directly acquired historical operating data is provided to the model creation unit 120. For example only, in an exemplary embodiment of the present invention, although not shown in fig. 1, the control device 100 may further include a data preprocessing unit configured to preprocess the historical operating data acquired by the data acquiring unit, so that the model creating unit 120 may create a prediction model using the preprocessed historical operating data to obtain a prediction model with better prediction effect. However, it should be understood that the data preprocessing unit may also be included in the data acquisition unit 110 as part of the data acquisition unit 110.
The process of preprocessing the historical operating data by the data preprocessing unit will be described in more detail below with reference to fig. 2.
Fig. 2 is a block diagram illustrating a data preprocessing unit 200 according to an exemplary embodiment of the present invention.
As shown in fig. 2, the data preprocessing unit 200 may include: a data cleansing module 210, a data serialization module 220, and a data normalization module 230.
The data cleansing module 210 may be used to perform a data cleansing process on the acquired historical operating data to eliminate singular items in the samples of the historical operating data.
In the exemplary embodiment of the invention, the singular term refers to sample data with obvious errors generated when the wind generating set is accidentally subjected to strong interference due to an extreme environment in operation, and the existence of the singular term has an important influence on the sample structure of historical operation data, and even may cause that a prediction model is not converged so that the model establishment fails. Therefore, the elimination of the singular items is beneficial to improving the prediction effect of the model.
In an exemplary embodiment of the present invention, the data cleansing module 210 may remove the singular item by using the temperature attribute of the sample because the temperature belongs to a slowly changing physical quantity, and no sudden change is possible, that is, if the temperature difference between the same temperature attribute of the current sample and the previous sample is much greater than a given threshold, the temperature value of the temperature attribute of the current sample may be considered as the singular item, and should be removed, and a corresponding data value may be supplemented as the data value of the temperature attribute of the current sample according to a preset difference scheme. In an exemplary embodiment of the present invention, if each sample includes a plurality of temperature attributes, the data cleansing process may be performed on the samples of historical operating data for each temperature attribute, respectively. This process will be described in detail below.
Assume that a sample value of a specific temperature attribute among temperature attributes of the t-th sample in the historical operating data is xtExpected value xt’＝Xt-1+(Xt-1-Xt-2) Where t is greater than or equal to 3 and less than or equal to M, M represents the total number of samples of historical operating data, Xt-1Refers to the specific temperature attribute of the t-1 th sample in the historical operating dataData value of (2), Xt-2Refers to the data value of the particular temperature attribute for the t-2 th sample in the historical operating data. The t-th sample, the t-1 th sample and the t-2 th sample mentioned here are samples taken at the t-th sampling time, the t-1 th sampling time and the t-2 th sampling time, respectively, in chronological order. Further, in the exemplary embodiment of the present invention, the data washing process for the subsequent sample may be performed by determining the sample value of the specific temperature attribute of the selected start sample and the sample subsequent to the start sample in the data washing process as its corresponding data value by default.
First, data washing module 210 may determine sample value x for the particular temperature attribute of the t-th sampletAnd the expected value xt' difference of | xt-xtWhether or not' | is greater than a first predetermined threshold W1. If the difference | xt-xt' | > W1, the sample value x may be determinedtFor singular terms, the data cleansing module 210 may now use the expected value xt' instead of the sample value xtData value X of the specific temperature attribute as the t-th sampletI.e. such that the data value X of the particular temperature property of the t-th samplet＝xt'. Conversely, if the difference | xt-xt' < W1, the data washing module 210 may determine the sample value xtNot singular terms, in which case the data cleansing module 210 may apply the sample value xtDirectly determining a data value X for the particular temperature attribute for the t-th sampletI.e. such that the data value X of the particular temperature property of the t-th samplet＝xt。
However, if a plurality of singular terms are detected in succession for the specific temperature property, this may mean that the sample value of the specific temperature property of the initial sample selected in the data cleansing process may itself be a singular term, or a calculation deviation occurs in the data cleansing process, resulting in an erroneous expected value being generated in the data cleansing process for the subsequent sample, thereby determining a succession of singular terms. In this regard, in an exemplary embodiment of the present invention, if the number of the plurality of samples successively determined to have the singular term is greater than the first predetermined number K for the specific temperature property, the data cleansing module 210 may determine whether a difference between sample values and expected values of the specific temperature property of samples subsequent to the plurality of samples successively determined to have the singular term is greater than a first predetermined threshold W1 and less than a second predetermined threshold W2. In an exemplary embodiment of the present invention, the first predetermined threshold value W1 and the second predetermined threshold value W2 may be set by a user according to experiments, experience, or other conditions, for example, W1 may be set to 0.5, and W2 may be set to 2 × W1.
