CN111396248A - Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction - Google Patents
Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction Download PDFInfo
- Publication number
- CN111396248A CN111396248A CN202010181463.8A CN202010181463A CN111396248A CN 111396248 A CN111396248 A CN 111396248A CN 202010181463 A CN202010181463 A CN 202010181463A CN 111396248 A CN111396248 A CN 111396248A
- Authority
- CN
- China
- Prior art keywords
- wind direction
- data
- value
- model
- wind
- 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
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000012360 testing method Methods 0.000 claims description 46
- 238000012952 Resampling Methods 0.000 claims description 20
- 238000001914 filtration Methods 0.000 claims description 16
- 230000000737 periodic effect Effects 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims description 8
- 230000009471 action Effects 0.000 claims description 5
- 238000013500 data storage Methods 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000004321 preservation Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000005215 recombination Methods 0.000 claims description 2
- 230000006798 recombination Effects 0.000 claims description 2
- 238000010977 unit operation Methods 0.000 claims description 2
- 238000012706 support-vector machine Methods 0.000 abstract 3
- 230000008859 change Effects 0.000 description 3
- 238000011217 control strategy Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D7/00—Controlling wind motors
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D7/00—Controlling wind motors
- F03D7/02—Controlling wind motors the wind motors having rotation axis substantially parallel to the air flow entering the rotor
- F03D7/0204—Controlling wind motors the wind motors having rotation axis substantially parallel to the air flow entering the rotor for orientation in relation to wind direction
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D7/00—Controlling wind motors
- F03D7/02—Controlling wind motors the wind motors having rotation axis substantially parallel to the air flow entering the rotor
- F03D7/04—Automatic control; Regulation
- F03D7/042—Automatic control; Regulation by means of an electrical or electronic controller
- F03D7/043—Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic
- F03D7/045—Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic with model-based controls
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
- F03D—WIND MOTORS
- F03D7/00—Controlling wind motors
- F03D7/02—Controlling wind motors the wind motors having rotation axis substantially parallel to the air flow entering the rotor
- F03D7/04—Automatic control; Regulation
- F03D7/042—Automatic control; Regulation by means of an electrical or electronic controller
- F03D7/048—Automatic control; Regulation by means of an electrical or electronic controller controlling wind farms
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
- F05B2270/00—Control
- F05B2270/30—Control parameters, e.g. input parameters
- F05B2270/32—Wind speeds
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
- F05B2270/00—Control
- F05B2270/30—Control parameters, e.g. input parameters
- F05B2270/321—Wind directions
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
- F05B2270/00—Control
- F05B2270/30—Control parameters, e.g. input parameters
- F05B2270/324—Air pressure
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
- F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
- F05B2270/00—Control
- F05B2270/30—Control parameters, e.g. input parameters
- F05B2270/325—Air temperature
-
- 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
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Sustainable Development (AREA)
- Sustainable Energy (AREA)
- Chemical & Material Sciences (AREA)
- Combustion & Propulsion (AREA)
- Mechanical Engineering (AREA)
- General Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a wind turbine generator intelligent yaw control method based on short-term wind direction prediction, which comprises the following steps: 1) acquiring a data source; 2) preprocessing and converting data; 3) establishing a model characteristic database; 4) selecting model parameters; 5) training and storing the model; 6) loading an SVR model and predicting data; 7) generating a wind direction angle value; 8) and controlling yaw. According to the method, historical wind direction data in a certain time is utilized, a wind direction prediction model is established through statistics of a support vector machine (SVR), yaw control improvement is carried out according to predicted short-term wind directions, wind assembly precision can be effectively improved, and the number of ineffective yaw starting and stopping times can be reduced.
Description
Technical Field
The invention relates to the technical field of wind power generation, in particular to an intelligent yaw control method of a wind turbine generator based on short-term wind direction prediction.
Background
In the prior art, the yaw control strategy of the current wind turbine generator set mainly compares a measured yaw error with a set threshold value to determine whether to start a yaw action, and because the measured wind direction is greatly influenced by the rotation of an impeller and the real-time change of the wind direction is quick, the measured yaw deviation is filtered in practice and then serves as a yaw error reference value, so that the reference yaw error has certain time delay compared with the current actual yaw error, and the total time delay is more prominent due to the time delay of a yaw actuator. By adopting the control strategy, certain loss of wind power can be brought on one hand, and invalid yaw times (short yaw time) can be increased, yaw starting and stopping times are increased, and loss of a mechanical system is increased. Based on the problems, the current optimization schemes mainly include two types, one type is that the wind direction change is obtained in advance by installing laser radar wind measuring equipment and measuring the wind direction in front of an impeller, the scheme has high accuracy and is used as a preferred scheme, but because the laser radar wind measuring equipment has higher cost, the possibility of configuring each unit is lower; and the other type of advanced control mode carries out short-term wind direction prediction on a unit without the laser radar wind measuring equipment, and determines whether the yawing action is carried out or not by carrying out short-term prediction on the wind direction and referring to the predicted wind direction change on the basis of the existing control strategy.