Here, it is assumed that the sample value of the above-described specific temperature property for K consecutive samples (where 3. ltoreq. i.ltoreq.M) that have existed before the ith sample is determined to be a singular term whose sample value is xiThe expected value is xi'. At this time, it may be determined whether W2 > | x is satisfiedi-xi' | > W1. If W2 > | x is satisfiedi-xi' | > W1, then data cleaning module 210 may not reject sample value x for the particular temperature attribute of the sampleiBut the sample value x of the specific temperature property of the ith sampleiDirectly determining a data value X for the particular temperature attribute for the ith samplei。
However, if | xi-xi' | ≧ W2, the data cleansing module 210 may sample the value x of the sample for the particular temperature attribute of the sampleiDetermined as singular terms and then taking the expected value x of the specific temperature property of the ith samplei' determining the data value X of the specific temperature attribute as the i-th samplei. In this case, if | x is satisfiedi-xi' i ≧ W2, and then another sample (e.g., the jth sample) is found again to satisfy W2 > | xj-xj’|≥W1(xjAnd xj' sample value and expected value of the jth sample, respectively), the data cleansing module 210 may automatically select a new sample among the samples of historical operating data as a starting sample for the data cleansing process, where the new sample may be the other sample or after the other sampleThe sample of (1).
Alternatively, if the number of the plurality of samples successively determined to have singular terms is greater than a second predetermined number L (L > K) for the specific temperature property, the data cleansing module 210 may also directly and automatically select a new sample as the starting sample for the data cleansing process, where the new sample may be a sample subsequent to the plurality of samples successively determined to have singular terms.
After completing the data cleansing process, the data serialization module 220 may serialize the historical operating data that has undergone the data cleansing process to obtain a corresponding M × N dimensional matrix, where N represents the number of attributes for each sample of the historical operating data.
For example only, each sample may include the following 7 attributes: the system comprises the environment temperature of the wind generating set, the temperature of the front end bearing inner ring, the temperature of the front end bearing outer ring, the temperature of the rear end bearing outer ring, output power and the running state of the wind generating set.
In addition, in the exemplary embodiment of the present invention, in the process of serializing the historical operating data, data of each sample of the historical operating data may be encoded separately,
alternatively, in an exemplary embodiment of the present invention, since the dimensions of the attributes are not consistent and there is a difference in magnitude, the data normalization module 230 may perform a normalization process on the data value related to each attribute in the matrix of M × N dimensions, where the normalization process is as follows:
in equation (1), XiData value, X, representing some property of the ith samplemaxAnd XminRespectively representing a maximum value and a minimum value among data values related to the attribute in the matrix of the M × N dimensions.
Through the normalization process, the following data sets in matrix form can be obtained:
m denotes the number of samples, N denotes the number of attributes per sample, XijAnd the value of the j attribute of the ith sample is expressed, wherein i is more than or equal to 1 and less than or equal to M, j is more than or equal to 1 and less than or equal to N, and i and j are integers. The data set may be used by the model creation unit 120 to create a predictive model, as will be described in more detail below in conjunction with FIG. 3. However, it should be understood that the above-described manner of data preprocessing and the expression of the data set are merely examples listed for easy understanding, the present invention is not limited thereto, more or less data preprocessing operations may be performed on the historical operating data acquired by the data acquisition unit 110 according to requirements, and different expression forms of the data set may be used according to the following prediction model creation method.
Fig. 3A is a block diagram illustrating the model creation unit 120 of the control apparatus 100 for a wind park according to an exemplary embodiment of the present invention. Fig. 3B is a schematic diagram illustrating the operation of the model creation unit 120 of the control device 100 for a wind park according to an exemplary embodiment of the present invention. In the diagram shown in fig. 3B, the flow indicated by the solid arrow is related to the creation and training process of the prediction model, and the flow indicated by the dotted arrow is related to the verification process of the prediction model.
As shown in fig. 3A, the model creation unit 120 may include a model training module 121, a model optimization module 122, and a dataset partitioning module 123.
The data set dividing module 123 may divide the received samples of the historical operating data into a training data set and a verification data set, for example, the data set dividing module 123 may divide a part of the received samples of the historical operating data into the training data set and divide the rest of the received samples into the training data set, and the ratio of the two may be set according to requirements.
The model training module 121 may create and train a prediction model using the training data set and obtain corresponding prediction model parameters, and the model optimization module 122 may optimize the corresponding prediction model parameters using the validation data set based on the created prediction model and determine the optimal prediction model parameters as the prediction model parameters of the created prediction model to obtain the optimal prediction model.