Disclosure of Invention
The invention aims to overcome the defects and shortcomings of the prior art, and provides an intelligent yaw control method of a wind turbine generator based on short-term wind direction prediction.
In order to achieve the purpose, the technical scheme provided by the invention is as follows: a wind turbine generator intelligent yaw control method based on short-term wind direction prediction comprises the following steps:
1) obtaining a data source
Acquiring unit operation data from a data storage server of a wind power plant centralized control center, wherein the sampled tag point data comprises wind speed, wind direction, air temperature and air pressure;
2) data preprocessing and conversion
Processing periodic wind direction data, converting by trigonometric functions sin and cos to obtain corresponding scalar data, and performing moving average and resampling on all label point data to form a new sampling data sequence set;
3) building a model feature database
Designing two prediction models respectively aiming at the wind direction sin value and the wind direction cos value, wherein the two prediction models are respectively an SVR model 1 and an SVR model 2, and then carrying out data recombination on the data sequence set generated in the step 2) to respectively form a characteristic database of the SVR models 1 and 2, and the method specifically comprises the following steps:
for the SVR model 1, each single input sample comprises sin values of 20 wind directions and the latest wind speed, temperature and air pressure values, the output sample is the wind direction sin value at the current moment and is used as a predicted value, and a characteristic database of the SVR model 1 is formed through data set arrangement;
for the SVR model 2, each single input sample comprises cos values of 20 wind directions and the latest wind speed, temperature and air pressure values, the output sample is the wind direction cos value at the current moment and is used as a predicted value, and a characteristic database of the SVR model 2 is formed through data set arrangement;
finally, splitting the characteristic databases of the two SVR models to respectively form a training set and a test set of the SVR models 1 and 2;
4) model parameter selection
Determining parameters of two prediction models and a kernel function;
5) model training and preservation
Training the SVR models 1 and 2 respectively by using the training sets of the two models, and storing the models;
6) SVR model loading, data prediction
Respectively testing the SVR models 1 and 2 by using the test sets of the two models, and generating a predicted value corresponding to an input sample of each test set;
7) generating wind direction angle values
Performing inverse operation according to sin scalar and cos scalar of the predicted wind direction to generate a predicted wind direction angle value, and defining the range of the wind direction angle value to be 0-360 degrees;
8) yaw control
And performing yaw control by taking the predicted wind direction angle value as a reference signal.
In the step 1), the sampled data are derived from a historical database of a data storage server of the wind power plant centralized control center, and the data sampling interval is 1 s.
In the step 2), wind direction data is an angular variable with a period of 360 degrees, a wind direction signal needs to be converted into a non-periodic signal, specifically, a trigonometric function sin and cos are adopted to convert the wind direction, so that the periodic signal of 0-360 degrees is converted into a scalar signal between-1 and + 1; the sin component and the cos component are integrated, the value of the wind direction can be uniquely determined, and the wind direction can be determined to be in the quadrant IV through the positive and negative of the sin component and the cos component;
let the wind direction be θ, and x1 and x2 represent the sin component and the cos component of the wind direction, respectively, then:
x1＝sin(θ)
x2＝cos(θ)
after the wind direction signal processing is completed, forming new data tags including x1, x2, wind speed, air temperature and air pressure;
performing sliding average filtering on all the label point data, wherein the filtering time constant is 5s, then performing resampling by taking 5s as a period, and generating a new sampling data sequence set, namely input1, wherein the data comprise a value obtained by resampling x1 filtering and a value obtained by resampling x2 filtering, and a value obtained by resampling wind speed filtering, a value obtained by resampling air temperature filtering and a value obtained by resampling air pressure filtering, which are named as y1, y2, y3, y4 and y5 respectively;
in step 3), recombining the input1 generated in the previous step, and establishing a model feature database, specifically:
taking 23 data in 20 historical values (y1-1, y1-2 … … y1-20) before the current time in the sequence of y3, y4, y5 and y1 at the latest time as an input sample; recombining input1 data by taking the current time y1 as an output sample to form an input and output sample set of the SVR model 1; the prediction model of y1 can be defined as shown in the following formula (1), wherein y1(t) is a predicted value, i.e., a current value, y1(t-1), y1(t-2),.. multidot.y 1(t-20) is 20 historical values, and the prediction model is f;
y1(t)＝f(y1(t-1),y1(t-2),...,y1(t-20),y3(t-1),y4(t-1),y5(t-1)) (1)
taking 23 data in 20 historical values (y2-1, y2-2 … … y2-20) before the current time in the sequence of y3, y4, y5 and y2 at the latest time as an input sample; recombining input1 data by taking the current time y2 as an output sample to form an input and output sample set of the SVR model 2; the prediction model of y2 is defined as the following formula (2), wherein y2(t) is the predicted value, i.e. the current value, y2(t-1), y2(t-2),.. the y2(t-20) is 20 historical values, and the prediction model is g;
y2(t)＝g(y2(t-1),y2(t-2),...,y2(t-20),y3(t-1),y4(t-1),y5(t-1)) (2)
after the processes are completed, splitting the sample sets of the SVR models 1 and 2 respectively to form training sets y1_ train _ input, y1_ train _ output, and testing sets y1_ test _ input and y1_ test _ output of the SVR model 1 respectively; training sets y2_ train _ input, y2_ train _ output, and test sets y2_ test _ input and y2_ test _ output of the SVR model 2; and the test set accounts for 25% of the total number of model sample sets.
In step 4), RBF is selected as a kernel function of the SVR model, gamma is set to 0.01, and penalty factor C is set to 1000.
In step 5), the SVR model 1 is trained, the input sample set is y1_ train _ input, the output sample set is y1_ train _ output, and the model is saved; SVR model 2 is trained with an input sample set of y2_ train _ input and an output sample set of y2_ train _ output, and the model is saved.
In step 6), testing the SVR model 1, wherein an input sample set is y1_ test _ input, an output sample set is y1_ test _ output, and each input sample generates a predicted value y 1' of y1, wherein y1 represents the value of the sin component filtered resample of the wind direction; the SVR model 2 was tested with an input sample set of y2_ test _ input and an output sample set of y2_ test _ output, each of which produced a predicted value y 2' of y2, where y2 represents the cos component filtered resample value of the wind direction.
In step 7), performing trigonometric function inverse operation according to the two predicted values y1 'and y 2' generated in step 6) to obtain predicted wind direction angle values
The calculation is shown in the following formula (3).
Wherein sign (×) is a sign function, y1 'represents a predicted value of a wind direction sin component, and y 2' represents a predicted value of a wind direction cos component;
and after the wind speed prediction of the test set is completed, wind direction angle error analysis is carried out, and the specified angle error is in the range of 0-360 degrees.
In step 8), replacing the measurement historical value adopted by the conventional control with the predicted wind direction angle value, then carrying out average value processing on the predicted wind direction angle value for 10s, 30s and 60s respectively, correspondingly setting three different thresholds A1, A2 and A3 aiming at the three average values, and triggering yaw action when any one wind direction average value exceeds the corresponding threshold.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the method converts the wind direction periodic angle value into the aperiodic scalar value, avoids the defect of a general data prediction algorithm in processing periodic signals, and greatly improves the prediction precision.
2. According to the method, on the basis of the wind direction sequence value, factors influencing the relationship such as wind speed, temperature and air pressure are increased, more data characteristics are utilized, and the accuracy of the model can be improved.
3. The yaw control related to the method of the invention refers to the predicted wind direction angle value, and does not measure the wind direction historical value to carry out yaw logic design, thereby reducing the total delay time of the yaw system, improving the wind alignment precision, reducing the loss of generated energy and simultaneously reducing the times of transient start-stop of the yaw system.
Drawings
FIG. 1 is a logic flow diagram of the present invention.
Detailed Description
The present invention will be further described with reference to the following specific examples.
As shown in fig. 1, the method for intelligently controlling yaw of a wind turbine generator based on short-term wind direction prediction provided in this embodiment has the following specific conditions:
1) obtaining a data source
The sampled data is from a historical database of a data storage server of the wind power plant centralized control center, the data sampling interval is 1s, and the sampled label point data comprises wind speed, wind direction, air temperature and air pressure.