For example only, when a predictive model is created using preprocessed historical operating data, a data set, for example, in the form of a matrix as shown in equation (2), may be divided into two data sets as follows:
wherein, the data set X(1)The P line data in (b) may be P line data (i.e., data of P samples) arbitrarily selected from M line data in the data set X in the form of a matrix shown in equation (2), not necessarily consecutive P line data. Data set X(2)The (M-P) line data in (M-P) is the (M-P) line data remaining after the P line data is removed from the M line data of the data set X (i.e., (M-P) sample data). The above two data sets X(1)And data set X(2)The training data set and the validation data set may be used separately and may be arbitrarily sized, e.g., data set X may be divided into the training data set and the validation data set in a 7:3 ratio.
Further, it should be understood that although the data set X is divided into two parts in the above description, that is, the data set X(1)And data set X(2)However, the present invention is not limited to this, and the data set X may be constituted by selecting an arbitrary number of rows from the data set X(1)And data set X(2)。
In an exemplary embodiment of the present invention, the model training module 121 may create a prediction model by training feature parameters of a Support Vector Machine (SVM) algorithm using a training data set based on an SVM algorithm, and acquire corresponding prediction model parameters. Each row of data in the training dataset may represent a training sample vector, and the prediction model may be represented as equation (3) below:
f(X)＝ωTφ(X)+b……(3)
wherein X is an input training sample vector, omega is a weight vector parameter of the prediction model, and omega is a weight vector of the prediction modelTIs the transpose of ω, φ (X) is a non-linear feature mapping from the input space to the feature space where each training sample vector can be represented as a corresponding point, and b is the residual term.
It should be understood that the SVM algorithm is only one example of an algorithm for creating the predictive model, and the present invention is not limited thereto, and various other algorithms may be used to create the predictive model, for example, a neural network algorithm, a machine learning algorithm, and the like.
After introducing the lagrange multiplier in equation (3) above, the SVM-based prediction model can be further expressed as:
wherein, the kernel function K adopts a gaussian radial basis kernel function form, namely:
wherein, XiFor the input training sample vector, σ is the kernel function parameter, λiIs a group of groups with K (X, X)i) The corresponding weight.
The model training module 121 may train the prediction model represented by equation (4) above using a training data set to obtain corresponding prediction model parameters σ and C, where C is an error penalty factor for the model and represents the tolerance of the prediction model to prediction errors. In an exemplary embodiment of the present invention, the parameter C is a weight for weighing the model loss and the classification interval of the model, and the initial value thereof may be selected by human according to the needs of the model generation tool (e.g., matlab), and may be continuously optimized by using the sample for training or verification during the building, training and subsequent verification of the prediction model. This will be further described below.
The model optimization module 122 may use the validation data set to validate the predictive effect of the predictive model and adjust and optimize variable parameters (e.g., σ and C) in the model training based on the predictive effect, determine the variable parameters that achieve the best predictive effect as the final parameters of the predictive model, thereby obtaining an optimal predictive model, as described in more detail below.
As shown in FIG. 3B, during the verification process, the model optimization module 122 may set desired results and input data based on the verification dataset. For example, model optimization module 122 may input the data of the t-th sample in the validation dataset as input data to the prediction model, predicting the data of the t + 1-th sample. Thereafter, the model optimization module 122 can compare the predicted results (i.e., the predicted data for the t +1 th sample) with respect to the input data to the expected results (i.e., the actual data for the t +1 th sample in the validation dataset).
By way of example only, in an exemplary embodiment of the invention, the model optimization module 122 may input the data of the t-th sample related to the attributes of the wind turbine generator system, such as the ambient temperature of the wind turbine generator system, the inner ring temperature of the front end bearing, the outer ring temperature of the rear end bearing, the output power, and the operation state of the wind turbine generator system, as input data to the prediction model, predict the data of the t + 1-th sample related to the attributes of the wind turbine generator system, such as the ambient temperature of the wind turbine generator system, the inner ring temperature of the front end bearing, the outer ring temperature of the rear end bearing, the output power, and the operation state of the wind turbine generator system, the predicted data relating to the above-mentioned attribute of the t +1 th sample is then compared with the data of the corresponding attribute of the t +1 th sample in the validation dataset, respectively.
If the hit rates of the predicted result and the expected result for a plurality of samples in the entire verification dataset are greater than a predetermined threshold (i.e., the ratio of the number of samples for which the predicted result is consistent with the expected result to the total number of samples in the verification dataset exceeds the predetermined threshold for a plurality of samples in the verification dataset), then the prediction model may be determined to be the optimal prediction model and its prediction model parameters are the optimal prediction model parameters.