2) Data preprocessing and conversion
Wind direction data is an angular variable with a period of 360 degrees, and a general prediction algorithm has insufficient effect on processing periodic signals such as wind direction signals, so that the wind direction signals need to be converted into non-periodic signals firstly, and the wind direction is converted by adopting trigonometric functions sin and cos in the invention, so that the periodic signals of 0-360 degrees are converted into scalar signals between-1 and + 1. By combining sin and cos components, the value of the wind direction can be uniquely determined, for example, when sin is negative and cos is positive, the wind direction can be determined to be in the second quadrant.
Let the wind direction be θ, and x1 and x2 represent the sin component and the cos component of the wind direction, respectively, then:
x1＝sin(θ)
x2＝cos(θ)
after the wind direction signal processing is completed, new data tags including x1, x2, wind speed, air temperature and air pressure are formed.
And performing sliding average filtering on all the label point data, wherein the filtering time constant is 5s, then performing resampling by taking 5s as a period, and generating a new sampling data sequence set, namely input1, wherein the included data comprise a value obtained by resampling an x1 filter and a value obtained by resampling an x2 filter, and a value obtained by resampling a wind speed filter, a value obtained by resampling an air temperature filter and a value obtained by resampling an air pressure filter, which are named as y1, y2, y3, y4 and y5 respectively.
3) Building a model feature database
Recombining the input1 generated in the previous step, and establishing a model feature database, which specifically comprises the following steps:
23 data in 20 historical values (y1-1, y1-2 … … y1-20) before the current time are taken as an input sample in the sequence of y3, y4, y5 and y1 at the latest time. And recombining the input1 data by taking the current time y1 as an output sample to form an input and output sample set of the SVR model 1. The prediction model of y1 can be defined as the following formula (1), where y1(t) is the predicted value, i.e., the current value, y1(t-1), y1(t-2),.. the y1(t-20) is 20 historical values, and the prediction model is f.
y1(t)＝f(y1(t-1),y1(t-2),...,y1(t-20),y3(t-1),y4(t-1),y5(t-1)) (1)
23 data in 20 historical values (y2-1, y2-2 … … y2-20) before the current time are taken as an input sample in the sequence of y3, y4, y5 and y2 at the latest time. And recombining the input1 data by taking the current time y2 as an output sample to form an input and output sample set of the SVR model 2. The prediction model of y2 is defined as the following formula (2), where y2(t) is the predicted value, i.e., the current value, y2(t-1), y2(t-2),.. the y2(t-20) is 20 historical values, and the prediction model is g.
y2(t)＝g(y2(t-1),y2(t-2),...,y2(t-20),y3(t-1),y4(t-1),y5(t-1)) (2)
After the above processes are completed, splitting the sample sets of the SVR models 1 and 2 respectively to form training sets y1_ train _ input, y1_ train _ output, and test sets y1_ test _ input and y1_ test _ output of the SVR model 1 respectively; the training set y2_ train _ input, y2_ train _ output, and the test set y2_ test _ input, y2_ test _ output of the SVR model 2. And the test set accounts for 25% of the total number of model sample sets.
4) Model parameter selection
Determining SVR model parameters, selecting RBF as a kernel function of the SVR, setting gamma to be 0.01 and setting a penalty factor C to be 1000.
5) Model training and preservation
The SVR model 1 is trained with an input sample set of y1_ train _ input and an output sample set of y1_ train _ output, and the model is saved.
SVR model 2 is trained with an input sample set of y2_ train _ input and an output sample set of y2_ train _ output, and the model is saved.
6) SVR model loading, data prediction
The SVR model 1 was tested with an input sample set of y1_ test _ input and an output sample set of y1_ test _ output, each of which produced a predicted value y 1' of y 1.
The SVR model 2 was tested with an input sample set of y2_ test _ input and an output sample set of y2_ test _ output, each input sample yielding a predicted value y 2' of y 2.
7) Generating wind direction angle values
According to the two predicted values generated in the step 6), performing trigonometric function inverse operation to obtain a predicted wind direction angle value
The operation formula is shown in the following formula (3).
Wherein sign (#) is a sign taking function.
And after the test set wind speed prediction is completed, wind direction angle error analysis is carried out, the specified angle error is in the range of 0-360 degrees, and other treatment can be carried out according to the empirical range of the error.
8) Yaw control
The method adopts the predicted wind direction angle value to improve the yaw control, and specifically comprises the following steps: the measured wind direction historical value adopted by the conventional control is replaced by the predicted wind direction angle value, then the predicted wind direction angle value is respectively subjected to average value processing of 10s, 30s and 60s, three different threshold values A1, A2 and A3 are correspondingly set for the three average values, and any one wind direction average value exceeds the corresponding threshold value, so that the yaw action can be triggered.
The above-mentioned embodiments are merely preferred embodiments of the present invention, and the scope of the present invention is not limited thereto, so that the changes in the shape and principle of the present invention should be covered within the protection scope of the present invention.
Claims (8)
1. A wind turbine generator intelligent yaw control method based on short-term wind direction prediction is characterized by comprising the following steps:
1) obtaining a data source
Acquiring unit operation data from a data storage server of a wind power plant centralized control center, wherein the sampled tag point data comprises wind speed, wind direction, air temperature and air pressure;
2) data preprocessing and conversion
Processing periodic wind direction data, converting by trigonometric functions sin and cos to obtain corresponding scalar data, and performing moving average and resampling on all label point data to form a new sampling data sequence set;
3) building a model feature database
Designing two prediction models respectively aiming at the wind direction sin value and the wind direction cos value, wherein the two prediction models are respectively an SVR model 1 and an SVR model 2, and then carrying out data recombination on the data sequence set generated in the step 2) to respectively form a characteristic database of the SVR models 1 and 2, and the method specifically comprises the following steps:
for the SVR model 1, each single input sample comprises sin values of 20 wind directions and the latest wind speed, temperature and air pressure values, the output sample is the wind direction sin value at the current moment and is used as a predicted value, and a characteristic database of the SVR model 1 is formed through data set arrangement;
for the SVR model 2, each single input sample comprises cos values of 20 wind directions and the latest wind speed, temperature and air pressure values, the output sample is the wind direction cos value at the current moment and is used as a predicted value, and a characteristic database of the SVR model 2 is formed through data set arrangement;
finally, splitting the characteristic databases of the two SVR models to respectively form a training set and a test set of the SVR models 1 and 2;
4) model parameter selection
Determining parameters of two prediction models and a kernel function;
5) model training and preservation
Training the SVR models 1 and 2 respectively by using the training sets of the two models, and storing the models;
6) SVR model loading, data prediction
Respectively testing the SVR models 1 and 2 by using the test sets of the two models, and generating a predicted value corresponding to an input sample of each test set;
7) generating wind direction angle values
Performing inverse operation according to sin scalar and cos scalar of the predicted wind direction to generate a predicted wind direction angle value, and defining the range of the wind direction angle value to be 0-360 degrees;
8) yaw control
And performing yaw control by taking the predicted wind direction angle value as a reference signal.
2. The intelligent yaw control method of the wind turbine generator based on short-term wind direction prediction as claimed in claim 1, wherein: in the step 1), the sampled data are derived from a historical database of a data storage server of the wind power plant centralized control center, and the data sampling interval is 1 s.
3. The intelligent yaw control method of the wind turbine generator based on short-term wind direction prediction as claimed in claim 1, wherein: in the step 2), wind direction data is an angular variable with a period of 360 degrees, a wind direction signal needs to be converted into a non-periodic signal, specifically, a trigonometric function sin and cos are adopted to convert the wind direction, so that the periodic signal of 0-360 degrees is converted into a scalar signal between-1 and + 1; the sin component and the cos component are integrated, the value of the wind direction can be uniquely determined, and the wind direction can be determined to be in the quadrant IV through the positive and negative of the sin component and the cos component;
let the wind direction be θ, and x1 and x2 represent the sin component and the cos component of the wind direction, respectively, then:
x1＝sin(θ)
x2＝cos(θ)
after the wind direction signal processing is completed, forming new data tags including x1, x2, wind speed, air temperature and air pressure;