If the hit rates of the predicted result and the expected result are not greater than the predetermined threshold for the plurality of samples in the entire verification dataset (i.e., the ratio of the number of samples with the same predicted result as the expected result to the total number of samples in the verification dataset does not exceed the predetermined threshold for the plurality of samples in the verification dataset), the model optimization module 122 may adjust the variable prediction model parameters that can be adjusted in the model training, then use the verification dataset to verify the prediction model with the adjusted variable prediction model parameters again, and repeat the above process until the hit rates of the predicted result and the expected result obtained when the prediction model is verified using the verification dataset are greater than the predetermined threshold. In an exemplary embodiment of the present invention, the predetermined threshold may be a threshold that is arbitrarily set by a user according to a demand, for example, 99%.
For example only, the model optimization module 122 may use the validation data set to optimize the aforementioned predictive model parameters σ and C.
Specifically, the ranges of the prediction model parameters σ and C may be set first. The parameter C is an error penalty factor aiming at the prediction model and represents the tolerance of the prediction model to prediction errors, the higher the C is, the more the error cannot be tolerated, so that overfitting is easy to cause, the smaller the C is, so that under-fitting is easy to cause, and the overlarge or undersize of the C can cause the deterioration of generalization capability, thereby causing the reduction of prediction accuracy. Sigma is a gaussian kernel function, and the parameter implicitly determines the distribution of data after mapping to a new feature space, the smaller sigma is, the fewer support vectors are, the poor classification effect on an unknown sample (verification data set) is, and thus the possibility that the accuracy of the training data set is high but the accuracy of the verification data set is not high (i.e. over-training) may occur, while the larger sigma is, the more support vectors are, in this case, a particularly high accuracy may not be obtained on the training data set, and the accuracy of the verification data set is further affected, and besides, the number of support vectors may also affect the speed of training and prediction. Therefore, in an exemplary embodiment of the present invention, the prediction model parameters σ and the range of C may be set empirically, for example, the range of C may be set empirically to 0.0001 to 10000, and the value of σ is set to be selected from the following sequence: 0.1, 0.2, 0.4, 0.5, 0.6, 0.8, 1.6, 3.2, 6.4, 12.8.
The model optimization module 122 may obtain prediction accuracy (i.e., hit rate) for different parameter value combinations of σ and C using the validation dataset, and compare the obtained prediction accuracy, thereby determining a set of σ and C parameter values with the highest prediction accuracy as the optimal prediction model parameters.
Alternatively, if there are multiple sets of parameter value combinations of σ and C corresponding to the same highest prediction accuracy, the model optimization module 122 may determine the parameter values of σ and C in the set of parameter value combinations having the smallest C among the multiple sets of parameter value combinations as the optimal prediction model parameters.
Then, the prediction unit 130 may predict the state of the wind turbine generator system by using the collected current operation data of the wind turbine generator system (e.g., the main shaft bearing temperature of the wind turbine generator system (including the front end bearing inner ring temperature, the front end bearing outer ring temperature, the rear end bearing inner ring temperature, the rear end bearing outer ring temperature, the ambient temperature, the output power, the operation state, etc.) based on the prediction model (e.g., the optimal prediction model with the optimal prediction model parameters obtained by using the above manner), and determine a corresponding control strategy based on the prediction result, such as whether the wind turbine generator system needs to be operated with reduced power, whether the wind turbine generator system needs to be shut down and overhauled, etc.
For example only, the prediction unit 130 may add the prediction model parameters of the prediction model obtained from the model creation unit 120 to a TCwind library, perform real-time analysis on the wind turbine generator set using the library, so that a corresponding control strategy (e.g., whether the wind turbine generator set needs to be powered down, whether shutdown for maintenance is required) may be determined, and notify a user of the determination.
Fig. 4 is a flowchart illustrating a power control prediction method for a wind park according to an exemplary embodiment of the present invention.
Referring to fig. 4, in step 401, historical operating data of the wind turbine generator set may be acquired by the data acquisition unit 110 of the control apparatus 100, and in an exemplary embodiment of the present invention, the historical operating data may include a plurality of samples, and the attribute of each sample may include: a temperature attribute indicative of a temperature of a main shaft bearing of the wind turbine generator set, an output power attribute indicative of an output power of the wind turbine generator set, a temperature attribute indicative of an ambient temperature of the wind turbine generator set, and a status attribute indicative of an operational status of the wind turbine generator set.
In an exemplary embodiment of the invention, the temperature property indicative of the temperature of the main shaft bearing of the wind park may comprise: the temperature control system comprises a first temperature attribute indicating the temperature of a front end bearing inner ring of the wind generating set, a second temperature attribute indicating the temperature of a front end bearing outer ring of the wind generating set, a third temperature attribute indicating the temperature of a rear end bearing inner ring of the wind generating set, a fourth temperature attribute indicating the temperature of a rear end bearing outer ring of the wind generating set and a fifth temperature attribute indicating the temperature of grease on a main shaft bearing.