performing sliding average filtering on all the label point data, wherein the filtering time constant is 5s, then performing resampling by taking 5s as a period, and generating a new sampling data sequence set, namely input1, wherein the data comprise a value obtained by resampling x1 filtering and a value obtained by resampling x2 filtering, and a value obtained by resampling wind speed filtering, a value obtained by resampling air temperature filtering and a value obtained by resampling air pressure filtering, which are named as y1, y2, y3, y4 and y5 respectively;
in step 3), recombining the input1 generated in the previous step, and establishing a model feature database, specifically:
taking 23 data in 20 historical values (y1-1, y1-2 … … y1-20) before the current time in the sequence of y3, y4, y5 and y1 at the latest time as an input sample; recombining input1 data by taking the current time y1 as an output sample to form an input and output sample set of the SVR model 1; the prediction model of y1 can be defined as shown in the following formula (1), wherein y1(t) is a predicted value, i.e., a current value, y1(t-1), y1(t-2),.. multidot.y 1(t-20) is 20 historical values, and the prediction model is f;
y1(t)＝f(y1(t-1),y1(t-2),...,y1(t-20),y3(t-1),y4(t-1),y5(t-1)) (1)
taking 23 data in 20 historical values (y2-1, y2-2 … … y2-20) before the current time in the sequence of y3, y4, y5 and y2 at the latest time as an input sample; recombining input1 data by taking the current time y2 as an output sample to form an input and output sample set of the SVR model 2; the prediction model of y2 is defined as the following formula (2), wherein y2(t) is the predicted value, i.e. the current value, y2(t-1), y2(t-2),.. the y2(t-20) is 20 historical values, and the prediction model is g;
y2(t)＝g(y2(t-1),y2(t-2),...,y2(t-20),y3(t-1),y4(t-1),y5(t-1)) (2)
after the processes are completed, splitting the sample sets of the SVR models 1 and 2 respectively to form training sets y1_ train _ input, y1_ train _ output, and testing sets y1_ test _ input and y1_ test _ output of the SVR model 1 respectively; training sets y2_ train _ input, y2_ train _ output, and test sets y2_ test _ input and y2_ test _ output of the SVR model 2; and the test set accounts for 25% of the total number of model sample sets.
4. The intelligent yaw control method of the wind turbine generator based on short-term wind direction prediction as claimed in claim 1, wherein: in step 4), RBF is selected as a kernel function of the SVR model, gamma is set to 0.01, and penalty factor C is set to 1000.
5. The intelligent yaw control method of the wind turbine generator based on short-term wind direction prediction as claimed in claim 1, wherein: in step 5), the SVR model 1 is trained, the input sample set is y1_ train _ input, the output sample set is y1_ train _ output, and the model is saved; SVR model 2 is trained with an input sample set of y2_ train _ input and an output sample set of y2_ train _ output, and the model is saved.
6. The intelligent yaw control method of the wind turbine generator based on short-term wind direction prediction as claimed in claim 1, wherein: in step 6), testing the SVR model 1, wherein an input sample set is y1_ test _ input, an output sample set is y1_ test _ output, and each input sample generates a predicted value y 1' of y1, wherein y1 represents the value of the sin component filtered resample of the wind direction; the SVR model 2 was tested with an input sample set of y2_ test _ input and an output sample set of y2_ test _ output, each of which produced a predicted value y 2' of y2, where y2 represents the cos component filtered resample value of the wind direction.
7. The intelligent yaw control method of the wind turbine generator based on short-term wind direction prediction as claimed in claim 1, wherein: in step 7), performing trigonometric function inverse operation according to the two predicted values y1 'and y 2' generated in step 6) to obtain a predicted wind direction angle value
The calculation is shown in the following formula (3).
Wherein sign (×) is a sign function, y1 'represents a predicted value of a wind direction sin component, and y 2' represents a predicted value of a wind direction cos component;
and after the test set wind speed prediction is finished, analyzing the wind direction angle prediction error, and setting the angle error within the range of 0-360 degrees.