Thereafter, at step 403, a prediction model may be created by the model creation unit 120 of the control apparatus 100 using the historical operation data. Here, the historical operating data used by the model creating unit 120 may be historical operating data directly obtained from the data obtaining unit 110, or preferably, may also be historical operating data preprocessed by the data preprocessing unit 200 shown in fig. 2, and the detailed operations of the data preprocessing unit 2000 have been described in detail above in connection with fig. 2, and thus will not be further explained herein for the sake of brevity.
In an exemplary embodiment of the present invention, the data set of the historical operating data may be first divided into a training data set and a validation data set by the data set dividing unit 123 of the model creating unit 120. Then, the model training module 121 creates and trains a prediction model using the training data set, and obtains corresponding prediction model parameters, and then optimizes the corresponding prediction model parameters using the validation data set based on the created prediction model by the model optimization module 121 of the model creation unit 120, and determines the optimal prediction model parameters as the prediction model parameters of the created prediction model, to obtain the optimal prediction model.
The operations of the model creation unit 120 have been described in detail above in conjunction with fig. 3A and 3B, and thus will not be further explained herein for the sake of brevity.
Thereafter, in step 405, the state of the wind park may be predicted by the prediction unit 130 of the control device 100 based on the prediction model created in step 403, using the collected current operating data of the wind park, and the required control strategy may be determined based on the prediction result.
The process of prediction by the prediction unit 130 has been described in detail above in conjunction with fig. 3A and 3B, and thus will not be further explained here for the sake of brevity.
By applying the control device and the method for the wind generating set, the operating condition of the wind generating set can be monitored more accurately by acquiring the temperatures of a plurality of positions of the main shaft bearing of the wind generating set, the operating state of the wind generating set can be predicted more accurately, and a corresponding control strategy can be formulated, so that the temperature of the wind generating set can be controlled by power reduction operation and other modes before the shutdown temperature of the wind generating set is reached, various faults which may occur can be predicted effectively, the shutdown time is reduced effectively, the loss of the wind generating set is reduced, and the wind power generation benefit is improved.
The above-described methods and/or operations may be recorded, stored, or fixed in one or more computer-readable storage media that include program instructions to be executed by a computer to cause a processor to execute or perform the program instructions. The media may also include program instructions, data files, data structures, etc. alone or in combination with the program instructions. Examples of the computer readable storage medium include magnetic media (e.g., hard disks, floppy disks, and magnetic tape), optical media (e.g., CD ROM disks and DVDs), magneto-optical media (e.g., optical disks), and hardware devices (e.g., Read Only Memories (ROMs), Random Access Memories (RAMs), flash memories, etc.) specially configured to store and execute program instructions. Examples of program instructions include both machine code, such as produced by a compiler, and files containing higher level code that may be executed by the computer using an interpreter. The described hardware devices may be configured to act as one or more software elements to perform the operations and methods described above, and vice versa. Furthermore, the computer readable storage medium can be distributed over network coupled computer systems and the computer readable code or program instructions can be stored and executed in a distributed fashion.
While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope of the present invention as defined by the following claims.
Claims (18)
1. A control device for a wind energy plant, characterized in that it comprises:
a data obtaining unit configured to obtain historical operating data of the wind turbine generator set, wherein the historical operating data comprises a plurality of samples, and an attribute of each sample comprises: the system comprises a temperature attribute indicating the temperature of a main shaft bearing of the wind generating set, an output power attribute indicating the output power of the wind generating set, a temperature attribute indicating the ambient temperature of the wind generating set and a state attribute indicating the running state of the wind generating set;
a data preprocessing unit configured to preprocess the historical operating data acquired by the data acquisition unit, a model creation unit configured to create a prediction model using the preprocessed historical operating data; and
a prediction unit configured to predict an operational state of the wind park based on the created prediction model using the collected current operational data of the wind park and to determine a corresponding control strategy based on the prediction result,
wherein the data preprocessing unit includes: a data cleaning module configured to perform data cleaning processing on the acquired historical operating data to remove singular items in a sample of the historical operating data,
the data cleaning module is used for performing data cleaning processing on the acquired historical operating data through the following operations so as to remove singular items in the historical operating data sample: determining, for a particular temperature attribute among temperature attributes of samples in historical run data, whether a sample in the historical run data has a singular term and a data value for the particular temperature attribute of the sample based on a sample value and an expected value for the particular temperature attribute of the sample,
wherein if the number of the plurality of samples that are successively determined to have singular terms for the particular temperature attribute is greater than a first predetermined number, the data cleansing module is further configured to: determining a data value of a particular temperature property of a particular sample subsequent to the plurality of samples that are successively determined to have singular terms based on a sample value and an expected value of the particular temperature property of the particular sample.
2. The apparatus of claim 1, wherein the temperature attribute indicative of a temperature of a main shaft bearing of the wind power plant comprises at least one of the following temperature attributes: the temperature control system comprises a first temperature attribute indicating the temperature of a front end bearing inner ring of the wind generating set, a second temperature attribute indicating the temperature of a front end bearing outer ring of the wind generating set, a third temperature attribute indicating the temperature of a rear end bearing inner ring of the wind generating set, a fourth temperature attribute indicating the temperature of a rear end bearing outer ring of the wind generating set and a fifth temperature attribute indicating the temperature of grease on a main shaft bearing.
3. The apparatus of claim 1, wherein the data pre-processing unit further comprises:
the data serialization module is configured to serialize the historical operating data subjected to the data cleaning processing to obtain a corresponding matrix of M multiplied by N dimensions, wherein M represents the total number of samples of the historical operating data, and N represents the number of attributes of each sample of the historical operating data;
a data normalization module configured to normalize the data values associated with each attribute in the M N dimensional matrix.
4. The apparatus of claim 3, wherein the data cleansing module determines whether a sample in the historical run data has singular terms and data values for the particular temperature attribute of the sample based on the sample values and expected values for the particular temperature attribute of the sample by:
determining whether a difference between a sample value of the particular temperature attribute of a t-th sample in historical operating data and an expected value is greater than a first predetermined threshold;
determining that the sample value is a singular term and replacing the sample value with the expected value as the data value for the particular temperature attribute for the t-th sample if the difference is greater than a first predetermined threshold,
determining the sample value as a data value for the particular temperature attribute for the t-th sample if the difference is not greater than a first predetermined threshold,
wherein the expected value of the specific temperature property for the t-th sample is determined based on the data value of the specific temperature property for the (t-1) th sample and the data value of the specific temperature property for the (t-2) th sample, and 3 ≦ t ≦ M.
5. The apparatus of claim 4, wherein the data cleansing module is to determine the data value for the particular temperature attribute for a particular sample subsequent to the plurality of samples that are successively determined to have singular terms based on the sample value and an expected value for the particular temperature attribute for the particular sample by:
determining whether a difference between a sample value of the particular temperature property and a desired value of a particular sample subsequent to the plurality of samples successively determined to have singular terms is greater than a first predetermined threshold and less than a second predetermined threshold,
determining a sample value of the particular temperature property of the particular sample as a data value of the particular temperature property of the particular sample if a difference between the sample value of the particular temperature property of the particular sample and an expected value is greater than a first predetermined threshold and less than a second predetermined threshold,
determining the desired value of the particular temperature property of the particular sample as the data value of the particular temperature property of the particular sample if the difference between the sample value of the particular temperature property of the particular sample and the desired value is greater than or equal to a second predetermined threshold.
6. The apparatus of claim 5, wherein if the number of the plurality of samples that are consecutively determined to have singular terms for the particular temperature attribute is greater than a second predetermined number, the data cleansing module is further configured to:
selecting a new sample among the samples of the historical operating data as a starting sample of the data cleansing process, wherein the second predetermined number is greater than the first predetermined number.
7. The apparatus of claim 5, wherein if another sample meeting a predetermined condition is found after the particular sample for which the difference between the sample value of the particular temperature attribute and the expected value is greater than or equal to a second predetermined threshold, the data cleansing module is further configured to:
a new sample is selected among the samples of historical operating data as a starting sample for the data cleaning process,
wherein the predetermined condition is: a difference between a sample value of the particular temperature attribute of the other sample and a desired value is greater than a first predetermined threshold and less than a second predetermined threshold.
8. The apparatus according to any one of claims 1-7, wherein the model creation unit comprises:
a data set partitioning module configured to partition the preprocessed historical operating data into a training data set and a validation data set;
a model training module configured to create and train a predictive model using a training dataset and obtain corresponding predictive model parameters; and
and the model optimization module is configured to optimize the corresponding prediction model parameters by using the verification data set based on the created prediction model, and determine the optimal prediction model parameters as the prediction model parameters of the created prediction model to obtain the optimal prediction model.
9. A control method for a wind park, the method comprising:
obtaining historical operating data of a wind generating set, wherein the historical operating data comprises a plurality of samples, and attributes of each sample comprise: the system comprises a temperature attribute indicating the temperature of a main shaft bearing of the wind generating set, an output power attribute indicating the output power of the wind generating set, a temperature attribute indicating the ambient temperature of the wind generating set and a state attribute indicating the running state of the wind generating set;
preprocessing the acquired historical operating data;
creating a predictive model using the preprocessed historical operating data; and
predicting the operation state of the wind generating set by using the collected current operation data of the wind generating set based on the created prediction model, determining a corresponding control strategy based on the prediction result,
the method for preprocessing the acquired historical operating data comprises the following steps: performing data cleaning processing on the acquired historical operating data to remove singular items in a sample of the historical operating data,
the method comprises the following steps of carrying out data cleaning processing on the acquired historical operating data to remove singular items in a historical operating data sample: determining, for a particular temperature attribute among temperature attributes of samples in historical run data, whether a sample in the historical run data has a singular term and a data value for the particular temperature attribute of the sample based on a sample value and an expected value for the particular temperature attribute of the sample,
wherein if the number of the plurality of samples successively determined to have singular items is greater than a first predetermined number for the particular temperature attribute, the data value of the particular temperature attribute of a particular sample subsequent to the plurality of samples successively determined to have singular items is determined based on the sample value and the expected value of the particular temperature attribute of the particular sample.