8. The intelligent yaw control method of the wind turbine generator based on short-term wind direction prediction as claimed in claim 1, wherein: in step 8), replacing the measured wind direction historical value adopted by the conventional control with the predicted wind direction angle value, then carrying out average value processing on the predicted wind direction angle value for 10s, 30s and 60s respectively, correspondingly setting three different threshold values A1, A2 and A3 aiming at the three average values, and triggering yaw action when any one wind direction average value exceeds the corresponding threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010181463.8A CN111396248A (en) | 2020-03-16 | 2020-03-16 | Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010181463.8A CN111396248A (en) | 2020-03-16 | 2020-03-16 | Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111396248A true CN111396248A (en) | 2020-07-10 |
Family
ID=71428821
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010181463.8A Pending CN111396248A (en) | 2020-03-16 | 2020-03-16 | Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111396248A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111984905A (en) * | 2020-07-17 | 2020-11-24 | 明阳智慧能源集团股份公司 | Wind turbine generator wind direction data filtering method based on fitting technology |
| CN112502899A (en) * | 2020-11-30 | 2021-03-16 | 东方电气风电有限公司 | Consumption reduction method for wind generating set |
| CN113482853A (en) * | 2021-08-06 | 2021-10-08 | 贵州大学 | Yaw control method, system, electronic equipment and storage medium |
| CN114263565A (en) * | 2020-09-16 | 2022-04-01 | 新疆金风科技股份有限公司 | Yaw control equipment and method of wind generating set |
| CN115600639A (en) * | 2022-09-30 | 2023-01-13 | 国网四川省电力公司眉山供电公司（Cn） | Wind speed sensor, power transmission line wind speed prediction method and early warning system |
| CN117287341A (en) * | 2023-07-11 | 2023-12-26 | 华能新能源股份有限公司山西分公司 | Yaw prediction system and method for wind turbine generator |
| CN117989054A (en) * | 2024-04-03 | 2024-05-07 | 东方电气风电股份有限公司 | Domestic fan intelligent control method, system and equipment |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007233639A (en) * | 2006-02-28 | 2007-09-13 | Tohoku Electric Power Co Inc | Wind power generation output prediction method, wind power generation output prediction unit, and program |
| CN103065202A (en) * | 2012-12-24 | 2013-04-24 | 电子科技大学 | Wind power plant ultrashort term wind speed prediction method based on combination kernel function |
| CN103616734A (en) * | 2013-12-11 | 2014-03-05 | 山东大学 | System and method for large-range synchronous real-time meteorological data measurement and wind speed and direction prediction |
| CN104196680A (en) * | 2014-09-05 | 2014-12-10 | 南京达沙信息科技有限公司 | Draught fan foreseeable yaw control system based on imminent prediction |
| CN106337778A (en) * | 2016-10-31 | 2017-01-18 | 湘电风能有限公司 | Control method for pre-start of wind generating set |
| CN108537372A (en) * | 2018-03-27 | 2018-09-14 | 中南大学 | A kind of Yaw control method of wind direction prediction technique and wind power generating set |
| CN108825432A (en) * | 2018-06-22 | 2018-11-16 | 北京金风科创风电设备有限公司 | Yaw control method and device, and computer readable storage medium |
| CN108843497A (en) * | 2018-06-29 | 2018-11-20 | 北京金风科创风电设备有限公司 | Yaw control method and equipment of wind generating set |
| KR101956717B1 (en) * | 2017-09-01 | 2019-03-11 | 군산대학교산학협력단 | Method and apparatus for prediction of wind direction, and method for yaw control of wind turbines using the same |
| CN110210641A (en) * | 2018-02-28 | 2019-09-06 | 北京金风科创风电设备有限公司 | Wind direction prediction method and device for wind power plant |
-
2020
- 2020-03-16 CN CN202010181463.8A patent/CN111396248A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007233639A (en) * | 2006-02-28 | 2007-09-13 | Tohoku Electric Power Co Inc | Wind power generation output prediction method, wind power generation output prediction unit, and program |
| CN103065202A (en) * | 2012-12-24 | 2013-04-24 | 电子科技大学 | Wind power plant ultrashort term wind speed prediction method based on combination kernel function |
| CN103616734A (en) * | 2013-12-11 | 2014-03-05 | 山东大学 | System and method for large-range synchronous real-time meteorological data measurement and wind speed and direction prediction |
| CN104196680A (en) * | 2014-09-05 | 2014-12-10 | 南京达沙信息科技有限公司 | Draught fan foreseeable yaw control system based on imminent prediction |
| CN106337778A (en) * | 2016-10-31 | 2017-01-18 | 湘电风能有限公司 | Control method for pre-start of wind generating set |
| KR101956717B1 (en) * | 2017-09-01 | 2019-03-11 | 군산대학교산학협력단 | Method and apparatus for prediction of wind direction, and method for yaw control of wind turbines using the same |
| CN110210641A (en) * | 2018-02-28 | 2019-09-06 | 北京金风科创风电设备有限公司 | Wind direction prediction method and device for wind power plant |
| CN108537372A (en) * | 2018-03-27 | 2018-09-14 | 中南大学 | A kind of Yaw control method of wind direction prediction technique and wind power generating set |