10. The method of claim 9, wherein the temperature attribute indicative of the temperature of the main shaft bearing of the wind park comprises at least one of the following temperature attributes: the temperature control system comprises a first temperature attribute indicating the temperature of a front end bearing inner ring of the wind generating set, a second temperature attribute indicating the temperature of a front end bearing outer ring of the wind generating set, a third temperature attribute indicating the temperature of a rear end bearing inner ring of the wind generating set, a fourth temperature attribute indicating the temperature of a rear end bearing outer ring of the wind generating set and a fifth temperature attribute indicating the temperature of grease on a main shaft bearing.
11. The method of claim 9, wherein the step of pre-processing the acquired historical operating data further comprises:
serializing the historical operating data subjected to data cleaning processing to obtain a corresponding matrix of M multiplied by N dimensions, wherein M represents the total number of samples of the historical operating data, and N represents the number of attributes of each sample of the historical operating data;
normalizing the data values associated with each attribute in the M N dimensional matrix.
12. The method of claim 11, wherein determining whether a sample in historical operating data has singular terms and data values for a particular temperature attribute of the sample based on sample values and expected values for the particular temperature attribute of the sample comprises:
determining whether a difference between a sample value of the particular temperature attribute of a t-th sample in historical operating data and an expected value is greater than a first predetermined threshold;
determining that the sample value is a singular term and replacing the sample value with the expected value as the data value for the particular temperature attribute for the t-th sample if the difference is greater than a first predetermined threshold,
determining the sample value as a data value for the particular temperature attribute for the t-th sample if the difference is not greater than a first predetermined threshold,
wherein the expected value of the specific temperature property for the t-th sample is determined based on the data value of the specific temperature property for the (t-1) th sample and the data value of the specific temperature property for the (t-2) th sample, and 3 ≦ t ≦ M.
13. The method of claim 12, wherein determining the data value for a particular temperature attribute for a particular sample subsequent to the plurality of samples that are successively determined to have singular terms based on the sample value and expected value for the particular temperature attribute for the particular sample comprises: determining whether a difference between a sample value of the particular temperature property and a desired value of a particular sample subsequent to the plurality of samples successively determined to have singular terms is greater than a first predetermined threshold and less than a second predetermined threshold,
determining a sample value of the particular temperature property of the particular sample as a data value of the particular temperature property of the particular sample if a difference between the sample value of the particular temperature property of the particular sample and an expected value is greater than a first predetermined threshold and less than a second predetermined threshold,
determining the desired value of the particular temperature property of the particular sample as the data value of the particular temperature property of the particular sample if the difference between the sample value of the particular temperature property of the particular sample and the desired value is greater than or equal to a second predetermined threshold.
14. The method of claim 13, wherein the method further comprises:
selecting a new sample among the samples of the historical operating data as a starting sample of the data cleansing process if the number of the plurality of samples successively determined to have singular terms is greater than a second predetermined number for the specific temperature attribute,
wherein the second predetermined number is greater than the first predetermined number.
15. The method of claim 13, wherein the method further comprises:
if another sample satisfying a predetermined condition is found after the specific sample having a difference between the sample value of the specific temperature attribute and the expected value greater than or equal to a second predetermined threshold value, selecting a new sample among the samples of the historical operating data as a starting sample of the data washing process,
wherein the predetermined condition is: a difference between a sample value of the particular temperature attribute of the other sample and a desired value is greater than a first predetermined threshold and less than a second predetermined threshold.
16. The method according to any one of claims 9 to 15,
the step of creating a predictive model using the preprocessed historical operating data includes:
dividing the preprocessed historical operating data into a training data set and a verification data set;
establishing and training a prediction model by using a training data set, and acquiring corresponding prediction model parameters;
and optimizing the corresponding prediction model parameters by using a verification data set based on the created prediction model, and determining the optimal prediction model parameters as the prediction model parameters of the created prediction model to obtain the optimal prediction model.
17. A computer-readable storage medium storing a computer program, the computer program comprising instructions for performing the method of any one of claims 9-16.