| CN108825432A (en) * | 2018-06-22 | 2018-11-16 | 北京金风科创风电设备有限公司 | Yaw control method and device, and computer readable storage medium |
| CN108843497A (en) * | 2018-06-29 | 2018-11-20 | 北京金风科创风电设备有限公司 | Yaw control method and equipment of wind generating set |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111984905A (en) * | 2020-07-17 | 2020-11-24 | 明阳智慧能源集团股份公司 | Wind turbine generator wind direction data filtering method based on fitting technology |
| CN111984905B (en) * | 2020-07-17 | 2023-11-28 | 明阳智慧能源集团股份公司 | Wind turbine generator wind direction data filtering method based on fitting technology |
| CN114263565B (en) * | 2020-09-16 | 2024-04-12 | 金风科技股份有限公司 | Yaw control equipment and method of wind generating set |
| CN114263565A (en) * | 2020-09-16 | 2022-04-01 | 新疆金风科技股份有限公司 | Yaw control equipment and method of wind generating set |
| CN112502899A (en) * | 2020-11-30 | 2021-03-16 | 东方电气风电有限公司 | Consumption reduction method for wind generating set |
| CN112502899B (en) * | 2020-11-30 | 2021-11-16 | 东方电气风电有限公司 | Consumption reduction method for wind generating set |
| CN113482853B (en) * | 2021-08-06 | 2023-02-24 | 贵州大学 | Yaw control method, system, electronic equipment and storage medium |
| CN113482853A (en) * | 2021-08-06 | 2021-10-08 | 贵州大学 | Yaw control method, system, electronic equipment and storage medium |
| CN115600639B (en) * | 2022-09-30 | 2023-11-14 | 国网四川省电力公司眉山供电公司 | Wind speed sensor, wind speed prediction method of power transmission line and early warning system |
| CN115600639A (en) * | 2022-09-30 | 2023-01-13 | 国网四川省电力公司眉山供电公司（Cn） | Wind speed sensor, power transmission line wind speed prediction method and early warning system |
| CN117287341A (en) * | 2023-07-11 | 2023-12-26 | 华能新能源股份有限公司山西分公司 | Yaw prediction system and method for wind turbine generator |
| CN117989054A (en) * | 2024-04-03 | 2024-05-07 | 东方电气风电股份有限公司 | Domestic fan intelligent control method, system and equipment |
| CN117989054B (en) * | 2024-04-03 | 2024-06-07 | 东方电气风电股份有限公司 | Domestic fan intelligent control method, system and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111396248A (en) | Wind turbine generator set intelligent yaw control method based on short-term wind direction prediction | |
| CN108334672B (en) | Maximum wind energy capturing method of variable-speed wind generating set based on effective wind speed estimation | |
| CN106815771B (en) | Long-term assessment method for wind farm load | |
| CN106979126B (en) | Wind power generating set high wind speed section effective wind speed estimation method based on SVR | |
| EP2869144A1 (en) | Simulation testing platform for wind power plant and testing method thereof | |
| CN102184453A (en) | Wind power combination predicting method based on fuzzy neural network and support vector machine | |
| CN104317283A (en) | Hardware-in-the-loop test platform and test method for wind power plant control system | |
| CN108869174B (en) | Nonlinear modeling wind driven generator blade natural frequency working condition compensation method | |
| CN110365053B (en) | Short-term wind power prediction method based on delay optimization strategy | |
| US20210054825A1 (en) | Wind turbine control method and device, controller, and control system | |
| CN113822418A (en) | Wind power plant power prediction method, system, device and storage medium | |
| CN106529706A (en) | Support-vector-machine-regression-based method for predicting wind speed of wind power plant | |
| EP3771822A1 (en) | A method for computer-implemented determination of a vertical speed wind profile of a wind field | |
| CN108547736A (en) | The Yaw control method of wind speed and direction prediction technique and wind power generating set | |
| CN113612237A (en) | Method for positioning resonance-induced subsynchronous oscillation source in offshore wind farm | |
| CN115221785A (en) | Neural network prediction model construction method, prediction method and system | |
| CN111878322A (en) | Wind power generator device | |
| CN118481905A (en) | Adjusting method of variable pitch system of wind driven generator | |
| CN207038185U (en) | Pressurized-water reactor nuclear power plant core power control device based on Takagi Sugeno fuzzy controls | |
| CN117494618A (en) | Wind farm intelligent image monitoring system based on real-time network | |
| CN113048012A (en) | Wind turbine generator yaw angle identification method and device based on Gaussian mixture model | |
| van der Walt et al. | A comparison of regression algorithms for wind speed forecasting at Alexander Bay | |
| CN115898787A (en) | Method and device for dynamically identifying static yaw error of wind turbine generator | |
| Yang et al. | Non-linear autoregressive neural network based wind direction prediction for the wind turbine yaw system | |
| CN114298512A (en) | Wind turbine generator set power characteristic analysis method |
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 |
Application publication date: 20200710 |
|
| RJ01 | Rejection of invention patent application after publication |