18. A computer device comprising a readable medium having a computer program stored thereon, wherein the computer program comprises instructions for performing the method of any one of claims 9-16.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811012110.4A CN110874665B (en) | 2018-08-31 | 2018-08-31 | Control device and method for wind generating set |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811012110.4A CN110874665B (en) | 2018-08-31 | 2018-08-31 | Control device and method for wind generating set |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110874665A CN110874665A (en) | 2020-03-10 |
| CN110874665B true CN110874665B (en) | 2022-07-08 |
Family
ID=69715506
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811012110.4A Active CN110874665B (en) | 2018-08-31 | 2018-08-31 | Control device and method for wind generating set |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110874665B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111878320B (en) * | 2020-07-15 | 2021-07-30 | 上海电气风电集团股份有限公司 | Monitoring method and system of wind generating set and computer readable storage medium |
| CN113984111A (en) * | 2021-09-30 | 2022-01-28 | 北京华能新锐控制技术有限公司 | Wind turbine generator control method and device based on external environment change |
| CN113790766B (en) * | 2021-11-15 | 2022-04-01 | 江苏未来智慧信息科技有限公司 | Intelligent detection system and method for power plant |
| CN115933503B (en) * | 2023-03-10 | 2023-05-30 | 山东盛日电力集团有限公司 | Intelligent regulation control method and system for power generation equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2948679A1 (en) * | 2013-01-25 | 2015-12-02 | Vestas Wind Systems A/S | Control of wind turbines |
| CN106229972A (en) * | 2016-08-16 | 2016-12-14 | 北京国能日新系统控制技术有限公司 | A kind of integrated based on many meteorological sources and the wind power forecasting method of segmentation modeling |
| CN106844826A (en) * | 2016-12-02 | 2017-06-13 | 上海电机学院 | A kind of method for the diagnosis of gearbox of wind turbine failure predication |
-
2018
- 2018-08-31 CN CN201811012110.4A patent/CN110874665B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2948679A1 (en) * | 2013-01-25 | 2015-12-02 | Vestas Wind Systems A/S | Control of wind turbines |
| CN106229972A (en) * | 2016-08-16 | 2016-12-14 | 北京国能日新系统控制技术有限公司 | A kind of integrated based on many meteorological sources and the wind power forecasting method of segmentation modeling |
| CN106844826A (en) * | 2016-12-02 | 2017-06-13 | 上海电机学院 | A kind of method for the diagnosis of gearbox of wind turbine failure predication |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110874665A (en) | 2020-03-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110874665B (en) | Control device and method for wind generating set | |
| Chen et al. | Anomaly detection for wind turbines based on the reconstruction of condition parameters using stacked denoising autoencoders | |
| US11409347B2 (en) | Method, system and storage medium for predicting power load probability density based on deep learning | |
| Zhang et al. | Anomaly detection and diagnosis for wind turbines using long short-term memory-based stacked denoising autoencoders and XGBoost | |
| Marvuglia et al. | Monitoring of wind farms’ power curves using machine learning techniques | |
| CN112784965A (en) | Large-scale multi-element time series data abnormity detection method oriented to cloud environment | |
| US11840998B2 (en) | Hydraulic turbine cavitation acoustic signal identification method based on big data machine learning | |
| CN110685868A (en) | Wind turbine generator fault detection method and device based on improved gradient elevator | |
| Louen et al. | A new framework for remaining useful life estimation using support vector machine classifier | |
| CN112418277A (en) | Method, system, medium, and apparatus for predicting remaining life of rotating machine component | |
| CN111340282A (en) | DA-TCN-based method and system for estimating residual service life of equipment | |
| CN114048688A (en) | Method for predicting service life of bearing of wind power generator | |
| Zhang et al. | A hybrid forecasting system with complexity identification and improved optimization for short-term wind speed prediction | |
| CN116010900A (en) | Multi-scale feature fusion gearbox fault diagnosis method based on self-attention mechanism | |
| Zaman et al. | Wind speed forecasting using ARMA and neural network models | |
| Liu et al. | A hybrid LSSVR-HMM based prognostics approach | |
| JP2005242803A (en) | Performance estimator, performance estimating method, and performance estimating program of machine | |
| CN112465184A (en) | Cloud energy storage system control method of small-sample generation type counterstudy network | |
| Li et al. | Short-term load forecasting using support vector regression-based local predictor | |
| CN115936303A (en) | Transient voltage safety analysis method based on machine learning model | |
| CN111461416B (en) | Wind speed prediction method, system, electronic equipment and storage medium | |
| Sun et al. | Analysis of the health status of railway vehicle bearings based on improved HDP-HMM | |
| CN112183814A (en) | Short-term wind speed prediction method | |
| CN116722548B (en) | Photovoltaic power generation prediction method based on time sequence model and related equipment | |
| Cancelliere et al. | Neural networks for wind power generation forecasting: a case study |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |