WO2018014825A1 - 一种风电机组对风误差自动校准方法及装置 - Google Patents

一种风电机组对风误差自动校准方法及装置 Download PDF

Info

Publication number
WO2018014825A1
WO2018014825A1 PCT/CN2017/093303 CN2017093303W WO2018014825A1 WO 2018014825 A1 WO2018014825 A1 WO 2018014825A1 CN 2017093303 W CN2017093303 W CN 2017093303W WO 2018014825 A1 WO2018014825 A1 WO 2018014825A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
target
wind
module
correction
Prior art date
Application number
PCT/CN2017/093303
Other languages
English (en)
French (fr)
Inventor
叶杭冶
潘东浩
王欣
吴根勇
应有
Original Assignee
浙江运达风电股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 浙江运达风电股份有限公司 filed Critical 浙江运达风电股份有限公司
Priority to DK17830459.8T priority Critical patent/DK3489508T3/da
Priority to EP17830459.8A priority patent/EP3489508B1/en
Priority to ES17830459T priority patent/ES2926639T3/es
Publication of WO2018014825A1 publication Critical patent/WO2018014825A1/zh

Links

Images

Classifications

    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D17/00Monitoring or testing of wind motors, e.g. diagnostics
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D7/00Controlling wind motors 
    • F03D7/02Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor
    • F03D7/0204Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor for orientation in relation to wind direction
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03DWIND MOTORS
    • F03D7/00Controlling wind motors 
    • F03D7/02Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor
    • F03D7/04Automatic control; Regulation
    • F03D7/042Automatic control; Regulation by means of an electrical or electronic controller
    • F03D7/043Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05BINDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00Control
    • F05B2270/30Control parameters, e.g. input parameters
    • F05B2270/321Wind directions
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05BINDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00Control
    • F05B2270/30Control parameters, e.g. input parameters
    • F05B2270/329Azimuth or yaw angle
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05BINDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00Control
    • F05B2270/30Control parameters, e.g. input parameters
    • F05B2270/335Output power or torque
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05BINDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00Control
    • F05B2270/80Devices generating input signals, e.g. transducers, sensors, cameras or strain gauges
    • F05B2270/802Calibration thereof
    • YGENERAL 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
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02EREDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00Energy generation through renewable energy sources
    • Y02E10/70Wind energy
    • Y02E10/72Wind turbines with rotation axis in wind direction

Definitions

  • the present invention provides a method and apparatus for automatically calibrating a wind turbine against wind error, which solves the problem in the prior art that the absolute zero of the wind vane is not parallel to the centerline of the nacelle for a number of reasons, so that the yaw system exists. “Inherent wind error”, which leads to the problem that the output power of the unit does not meet the design requirements.
  • the technical solutions are as follows:
  • a wind turbine automatic calibration method for wind error comprising:
  • a zero parameter in the yaw system of the target unit is corrected by the inherent wind error.
  • the extracting the abnormal data from the historical running data of the target unit includes:
  • the performing clustering processing on the historical running data includes:
  • the performing the dimension reduction processing on the target data, and determining the relationship between the wind error and the output performance by using the data after the dimension reduction processing including:
  • the wind power curve is processed into a relationship between wind error and output performance.
  • the method further includes:
  • the change in the output performance of the target unit before and after the correction is evaluated based on the operational data deep purification technique.
  • the evaluation of the output performance of the target unit before and after the correction based on the operational data deep purification technology comprises:
  • the abnormal data is removed from the two sets of pre-correction operation data, and the two sets of pre-correction operation data after the abnormal data is eliminated are combined to obtain the pre-correction target operation data, and the abnormal data is removed from the two sets of corrected operation data, and the abnormal data is removed.
  • the two sets of corrected operational data are combined to obtain corrected target operational data;
  • the wind power curve is linearly interpolated for the nodes to obtain the theoretical active power corresponding to each wind speed
  • the data points whose probability density is not within the preset range are removed, and then returning to the rated wind speed to divide the pre-correction operation data into two groups, and the corrected operation is performed with the rated wind speed as a boundary.
  • the data is divided into two groups, until the wind power curve obtained by the fitting no longer changes;
  • the wind power curve before and after the correction is converted into the quantized value of the output performance index before and after the correction by separately integrating the wind power curves before and after the correction;
  • a wind turbine automatic wind calibration device comprising:
  • a data acquisition module configured to acquire historical operation data of the target unit within a preset time period
  • a data culling module configured to remove abnormal data from the historical running data acquired by the data acquiring module, and obtain target data
  • a dimension reduction processing module configured to perform dimension reduction processing on the target data obtained after the data culling module rejects the abnormal data
  • a curve determining module configured to determine a relationship between wind error and output performance by using the dimension reduction processing module of the dimensionality reduction processing module
  • An inherent wind error determining module configured to determine an inherent wind error by a relationship between the wind error and the output performance determined by the curve determining module;
  • a correction module configured to correct a zero parameter in the yaw system of the target unit by the inherent wind error determined by the inherent wind error determination module.
  • the data culling module includes:
  • a clustering module configured to perform clustering processing on the historical running data to obtain a clustering result
  • a culling module for culling the abnormal data based on the clustering result of the clustering process.
  • the dimension reduction processing module includes:
  • the probability statistics sub-module is configured to perform probability statistics on the wind error in the target data, and determine a target wind error range based on the probability statistics result;
  • the data culling sub-module is configured to remove data from the target data that is not within the target wind error range, and obtain first target data;
  • the curve determining module includes: a first curve determining submodule and a second curve determining submodule;
  • the first curve determining submodule is configured to determine, by using the first target data, a wind power curve corresponding to each pair of wind errors, where the wind power curve is a relationship between wind speed and active power;
  • the second curve determining sub-module is configured to process the wind power curve into a relationship between wind error and output performance.
  • the device also includes:
  • An evaluation module for evaluating the target unit before and after correction based on the operational data deep purification technique The change in performance of the output;
  • the evaluation module includes:
  • a data acquisition sub-module configured to obtain pre-correction operation data of the target unit and corrected operation data
  • a data grouping module block configured to divide the pre-correction operation data and the corrected operation data into two groups respectively according to a rated wind speed
  • the abnormal data culling sub-module is configured to remove the abnormal data from the two sets of pre-corrected running data, and remove the abnormal data from the two sets of corrected running data;
  • the data merging sub-module is configured to combine the two sets of pre-correction operation data after the abnormal data is removed to obtain the pre-correction target operation data, and combine the two sets of corrected operation data after the abnormal data is removed to obtain the corrected target operation data;
  • a data fitting submodule configured to respectively fit the pre-correction target operational data and the corrected target operational data to the wind power curve
  • a data interpolation sub-module configured to linearly interpolate the wind speed in the pre-correction target operation data and the corrected target operation data, respectively, to obtain a theoretical active power corresponding to each wind speed
  • a power deviation calculation sub-module configured to separately calculate a power deviation between the actual active power and the theoretical active power of each data point in the pre-correction target operation data and the corrected target operation data, and obtain corresponding to each data point Power deviation
  • the abnormal data culling sub-module is configured to perform probability statistics on the power deviation corresponding to each data point, and remove data points whose probability density is not within the preset range according to the probability statistical result, to obtain new operational data before correction and New corrected operational data, then triggering the data packet module block to divide the pre-correction operational data into two groups at a nominal wind speed, until the fitted wind power curve no longer changes;
  • a wind power curve processing sub-module configured to respectively perform integral processing on the wind power curve before and after the correction, and convert the wind power curve before and after the correction into a quantized value of the output performance index before and after the correction;
  • the comparison sub-module is used to compare the magnitude of the quantitative value of the output performance index before and after the correction, and obtain the change of the output performance of the target unit before and after the correction.
  • the wind turbine automatic wind calibration method and device fully utilizes the historical operation data of the wind turbine, and automatically identifies the inherent wind of the yaw system of the wind turbine by data mining and analysis of the historical operation data of the wind turbine.
  • the error in turn, can automatically adjust the zero parameter in the yaw control system according to the inherent wind error, which enhances the self-adaptive capability of the yaw system and improves the actual output performance of the unit, so that the output power of the unit can be achieved. Design requirements.
  • Figure 1 is a schematic diagram showing the relationship between the actual wind direction and the position of the nacelle when there is an inherent wind error
  • FIG. 2 is a schematic flow chart of a method for automatically adjusting a wind error of a wind turbine according to an embodiment of the present invention
  • FIG. 3 is a schematic flow chart of another method for automatically adjusting a wind error of a wind turbine according to an embodiment of the present invention
  • FIG. 4 is a schematic flowchart of an implementation process of clustering processing historical operation data in a method for automatically adjusting a wind error of a wind turbine set according to an embodiment of the present invention
  • 5 is a scattergram of historical running data before clustering processing, and a scattergram after culling abnormal data by clustering the historical running data by using the improved DBSCAN clustering algorithm;
  • FIG. 6 is a flow chart of an implementation manner of determining a relationship between wind error and output performance by reducing dimensionality on target data in a method for automatically adjusting wind error of a wind turbine according to an embodiment of the present invention
  • FIG. 7 is a schematic diagram of probability and statistics of wind error and target wind error range according to an embodiment of the present invention.
  • FIG. 8 is a wind power curve of different pairs of wind errors according to an embodiment of the present invention.
  • FIG. 9 is a graph showing relationship between wind error and output performance according to an embodiment of the present invention.
  • FIG. 10 is a schematic diagram of power deviation between actual active power and theoretical active power of a data point according to an embodiment of the present invention.
  • FIG. 11 is a schematic diagram showing statistical results of power deviations below a rated wind speed and above a rated wind speed according to an embodiment of the present invention
  • FIG. 13 is a diagram showing an evaluation result of an output performance of a target unit before and after correction according to an embodiment of the present invention
  • FIG. 14 is a schematic structural diagram of a wind speed automatic calibration device for a wind turbine according to an embodiment of the present invention.
  • the wind vane is mounted on the weather rack at the top rear of the nacelle cover to measure the deviation angle of the wind direction from the centerline of the nacelle, ie the wind error. Due to the difference in magnitude between the feature size of the wind vane and the top plane of the nacelle cover, it is difficult to ensure that the wind direction mark zero line is parallel to the centerline of the nacelle by visual inspection by a worker or by using a simple wind vane to find a zero tooling. According to engineering experience, during the operation of the wind farm, there are still many factors that cause the wind vane to shift or loosen. As shown in Fig.
  • the wind turbine automatic wind calibration method provided by the embodiment of the present invention fully utilizes the historical operation data of the wind turbine, and automatically determines the inherent wind of the yaw system of the wind turbine by performing data mining and analysis on the data. The error, and then automatically adjust the zero parameter in the yaw control system according to the inherent wind error, which enhances the self-adaptive capability of the yaw system and improves the actual output performance of the unit.
  • the method may include:
  • Step S201 Acquire historical operation data of the target unit in the preset time period.
  • the historical operation data of the target unit may include sampling time of the data, fan status, wind speed, active power, generator speed, blade pitch angle, and wind error.
  • Step S202 Excluding the abnormal data from the historical running data of the target unit to obtain the target data.
  • Step S203 Perform dimensionality reduction processing on the target data, and determine a relationship curve between the wind error and the output performance by the dimensionality-reduced data.
  • Step S204 Determine the inherent wind error by comparing the relationship between the wind error and the output performance.
  • the wind error corresponding to the highest point in the relationship between the wind error and the output performance is taken as the inherent wind error of the target unit.
  • Step S205 Correcting the zero parameter in the yaw system of the target unit by the inherent wind error.
  • the wind turbine automatic wind calibration method obtains the historical operation data of the unit, first removes the abnormal data, and then determines the relationship between the wind error and the output performance by performing dimensional reduction processing on the data after the abnormal data is removed. Then, the inherent wind error is determined by the relationship between the wind error and the output performance, and then the zero parameter in the yaw system of the target unit is corrected by the inherent wind error.
  • the method provided by the embodiment of the invention can determine the inherent wind error of the yaw system of the wind turbine by clustering and dimension reduction of the historical operation data, and can automatically adjust the yaw control system according to the inherent wind error.
  • the zero position parameter which makes the adaptive ability of the unit yaw system enhanced, and the actual output performance of the unit is improved.
  • FIG. 3 another flow diagram of a method for automatically adjusting wind error of a wind turbine according to an embodiment of the present invention is shown.
  • the method may include:
  • Step S301 Acquire historical operation data of the target unit in the preset time period.
  • the historical operation data of the target unit may include sampling time of the data, fan status, wind speed, active power, generator speed, blade pitch angle, and wind error.
  • Step S302 Excluding the abnormal data from the historical running data of the target unit to obtain the target data.
  • Step S303 Perform dimensionality reduction processing on the target data, and determine a relationship curve between the wind error and the output performance by the dimension reduction processed data.
  • Step S304 Determine the inherent wind error by comparing the relationship between the wind error and the output performance.
  • the wind error corresponding to the highest point in the relationship between the wind error and the output performance is taken as the inherent wind error of the target unit.
  • Step S305 Correct the zero parameter in the yaw system of the target unit by the inherent wind error.
  • Step S306 The change of the output performance of the target unit before and after the correction is evaluated based on the operational data deep purification technology.
  • the method for automatically adjusting the wind error of the wind turbine obtains the historical operation data of the unit, first removes the abnormal data, and then determines the wind error and the output performance by performing dimensionality reduction on the data after the abnormal data is removed.
  • the relationship curve determines the inherent wind error by the relationship between the wind error and the output performance, and then corrects the zero parameter in the yaw system of the target unit through the inherent wind error, and finally the output of the target unit before and after the correction. Performance is evaluated.
  • the method provided by the embodiment of the invention can determine the inherent wind error of the yaw system of the wind turbine by clustering and dimension reduction of the historical operation data, and can automatically adjust the yaw control system according to the inherent wind error.
  • the zero parameter of the unit makes the adaptive ability of the yaw system of the unit enhanced, and the actual output performance of the unit is improved by the evaluation result.
  • the present invention removes the anomaly data from the historical operational data of the wind turbine.
  • the historical running data may be clustered, and the abnormal data may be eliminated based on the clustering result of the clustering processing.
  • the improved DBSCAN clustering algorithm is used to cluster the historical running data of the target unit, and the abnormal data is eliminated by the clustering result.
  • FIG. 4 a schematic flowchart of a specific implementation process of performing clustering processing on historical running data may be included, which may include:
  • Step S401 Perform normalization processing on the historical running data to obtain target historical running data.
  • Step S402 Determine the target noise data ratio ⁇ -noise according to the number of data objects in the target historical operation data.
  • Step S403 Calculate a geometric distance between each data object and the target object in the target historical running data, and obtain a distance set.
  • the target object is a data object whose target geometric data is close to the geometric distance k of the data object, and the initial value of k is 2.
  • Step S404 Perform probability statistics on each element in the distance set, and form a new distance set by the elements in the distance set whose probability values are within the preset probability range.
  • elements in the distance set whose probability values are within the probability range of 5%-95% may be grouped into a new distance set.
  • Step S405 Determine the mathematical expectation value of the new distance set as the parameter Eps k of the DBSCAN clustering algorithm.
  • Step S407 Perform probability statistics on the elements in the number set, and form elements of the number set whose probability values are within the preset probability range into a new set of numbers.
  • Step S408 Determine the mathematical expectation value of the new number set as the parameter Minpts k of the DBSCAN clustering algorithm.
  • the primes of the probability set in the set of numbers in the probability range of 5%-95% can be composed into a new set of numbers P k *, and the mathematical expectation of P k * can be assigned to Minpts k .
  • Step S409 Perform DBSCAN clustering processing on the target historical running data with the parameters Minpts k and Eps k .
  • Step S410 Calculate the current noise data ratio Ratio-noise k according to the clustering result of the DBSCAN clustering process.
  • Step S411 determining whether k satisfies k>2, and whether Ratio-noise k satisfies:
  • FIG. 5-a shows a scatter plot of the historical running data before the clustering process
  • FIG. 5-b shows the clustering process of the historical running data after the abnormal data is removed by using the improved DBSCAN clustering algorithm.
  • DBSCAN clustering algorithm is a classic density-based clustering algorithm. It has fast clustering speed and can find clusters of arbitrary shapes in datasets with abnormal data.
  • Eps and Minpts In the clustering process, the Eps and Minpt parameters of the DBSCAN algorithm are adaptively adjusted in the clustering process to ensure the reliability and accuracy of the clustering result, thereby fully eliminating abnormal data in wind power.
  • the abnormal data can be eliminated, and the data can be filtered based on the normal variation range of the main parameters in the historical running data, and then the data is reconstructed by linear interpolation.
  • Get target data may be obtained based on a method of eliminating abnormal data such as a quartile method or a k-means clustering algorithm.
  • the wind turbine error correction method for the wind turbine provided in the above embodiment is performed, and the target data is subjected to dimensionality reduction processing, and the relationship between the wind error and the output performance is determined by the dimension reduction processed data.
  • a schematic flow diagram of an implementation manner which may include:
  • Step S601 Perform probability statistics on the wind error in the target data, and determine a target wind error range based on the probability statistics.
  • the wind error range determined by the wind error within a probability range of 10% to 90% can be determined as the target wind error range, as shown in FIG. 7 .
  • Step S602 Excluding data that is not within the target wind error range from the target data, and obtaining the first target data.
  • Step S603 Determine, by using the first target data, a wind power curve corresponding to each pair of wind errors.
  • the wind power curve is the relationship between wind speed and active power.
  • the Bin method is used to reduce the dimensionality of the data, and the wind power curve under different wind errors as shown in FIG. 8 is obtained, and the wind power curve fitting method can be referred to. IEC 61400-12-1-2005 standard implementation.
  • Step S604 The wind power curve is processed into a relationship between the wind error and the output performance.
  • the wind power curve under different wind errors is integrated and converted into the output performance index quantitative value EOH (the equivalent annual full-time hours), and the wind error and output performance as shown in FIG. 9 are obtained.
  • EOH the equivalent annual full-time hours
  • the relationship curve, the wind error corresponding to the highest point of the curve is taken as the inherent wind error of the target unit.
  • the embodiment of the present invention further evaluates the change in the output performance of the target unit before and after the correction based on the operational data deep purification technique.
  • the specific process for evaluating the change of the output performance of the target unit before and after correction based on the operational data deep purification technology includes:
  • the pre-correction operation data and the corrected operation data are respectively divided into two groups based on the rated wind speed.
  • Figures 11-a and 11-b show the statistical results of the power deviation below the rated wind speed and above the rated wind speed, respectively.
  • step (2) the data points whose probability density is not within the preset range are removed, and the new pre-correction target operation data and the new corrected target operation data are obtained, and then the process proceeds to step (2) until the fitting is obtained.
  • the wind power curve no longer changes. Specifically, data points with probability density less than 10% and probability density greater than 90% are regarded as outlier data and edge data, and these data are eliminated.
  • an embodiment of the present invention further provides a wind turbine automatic wind calibration device.
  • the device may include: a data acquisition module 1401 and an abnormal data.
  • the data acquisition module 1401 is configured to acquire historical operation data of the target unit within a preset time period.
  • the data culling module 1402 is configured to remove abnormal data from the historical running data to obtain target data.
  • the dimension reduction processing module 1403 is configured to perform dimension reduction processing on the target data obtained by removing the abnormal data from the data culling module 1402.
  • the curve determination module 1404 is configured to determine a relationship between the wind error and the output performance by the dimension reduction processing data of the dimensionality reduction processing module 1403.
  • the inherent wind error determination module 1405 is configured to determine the inherent wind error by the relationship between the wind error and the output performance determined by the curve determination module 1404.
  • the correction module 1406 is configured to correct the zero parameter in the yaw system of the target unit by the inherent wind error determined by the inherent wind error determination module 1405.
  • the wind turbine automatic wind calibration device obtains the historical operation data of the unit, first removes the abnormal data, and then determines the relationship between the wind error and the output performance by performing dimensional reduction processing on the data after the abnormal data is removed. Then, the inherent wind error is determined by the relationship between the wind error and the output performance, and then the zero parameter in the yaw system of the target unit is corrected by the inherent wind error.
  • the device provided by the embodiment of the present invention can gather historical operation data Class, dimension reduction and other processes determine the inherent wind error of the yaw system of the wind turbine, and then automatically adjust the zero parameter in the yaw control system according to the inherent wind error, which makes the yaw system adaptive ability It is enhanced and the actual output performance of the unit is improved.
  • the data culling module comprises: a clustering module and a culling module. among them:
  • the clustering module is configured to perform clustering processing on historical running data to obtain clustering results.
  • the culling module is used for culling abnormal data based on the clustering result of the clustering process.
  • the clustering module includes: a normalization processing sub-module, a noise data ratio determination sub-module, a distance calculation sub-module, a first probability statistics sub-module, a distance set determination sub-module, a first parameter determination sub-module, and a number set determiner. a module, a second probability statistics module, a first number set determining submodule, a second probability statistics submodule, a second number set determining submodule, a second parameter determining submodule, a clustering processing submodule, and a noise data ratio calculating subroutine Modules, judgment submodules, and assignment submodules. among them:
  • the standardized processing sub-module is used for standardizing the historical running data to obtain the target historical running data.
  • the noise data ratio determining submodule is configured to determine the target noise data ratio ⁇ -noise according to the number of data objects in the target historical running data.
  • the distance calculation sub-module is configured to calculate a geometric distance between each data object and the target object in the target historical running data, and obtain a distance set.
  • the target object is a data object whose target geometric data is close to the geometric distance k of the data object, and the initial value of k is 2.
  • a first probability statistics sub-module configured to perform probability statistics on each element in the distance set
  • the distance determination determining sub-module is configured to form a new distance set by the elements in the distance set whose probability values are within a preset probability range.
  • the first parameter determining sub-module is configured to determine a mathematical expectation value of the new distance set as the parameter Eps k of the DBSCAN clustering algorithm.
  • the first number set determining submodule is configured to determine a set of the number of points in the Epsk domain of all data objects in the target historical running data.
  • the second probability statistics sub-module is configured to perform probability statistics on the elements in the number set.
  • the second number set determining submodule is configured to form a new set of numbers of elements in the set of numbers whose probability values are within a preset probability range.
  • the second parameter determining sub-module is configured to determine a mathematical expectation value of the new number set as a parameter Minpts k of the DBSCAN clustering algorithm.
  • the clustering processing sub-module is configured to perform DBSCAN clustering processing on the target historical running data with the parameters Minptsk and Epsk.
  • the noise data ratio calculation sub-module is configured to calculate the current noise data ratio Ratio-noise k according to the clustering result of the DBSCAN clustering process.
  • a judging submodule for judging whether k is greater than 2 and Ratio-noisek satisfies
  • the dimensionality reduction processing module comprises: a probability statistics sub-module and a data rejection sub-module. among them:
  • the probability statistics sub-module is configured to perform probability statistics on the wind error in the target data, and determine the target wind error range based on the probability statistics.
  • the data culling sub-module is configured to remove the data that the wind error is not within the target wind error range from the target data, and obtain the first target data.
  • the curve determination module includes: a first curve determination sub-module and a second curve determination sub-module. among them:
  • the first curve determining submodule is configured to determine, by using the first target data, a wind power curve corresponding to each pair of wind errors, where the wind power curve is a relationship between the wind speed and the active power.
  • the second curve determination sub-module is configured to process the wind power curve into a relationship between the wind error and the output performance.
  • the wind turbine pair wind error automatic calibration apparatus further includes: an evaluation module. among them:
  • the evaluation module is used to evaluate the change of the output performance of the target unit before and after the calibration according to the preset evaluation method.
  • the further evaluation module comprises: a data acquisition sub-module, a data grouping module block, an abnormal data culling sub-module, a data merging sub-module, a data fitting sub-module, a data interpolation sub-module, a power deviation calculation sub-module, an abnormal data culling sub-module, Wind power curve processing sub-module and comparison sub-module.
  • the data acquisition sub-module is used to obtain the pre-correction operation data of the target unit and the corrected operation data.
  • the data grouping module block is configured to divide the pre-correction operation data and the corrected operation data into two groups respectively at a rated wind speed.
  • the abnormal data culling sub-module is used to remove the abnormal data from the two sets of pre-corrected running data, and to remove the abnormal data from the two sets of corrected running data.
  • the data merging sub-module is configured to combine the two sets of pre-correction operation data after the abnormal data is removed to obtain the pre-correction target operation data, and combine the two sets of corrected operation data after the abnormal data is removed to obtain the corrected target operation data.
  • a data fitting submodule for fitting the pre-correction target operational data and the corrected target operational data to the wind power curve, respectively.
  • the data interpolation sub-module is configured to linearly interpolate the wind speed in the target operating data before the correction and the corrected target operating data, and obtain the theoretical active power corresponding to each wind speed.
  • the power deviation calculation sub-module is configured to separately calculate a power deviation between the actual active power and the theoretical active power of each data point in the corrected target operation data and the corrected target operation data, and obtain a power deviation corresponding to each data point.
  • the abnormal data culling sub-module is configured to perform probability statistics on the power deviation corresponding to each data point, and according to the probability statistical result, the data points whose probability density is not within the preset range are eliminated, and the new pre-corrected operational data and the new one are obtained.
  • the corrected operational data is then triggered to divide the pre-correction operational data into two groups at a nominal wind speed boundary until the fitted wind power curve no longer changes.
  • the wind power curve processing sub-module is configured to separately integrate the wind power curve before and after the correction, and convert the wind power curve before and after the correction into the quantized value of the output performance index before and after the correction.
  • the comparison sub-module is used to compare the magnitude of the quantitative value of the output performance index before and after the correction, and obtain the change of the output performance of the target unit before and after the correction.
  • the method and device for automatically calibrating the wind error of the wind turbine provided by the invention, clustering the operation data of the unit through the improved DBSCAN algorithm, eliminating the abnormal data, completely eliminating the anomaly data automatically in the wind turbine yaw system inherent wind error
  • the adverse effects in the identification process the accuracy of online identification of the inherent wind error of the yaw system is improved by probability statistics and dimensionality reduction, and timely feedback can be achieved when the wind direction indicator is again offset by external disturbances; Automatically adjusting the zero parameter in the yaw control system according to the inherent wind error can enhance the adaptive level of the yaw system of the unit and improve the wind efficiency. At the same time, it can also eliminate the periodic review of the wind direction indicator by the operation and maintenance personnel and reduce the human error.
  • the disclosed methods, apparatus, and devices may be implemented in other manners.
  • the device embodiments described above are merely illustrative.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored or not executed.
  • Another point that is shown or discussed between each other The coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some communication interface, device or unit, and may be in an electrical, mechanical or other form.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the functions may be stored in a computer readable storage medium if implemented in the form of a software functional unit and sold or used as a standalone product.
  • the technical solution of the present invention which is essential or contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product, which is stored in a storage medium, including
  • the instructions are used to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • the foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like. .

Abstract

一种风电机组对风误差自动校准方法及装置。方法包括:获取预设时段内目标机组的历史运行数据;从目标机组的历史运行数据中剔除异常数据;对目标数据进行降维处理,通过降维处理后的数据确定对风误差与出力性能的关系曲线;通过对风误差与出力性能的关系曲线确定固有对风误差;通过固有对风误差校正目标机组的偏航系统中的零位参数。装置具有多个模块,每个模块的功能和方法相对应。上述方法和装置利用风电机组的历史运行数据,通过对其进行数据挖掘和分析自动辨识出目标机组偏航系统的固有对风误差,进而能根据该固有对风误差自动调整偏航系统中的零位参数,增强了机组偏航系统的自适应能力,提升了机组的实际出力性能。

Description

一种风电机组对风误差自动校准方法及装置 技术领域
本申请要求于2016年7月21日提交中国专利局、申请号为201610579606.4、发明名称为“一种风电机组对风误差自动校准方法及装置”的国内申请的优先权,其全部内容通过引用结合在本申请中。
背景技术
近年来,风电行业中机舱尾部的风向标测风失准的问题逐渐引起了业内关注。风向标安装时校准方法不当、运维人员的操作误差、台风等极限工况扰动等诸多因素都可能造成风向标的绝对零位与机舱中心线不平行,使得偏航系统存在“固有对风误差”,从而导致机组输出功率达不到设计要求。
发明内容
有鉴于此,本发明提供了一种风电机组对风误差自动校准方法及装置,用以解决现有技术中,由于诸多原因导致风向标的绝对零位与机舱中心线不平行,使得偏航系统存在“固有对风误差”,进而导致机组输出功率达不到设计要求的问题,其技术方案如下:
一种风电机组对风误差自动校准方法,所述方法包括:
获取预设时段内目标机组的历史运行数据;
从所述目标机组的历史运行数据中剔除异常数据,获得目标数据;
对所述目标数据进行降维处理,并通过所述降维处理后的数据确定对风误差与出力性能的关系曲线;
通过所述对风误差与出力性能的关系曲线确定固有对风误差;
通过所述固有对风误差校正所述目标机组的偏航系统中的零位参数。
其中,所述从所述目标机组的历史运行数据中剔除异常数据,包括:
对所述历史运行数据进行聚类处理,并基于所述聚类处理的聚类结果剔除所述异常数据。
其中,所述对所述历史运行数据进行聚类处理,包括:
对所述历史运行数据进行标准化处理,获得目标历史运行数据;
根据所述目标历史运行数据中数据对象的数量确定目标噪声数据占比ε-noise;
计算所述目标历史运行数据中每个数据对象与目标对象的几何距离,获得距离集合,其中,所述目标对象为所述目标历史数据中与所述数据对象的几何距离第k近的数据对象,所述k的初始取值为2;
对所述距离集合中的各个元素进行概率统计,将所述距离集合中概率值在预设概率范围内的元素组成新的距离集合;
确定所述新的距离集合的数学期望值作为DBSCAN聚类算法的参数Epsk
确定所述目标历史运行数据中所有数据对象的Epsk域内点的数目集合;
对所述数目集合中的元素进行概率统计,将所述数目集合中概率值在预设概率范围内的元素组成新的数目集合;
确定所述新的数目集合的数学期望值作为所述DBSCAN聚类算法的参数Minptsk
以参数Minptsk和Epsk对所述目标历史运行数据进行DBSCAN聚类处理;
根据所述DBSCAN聚类处理的聚类结果计算当前噪声数据占比Ratio-noisek
当k>2,且Ratio-noisek满足|Ratio-noisek-1-Ratio-noisek|≤ε-noise时,结束聚类处理,否则,将k+1赋值给k,然后返回执行所述计算所述目标历史运行数据中每个数据对象与目标对象的几何距离这一步骤。
其中,所述对所述目标数据进行降维处理,并通过所述降维处理后的数据确定对风误差与出力性能的关系曲线,包括:
对所述目标数据中的对风误差进行概率统计,并基于概率统计结果确定目标对风误差范围;
从所述目标数据中剔除对风误差不在所述目标对风误差范围内的数据,获得第一目标数据;
利用所述第一目标数据确定与各个对风误差对应的风功率曲线,风功率曲线为风速与有功功率的关系曲线;
将所述风功率曲线处理成对风误差与出力性能的关系曲线。
优选地,所述方法还包括:
基于运行数据深度净化技术评估校正前后所述目标机组的出力性能的变化情况。
其中,所述基于运行数据深度净化技术评估校正前后所述目标机组的出力性能的变化情况,包括:
获取目标机组的校正前运行数据以及校正后运行数据;
以额定风速为界分别将所述校正前运行数据和所述校正后运行数据分成两组;
从两组校正前运行数据中剔除异常数据,将剔除异常数据后的两组校正前运行数据合并得到校正前目标运行数据,并从两组校正后运行数据中剔除异常数据,将剔除异常数据后的两组校正后运行数据合并得到校正后目标运行数据;
分别将所述校正前目标运行数据和所述校正后目标运行数据拟合到所述风功率曲线,并分别对所述校正前目标运行数据和所述校正后目标运行数据中的风速以所述风功率曲线为节点进行线性插值,获得与每个风速对应的理论有功功率;
分别计算所述校正前目标运行数据和所述校正后目标运行数据中每个数据点的实际有功功率与理论有功功率间的功率偏差,获得与各个数据点对应的功率偏差;
对所述与各个数据点对应的功率偏差进行概率统计;
根据概率统计结果,将概率密度不在预设范围内的数据点剔除,然后返回所述以额定风速为界将所述校正前运行数据分成两组,并以额定风速为界将所述校正后运行数据分成两组这一步骤,直至拟合得到的风功率曲线不再变化;
通过分别对校正前后的风功率曲线进行积分处理,将所述校正前后的风功率曲线转换为校正前后的出力性能指标量化值;
对比校正前后的出力性能指标量化值的大小,获得校正前后所述目标机组出力性能的变化情况。
一种风电机组对风误差自动校准装置,所述装置包括:
数据获取模块,用于获取预设时段内目标机组的历史运行数据;
数据剔除模块,用于从所述数据获取模块获取的所述历史运行数据中剔除异常数据,获得目标数据;
降维处理模块,用于对所述数据剔除模块剔除异常数据后所得到的所述目标数据进行降维处理;
曲线确定模块,用于通过所述降维处理模块降维处理后的数据确定对风误差与出力性能的关系曲线;
固有对风误差确定模块,用于通过所述曲线确定模块确定出的所述对风误差与出力性能的关系曲线确定固有对风误差;
校正模块,用于通过所述固有对风误差确定模块确定的所述固有对风误差校正所述目标机组的偏航系统中的零位参数。
其中,所述数据剔除模块包括:
聚类模块,用于对所述历史运行数据进行聚类处理,获得聚类结果;
剔除模块,用于基于所述聚类处理的聚类结果剔除所述异常数据。
其中,所述降维处理模块包括:
所述概率统计子模块,用于对所述目标数据中的对风误差进行概率统计,并基于概率统计结果确定目标对风误差范围;
所述数据剔除子模块,用于从所述目标数据中剔除对风误差不在所述目标对风误差范围内的数据,获得第一目标数据;
所述曲线确定模块包括:第一曲线确定子模块和第二曲线确定子模块;
所述第一曲线确定子模块,用于利用所述第一目标数据确定与各个对风误差对应的风功率曲线,风功率曲线为风速与有功功率的关系曲线;
所述第二曲线确定子模块,用于将所述风功率曲线处理成对风误差与出力性能的关系曲线。
所述装置还包括:
评估模块,用于基于运行数据深度净化技术评估校正前后所述目标机组 的出力性能的变化情况;
所述评估模块包括:
数据获取子模块,用于获取目标机组的校正前运行数据以及校正后运行数据;
数据分组模子块,用于以额定风速为界分别将所述校正前运行数据和所述校正后运行数据分成两组;
异常数据剔除子模块,用于从两组校正前运行数据中剔除异常数据,并从两组校正后运行数据中剔除异常数据;
数据合并子模块,用于将剔除异常数据后的两组校正前运行数据合并得到校正前目标运行数据,并将剔除异常数据后的两组校正后运行数据合并得到校正后目标运行数据;
数据拟合子模块,用于分别将所述校正前目标运行数据和所述校正后目标运行数据拟合到所述风功率曲线;
数据插值子模块,用于分别对所述校正前目标运行数据和所述校正后目标运行数据中的风速以所述风功率曲线为节点进行线性插值,获得与每个风速对应的理论有功功率;
功率偏差计算子模块,用于分别计算所述校正前目标运行数据和所述校正后目标运行数据中每个数据点的实际有功功率与理论有功功率间的功率偏差,获得与各个数据点对应的功率偏差;
异常数据剔除子模块,用于对所述与各个数据点对应的功率偏差进行概率统计,并根据概率统计结果将概率密度不在预设范围内的数据点剔除,获得新的校正前的运行数据和新的校正后的运行数据,然后触发数据分组模子块所述以额定风速为界将所述校正前运行数据分成两组,直至拟合得到的风功率曲线不再变化;
风功率曲线处理子模块,用于分别对校正前后的风功率曲线进行积分处理,将所述校正前后的风功率曲线转换为校正前后的出力性能指标量化值;
对比子模块,用于对比校正前后的出力性能指标量化值的大小,获得校正前后所述目标机组出力性能的变化情况。
上述技术方案具有如下有益效果:
本发明提供的风电机组对风误差自动校准方法及装置,充分利用风电机组的历史运行数据,通过对风电机组的历史运行数据进行数据挖掘和分析自动辨识出风电机组的偏航系统的固有对风误差,进而能够根据该固有对风误差自动调整偏航控制系统中的零位参数,这使得机组偏航系统的自适应能力得到增强,并且机组的实际出力性能得到提升,使机组输出功率能够达到设计要求。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本发明的实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据提供的附图获得其他的附图。
图1为存在固有对风误差时实际风向与机舱位置关系示意图;
图2为本发明实施例提供的风电机组对风误差自动校准方法的一流程示意图;
图3为本发明实施例提供的风电机组对风误差自动校准方法的另一流程示意图;
图4为本发明实施例提供的风电机组对风误差自动校准方法中,对历史运行数据进行聚类处理的实现过程的流程示意图;
图5为聚类处理前历史运行数据的散点图,以及采用改进的DBSCAN聚类算法对历史运行数据进行聚类处理后剔除异常数据后的散点图;
图6为本发明实施例提供的风电机组对风误差自动校准方法中,对目标数据进行降维处理,并通过降维处理后的数据确定对风误差与出力性能的关系曲线的实现方式的流程示意图;
图7为本发明实施例提供的对风误差的概率统计及目标对风误差范围的示意图;
图8为本发明实施例提供的不同对风误差下的风功率曲线;
图9为本发明实施例提供的对风误差与出力性能的关系曲线;
图10为本发明实施例提供的数据点的实际有功功率与理论有功功率间的功率偏差示意图;
图11为本发明实施例提供的额定风速以下以及额定风速以上功率偏差的统计结果示意图;
图12为本发明实施例提供的校正前后机组的风功率曲线对比图;
图13为本发明实施例提供的校正前后目标机组出力性能评估结果图;
图14为本发明实施例提供的风电机组对风误差自动校准装置的一结构示意图。
具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本发明一部分实施例,而不是全部的实施例。基于本发明中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本发明保护的范围。
风向标安装在机舱罩顶部后端的气象架上,用来测量风向与机舱中心线的偏差角,即对风误差。由于风向标特征尺寸与机舱罩顶部平面的特征尺寸存在量级差异,通过工作人员目测或采用简单的风向标找零工装很难保证风向标零刻线与机舱中心线平行。根据工程经验,在风电场运行期间,还存在很多因素会导致风向标发生偏移或松动。如图1所示,假设风向标零刻线与机舱中心线夹角为α0,风向标测量的对风误差为α,那么机舱中心线与来流风向的夹角β=α+α0,α0被看作偏航系统的“固有对风误差”。由于“固有对风误差”的存在使机舱不能准确对风,将直接影响机组的实际出力性能。
同时,大数据处理和云计算技术的迅速崛起和应用对风电行业产生了巨大的冲击。考虑到风电场具备数量庞大的数据信息,但这些数据很大程度上都停留在无序的源数据形态,目前这些“闲置”数据并没有为风电场的运营过程提供更多的价值。
从上述考虑出发,本发明实施例提供的风电机组对风误差自动校准方法充分利用风电机组的历史运行数据,通过对这些数据进行数据挖掘和分析自动确定出风电机组的偏航系统的固有对风误差,进而根据该固有对风误差自动调整偏航控制系统中的零位参数,这使得机组偏航系统的自适应能力得到增强,机组的实际出力性能得到提升。
请参阅图2,示出了本发明实施例提供的风电机组对风误差自动校准方法的一流程示意图,该方法可以包括:
步骤S201:获取预设时段内目标机组的历史运行数据。
其中,目标机组的历史运行数据可以包括数据的采样时间、风机状态、风速、有功功率、发电机转速、桨叶桨距角及对风误差。
步骤S202:从目标机组的历史运行数据中剔除异常数据,获得目标数据。
步骤S203:对目标数据进行降维处理,并通过降维处理后的数据确定对风误差与出力性能的关系曲线。
步骤S204:通过对风误差与出力性能的关系曲线确定固有对风误差。
在本实施例中,以对风误差与出力性能的关系曲线中最高点对应的对风误差作为目标机组的固有对风误差。
步骤S205:通过固有对风误差校正目标机组的偏航系统中的零位参数。
本发明提供的风电机组对风误差自动校准方法,在获取到机组的历史运行数据,先剔除异常数据,然后通过对剔除异常数据后的数据进行降维处理确定对风误差与出力性能的关系曲线,接着通过对风误差与出力性能的关系曲线确定出固有对风误差,进而通过固有对风误差校正目标机组的偏航系统中的零位参数。本发明实施例提供的方法能够通过对历史运行数据进行聚类、降维等处理确定出风电机组的偏航系统的固有对风误差,进而能够根据该固有对风误差自动调整偏航控制系统中的零位参数,这使得机组偏航系统的自适应能力得到增强,机组的实际出力性能得到提升。
请参阅图3,示出了本发明实施例提供的风电机组对风误差自动校准方法的另一流程示意图,该方法可以包括:
步骤S301:获取预设时段内目标机组的历史运行数据。
其中,目标机组的历史运行数据可以包括数据的采样时间、风机状态、风速、有功功率、发电机转速、桨叶桨距角及对风误差。
步骤S302:从目标机组的历史运行数据中剔除异常数据,获得目标数据。
步骤S303:对目标数据进行降维处理,并通过降维处理后的数据确定对风误差与出力性能的关系曲线。
步骤S304:通过对风误差与出力性能的关系曲线确定固有对风误差。
在本实施例中,以对风误差与出力性能的关系曲线中最高点对应的对风误差作为目标机组的固有对风误差。
步骤S305:通过固有对风误差校正目标机组的偏航系统中的零位参数。
步骤S306:基于运行数据深度净化技术评估校正前后目标机组的出力性能的变化情况。
本发明实施例提供的风电机组对风误差自动校准方法,在获取到机组的历史运行数据,先剔除异常数据,然后通过对剔除异常数据后的数据进行降维处理确定对风误差与出力性能的关系曲线,接着通过对风误差与出力性能的关系曲线确定出固有对风误差,进而通过固有对风误差校正目标机组的偏航系统中的零位参数,最后还可对校正前后目标机组的出力性能进行评估。本发明实施例提供的方法能够通过对历史运行数据进行聚类、降维等处理确定出风电机组的偏航系统的固有对风误差,进而能够根据该固有对风误差自动调整偏航控制系统中的零位参数,这使得机组偏航系统的自适应能力得到增强,并且通过评估结果确定出机组的实际出力性能得到了提升。
通常情况下,风电机组的历史运行数据存在一些异常数据,而这些异常数据会对固有对风误差的确定产生的不利影响,为了避免异常数据对固有对风误差的确定产生的不利影响,本发明实施例将异常数据从风电机组的历史运行数据中剔除。剔除异常数据的实现方式有多种,在一种可能的实现方式中,可对历史运行数据进行聚类处理,并基于所述聚类处理的聚类结果剔除所述异常数据。优选的,可采用改进的DBSCAN聚类算法对目标机组的历史运行数据进行聚类处理,并通过聚类结果剔除异常数据。请参阅图4,示出了对历史运行数据进行聚类处理的具体实现过程的流程示意图,可以包括:
步骤S401:对历史运行数据进行标准化处理,获得目标历史运行数据。
需要说明的是,不同类型的运行数据量纲不同,取值范围差异较大,在基于数据密度的聚类算法(即DBSCAN聚类算法)进行聚类处理时,通常需要预先对其进行标准化处理,将数据折算到0-1之间。
步骤S402:根据目标历史运行数据中数据对象的数量确定目标噪声数据占比ε-noise。
步骤S403:计算目标历史运行数据中每个数据对象与目标对象的几何距离,获得距离集合。
其中,目标对象为目标历史数据中与数据对象的几何距离第k近的数据对象,k的初始取值为2。
假设目标历史运行数据中有n个对象,每个数据对象与目标对象的几何距离为k-dis,那么,距离集合为Distk={k-dis1,k-dis2,…,k-disn}。
步骤S404:对距离集合中的各个元素进行概率统计,将距离集合中概率值在预设概率范围内的元素组成新的距离集合。
在本实施例中,可将距离集合中概率值在5%-95%这一概率范围内的元素组成新的距离集合。
步骤S405:确定新的距离集合的数学期望值作为DBSCAN聚类算法的参数Epsk
步骤S406:确定目标历史运行数据中所有数据对象的Epsk域内点的数目集合Pk={p1,p2,…,pn}。
步骤S407:对数目集合中的元素进行概率统计,并将数目集合中概率值在预设概率范围内的元素组成新的数目集合。
步骤S408:确定新的数目集合的数学期望值作为DBSCAN聚类算法的参数Minptsk
在本实施例中,可将数目集合中概率值在5%-95%这一概率范围内的素组成新的数目集合Pk*,将Pk*的数学期望值赋值给Minptsk
步骤S409:以参数Minptsk和Epsk对目标历史运行数据进行DBSCAN聚类处理。
步骤S410:根据DBSCAN聚类处理的聚类结果计算当前噪声数据占比Ratio-noisek
步骤S411:判断k是否满足k>2,且Ratio-noisek是否满足:
|Ratio-noisek-1-Ratio-noisek|≤ε-noise
如果是,则结束聚类处理,否则,将k+1赋值给k,然后返回步骤S403。
请参阅图5,图5-a示出了聚类处理前历史运行数据的散点图,图5-b为采用改进的DBSCAN聚类算法对历史运行数据进行聚类处理后剔除异常数据后的散点图。DBSCAN聚类算法是一种经典的基于密度的聚类算法,该算法聚类速度快,能在带有异常数据的数据集中发现任意形状的聚类,DBSCAN聚类算法的准确性与Eps和Minpts这两个参数的选择有关,本发明实施例在聚类处理过程中通过自适应整定DBSCAN算法的Eps和Minpt参数,以便保证聚类结果的可靠性及准确性,进而能够充分消除异常数据在风电机组偏航系统的固有对风误差的确定过程中所产生的不利影响。
需要说明的是,除了通过上述改进的DBSCAN聚类算法剔除异常数据之外,还可基于历史运行数据中主要参数的正常变化范围对数据进行筛选,然后通过线性插值的方法对数据进行重构,获得目标数据。另外,也可基于四分位法、k-means聚类算法等剔除异常数据的方法,获得目标数据。
请参阅图6,示出了上述实施例提供的风电机组对风误差自动校准方法中,对目标数据进行降维处理,并通过降维处理后的数据确定对风误差与出力性能的关系曲线的实现方式的流程示意图,可以包括:
步骤S601:对目标数据中的对风误差进行概率统计,并基于概率统计结果确定目标对风误差范围。
在本实施例中,可将概率值在10%~90%这一概率范围内的对风误差所确定的对风误差范围确定为目标对风误差范围,如图7所示。
步骤S602:从目标数据中剔除对风误差不在目标对风误差范围内的数据,获得第一目标数据。
步骤S603:利用第一目标数据确定与各个对风误差对应的风功率曲线。
其中,风功率曲线为风速与有功功率的关系曲线。
在本实施例中,在目标对风误差范围,采用Bin法对数据进行降维处理,拟合得到如图8所示的不同对风误差下的风功率曲线,风功率曲线拟合方法可参照IEC 61400-12-1-2005标准执行。
步骤S604:将风功率曲线处理成对风误差与出力性能的关系曲线。
具体的,对不同对风误差下的风功率曲线进行积分处理,折算为出力性能指标量化值EOH(理论年等效满发小时数),得到如图9所示的对风误差与出力性能的关系曲线,以曲线最高点对应的对风误差作为目标机组的固有对风误差。
在确定出固有对风误差过后,可通过固有对风误差校正目标机组的偏航系统中的零位参数。为了确定校正效果,本发明实施例进一步基于运行数据深度净化技术评估校正前后目标机组的出力性能的变化情况。
具体的,基于运行数据深度净化技术评估校正前后目标机组的出力性能的变化情况的具体过程包括:
(1)获取目标机组的校正前运行数据以及校正后运行数据。
(2)以额定风速为界分别将校正前运行数据和所述校正后运行数据分成两组。
(3)从两组校正前运行数据中剔除异常数据,将剔除异常数据后的两组校正前运行数据合并得到校正前目标运行数据,并从两组校正后运行数据中剔除异常数据,将剔除异常数据后的两组校正后运行数据合并得到校正后目标运行数据。
(4)分别将校正前目标运行数据和校正后目标运行数据拟合到风功率曲线,并分别对校正前目标运行数据和校正后目标运行数据中的风速以风功率曲线为节点进行线性插值,获得与每个风速对应的理论有功功率。
(5)分别计算校正前目标运行数据和校正后目标运行数据中每个数据点的实际有功功率与理论有功功率间的功率偏差,获得与各个数据点对应的功率偏差。图10示出了一数据点的实际有功功率与理论有功功率间的功率偏差。
(6)对与各个数据点对应的功率偏差进行概率统计。图11-a、图11-b分别示出了额定风速以下以及额定风速以上功率偏差的统计结果。
(7)根据概率统计结果,将概率密度不在预设范围内的数据点剔除,得到新的校正前目标运行数据和新的校正后目标运行数据,然后转入步骤(2),直至拟合得到的风功率曲线不再变化。具体的,将概率密度小于10%以及概率密度大于90%的数据点视为离群数据和边缘数据,将这些数据剔除。
(8)通过分别对校正前后的风功率曲线进行积分处理,将校正前后的风功率曲线转换为校正前后的出力性能指标量化值。图12为校正前后机组的风功率曲线对比图。
(9)对比校正前后的出力性能指标量化值的大小,获得校正前后目标机组出力性能的变化情况。图13为校正前后目标机组出力性能评估结果图。
与上述方法相对应,本发明实施例还提供了一种风电机组对风误差自动校准装置,请参阅图14,示出了该装置的结构示意图,该装置可以包括:数据获取模块1401、异常数据剔除模块1402、降维处理模块1403、曲线确定模块1404、固有对风误差确定模块1405和校正模块1406。
数据获取模块1401,用于获取预设时段内目标机组的历史运行数据。
数据剔除模块1402,用于从历史运行数据中剔除异常数据,获得目标数据。
降维处理模块1403,用于对数据剔除模块1402剔除异常数据后所得到的目标数据进行降维处理。
曲线确定模块1404,用于通过降维处理模块1403降维处理后的数据确定对风误差与出力性能的关系曲线。
固有对风误差确定模块1405,用于通过曲线确定模块1404确定出的对风误差与出力性能的关系曲线确定固有对风误差。
校正模块1406,用于通过固有对风误差确定模块1405确定的固有对风误差校正目标机组的偏航系统中的零位参数。
本发明提供的风电机组对风误差自动校准装置,在获取到机组的历史运行数据,先剔除异常数据,然后通过对剔除异常数据后的数据进行降维处理确定对风误差与出力性能的关系曲线,接着通过对风误差与出力性能的关系曲线确定出固有对风误差,进而通过固有对风误差校正目标机组的偏航系统中的零位参数。本发明实施例提供的装置能够通过对历史运行数据进行聚 类、降维等处理确定出风电机组的偏航系统的固有对风误差,进而能够根据该固有对风误差自动调整偏航控制系统中的零位参数,这使得机组偏航系统的自适应能力得到增强,机组的实际出力性能得到提升。
在上述实施例提供的风电机组对风误差自动校准装置中,数据剔除模块包括:聚类模块和剔除模块。其中:
聚类模块,用于对历史运行数据进行聚类处理,获得聚类结果。
剔除模块,用于基于聚类处理的聚类结果剔除异常数据。
进一步的,聚类模块包括:标准化处理子模块、噪声数据占比确定子模块、距离计算子模块、第一概率统计子模块、距离集合确定子模块、第一参数确定子模块、数目集合确定子模块、第二概率统计模块、第一数目集合确定子模块、第二概率统计子模块、第二数目集合确定子模块、第二参数确定子模块、聚类处理子模块、噪声数据占比计算子模块、判断子模块和赋值子模块。其中:
标准化处理子模块,用于对历史运行数据进行标准化处理,获得目标历史运行数据。
噪声数据占比确定子模块,用于根据目标历史运行数据中数据对象的数量确定目标噪声数据占比ε-noise。
距离计算子模块,用于计算目标历史运行数据中每个数据对象与目标对象的几何距离,获得距离集合。其中,目标对象为目标历史数据中与数据对象的几何距离第k近的数据对象,k的初始取值为2。
第一概率统计子模块,用于对距离集合中的各个元素进行概率统计;
距离集合确定子模块,用于将距离集合中概率值在预设概率范围内的元素组成新的距离集合。
第一参数确定子模块,用于确定新的距离集合的数学期望值作为DBSCAN聚类算法的参数Epsk
第一数目集合确定子模块,用于确定目标历史运行数据中所有数据对象的Epsk域内点的数目集合。
第二概率统计子模块,用于对数目集合中的元素进行概率统计。
第二数目集合确定子模块,用于将数目集合中概率值在预设概率范围内的元素组成新的数目集合。
第二参数确定子模块,用于确定新的数目集合的数学期望值作为DBSCAN聚类算法的参数Minptsk
聚类处理子模块,用于以参数Minptsk和Epsk对目标历史运行数据进行DBSCAN聚类处理。
噪声数据占比计算子模块,用于根据DBSCAN聚类处理的聚类结果计算当前噪声数据占比Ratio-noisek
判断子模块,用于判断k是否大于2且Ratio-noisek满足|Ratio-noisek-1-Ratio-noisek|≤ε-noise,如果是,则结束聚类处理,如果否则触发赋值子模块将k+1赋值给k,并触发距离计算子模块计算目标历史运行数据中每个数据对象与目标对象的几何距离。
在上述实施例提供的风电机组对风误差自动校准装置中,降维处理模块包括:概率统计子模块和数据剔除子模块。其中:
概率统计子模块,用于对目标数据中的对风误差进行概率统计,并基于概率统计结果确定目标对风误差范围。
数据剔除子模块,用于从目标数据中剔除对风误差不在目标对风误差范围内的数据,获得第一目标数据。
曲线确定模块包括:第一曲线确定子模块和第二曲线确定子模块。其中:
第一曲线确定子模块,用于利用第一目标数据确定与各个对风误差对应的风功率曲线,风功率曲线为风速与有功功率的关系曲线。
第二曲线确定子模块,用于将风功率曲线处理成对风误差与出力性能的关系曲线。
在上述实施例提供的风电机组对风误差自动校准装置还包括:评估模块。其中:
评估模块,用于按预设评估方法评估校正前后目标机组的出力性能的变化情况。
进一步的评估模块包括:数据获取子模块、数据分组模子块、异常数据剔除子模块、数据合并子模块、数据拟合子模块、数据插值子模块、功率偏差计算子模块、异常数据剔除子模块、风功率曲线处理子模块和对比子模块。
数据获取子模块,用于获取目标机组的校正前运行数据以及校正后运行数据。
数据分组模子块,用于以额定风速为界分别将所述校正前运行数据和校正后运行数据分成两组。
异常数据剔除子模块,用于从两组校正前运行数据中剔除异常数据,并从两组校正后运行数据中剔除异常数据。
数据合并子模块,用于将剔除异常数据后的两组校正前运行数据合并得到校正前目标运行数据,并将剔除异常数据后的两组校正后运行数据合并得到校正后目标运行数据。
数据拟合子模块,用于分别将校正前目标运行数据和所述校正后目标运行数据拟合到所述风功率曲线。
数据插值子模块,用于分别对校正前目标运行数据和校正后目标运行数据中的风速以风功率曲线为节点进行线性插值,获得与每个风速对应的理论有功功率。
功率偏差计算子模块,用于分别计算校正前目标运行数据和校正后目标运行数据中每个数据点的实际有功功率与理论有功功率间的功率偏差,获得与各个数据点对应的功率偏差。
异常数据剔除子模块,用于对与各个数据点对应的功率偏差进行概率统计,并根据概率统计结果将概率密度不在预设范围内的数据点剔除,获得新的校正前的运行数据和新的校正后的运行数据,然后触发数据分组模子块以额定风速为界将所述校正前运行数据分成两组,直至拟合得到的风功率曲线不再变化。
风功率曲线处理子模块,用于分别对校正前后的风功率曲线进行积分处理,将校正前后的风功率曲线转换为校正前后的出力性能指标量化值。
对比子模块,用于对比校正前后的出力性能指标量化值的大小,获得校正前后目标机组出力性能的变化情况。
本发明提供的风电机组对风误差自动校准方法及装置,通过改进的DBSCAN算法对机组的运行数据进行聚类处理,剔除异常数据,充分消除了异常数据在风电机组偏航系统固有对风误差自动辨识过程中的不利影响;通过概率统计和降维处理提高了对偏航系统的固有对风误差进行在线辨识的辨识精度,并且当风向标受外部扰动再次发生偏移时,能够做到及时反馈;根据固有对风误差自动调整偏航控制系统中的零位参数,可以增强机组偏航系统的自适应水平,提高对风效率,同时还能免除运维人员对风向标的定期复核工作,减少人为误差,提高机组偏航系统的可靠性;基于运行数据深度净化技术对校正前后风电机组的出力性能进行评估时,由于剔除了异常数据,保留了能够表征机组稳态性能的纯净数据,因此可以实现对风电机组出力性能的准确评估。
本说明书中各个实施例采用递进的方式描述,每个实施例重点说明的都是与其他实施例的不同之处,各个实施例之间相同相似部分互相参见即可。
在本申请所提供的几个实施例中,应该理解到,所揭露的方法、装置和设备,可以通过其它的方式实现。例如,以上所描述的装置实施例仅仅是示意性的,例如,所述单元的划分,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式,例如多个单元或组件可以结合或者可以集成到另一个系统,或一些特征可以忽略,或不执行。另一点,所显示或讨论的相互之间 的耦合或直接耦合或通信连接可以是通过一些通信接口,装置或单元的间接耦合或通信连接,可以是电性,机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的,作为单元显示的部件可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。另外,在本发明各个实施例中的各功能单元可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读取存储介质中。基于这样的理解,本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括:U盘、移动硬盘、只读存储器(ROM,Read-Only Memory)、随机存取存储器(RAM,Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
对所公开的实施例的上述说明,使本领域专业技术人员能够实现或使用本发明。对这些实施例的多种修改对本领域的专业技术人员来说将是显而易见的,本文中所定义的一般原理可以在不脱离本发明的精神或范围的情况下,在其它实施例中实现。因此,本发明将不会被限制于本文所示的这些实施例,而是要符合与本文所公开的原理和新颖特点相一致的最宽的范围。

Claims (10)

  1. 一种风电机组对风误差自动校准方法,其特征在于,所述方法包括:
    获取预设时段内目标机组的历史运行数据;
    从所述目标机组的历史运行数据中剔除异常数据,获得目标数据;
    对所述目标数据进行降维处理,并通过所述降维处理后的数据确定对风误差与出力性能的关系曲线;
    通过所述对风误差与出力性能的关系曲线确定固有对风误差;
    通过所述固有对风误差校正所述目标机组的偏航系统中的零位参数。
  2. 根据权利要求1所述的自动校准方法,其特征在于,所述从所述目标机组的历史运行数据中剔除异常数据,包括:
    对所述历史运行数据进行聚类处理,并基于所述聚类处理的聚类结果剔除所述异常数据。
  3. 根据权利要求2所述的方法,其特征在于,所述对所述历史运行数据进行聚类处理,包括:
    对所述历史运行数据进行标准化处理,获得目标历史运行数据;
    根据所述目标历史运行数据中数据对象的数量确定目标噪声数据占比ε-noise;
    计算所述目标历史运行数据中每个数据对象与目标对象的几何距离,获得距离集合,其中,所述目标对象为所述目标历史数据中与所述数据对象的几何距离第k近的数据对象,所述k的初始取值为2;
    对所述距离集合中的各个元素进行概率统计,将所述距离集合中概率 值在预设概率范围内的元素组成新的距离集合;
    确定所述新的距离集合的数学期望值作为DBSCAN聚类算法的参数Epsk
    确定所述目标历史运行数据中所有数据对象的Epsk域内点的数目集合;
    对所述数目集合中的元素进行概率统计,将所述数目集合中概率值在预设概率范围内的元素组成新的数目集合;
    确定所述新的数目集合的数学期望值作为所述DBSCAN聚类算法的参数Minptsk
    以参数Minptsk和Epsk对所述目标历史运行数据进行DBSCAN聚类处理;
    根据所述DBSCAN聚类处理的聚类结果计算当前噪声数据占比Ratio-noisek
    当k>2,且Ratio-noisek满足|Ratio-noisek-1-Ratio-noisek|≤ε-noise时,结束聚类处理,否则,将k+1赋值给k,然后返回执行所述计算所述目标历史运行数据中每个数据对象与目标对象的几何距离这一步骤。
  4. 根据权利要求1所述的方法,其特征在于,所述对所述目标数据进行降维处理,并通过所述降维处理后的数据确定对风误差与出力性能的关系曲线,包括:
    对所述目标数据中的对风误差进行概率统计,并基于概率统计结果确定目标对风误差范围;
    从所述目标数据中剔除对风误差不在所述目标对风误差范围内的数 据,获得第一目标数据;
    利用所述第一目标数据确定与各个对风误差对应的风功率曲线,风功率曲线为风速与有功功率的关系曲线;
    将所述风功率曲线处理成对风误差与出力性能的关系曲线。
  5. 根据权利要求1-4中任意一项所述的方法,其特征在于,所述方法还包括:
    基于运行数据深度净化技术评估校正前后所述目标机组的出力性能的变化情况。
  6. 根据权利要求5所述的方法,其特征在于,所述基于运行数据深度净化技术评估校正前后所述目标机组的出力性能的变化情况,包括:
    获取目标机组的校正前运行数据以及校正后运行数据;
    以额定风速为界分别将所述校正前运行数据和所述校正后运行数据分成两组;
    从两组校正前运行数据中剔除异常数据,将剔除异常数据后的两组校正前运行数据合并得到校正前目标运行数据,并从两组校正后运行数据中剔除异常数据,将剔除异常数据后的两组校正后运行数据合并得到校正后目标运行数据;
    分别将所述校正前目标运行数据和所述校正后目标运行数据拟合到所述风功率曲线,并分别对所述校正前目标运行数据和所述校正后目标运行数据中的风速以所述风功率曲线为节点进行线性插值,获得与每个风速对应的理论有功功率;
    分别计算所述校正前目标运行数据和所述校正后目标运行数据中每个 数据点的实际有功功率与理论有功功率间的功率偏差,获得与各个数据点对应的功率偏差;
    对所述与各个数据点对应的功率偏差进行概率统计;
    根据概率统计结果,将概率密度不在预设范围内的数据点剔除,然后返回所述以额定风速为界将所述校正前运行数据分成两组,并以额定风速为界将所述校正后运行数据分成两组这一步骤,直至拟合得到的风功率曲线不再变化;
    通过分别对校正前后的风功率曲线进行积分处理,将所述校正前后的风功率曲线转换为校正前后的出力性能指标量化值;
    对比校正前后的出力性能指标量化值的大小,获得校正前后所述目标机组出力性能的变化情况。
  7. 一种风电机组对风误差自动校准装置,其特征在于,所述装置包括:
    数据获取模块,用于获取预设时段内目标机组的历史运行数据;
    数据剔除模块,用于从所述数据获取模块获取的所述历史运行数据中剔除异常数据,获得目标数据;
    降维处理模块,用于对所述数据剔除模块剔除异常数据后所得到的所述目标数据进行降维处理;
    曲线确定模块,用于通过所述降维处理模块降维处理后的数据确定对风误差与出力性能的关系曲线;
    固有对风误差确定模块,用于通过所述曲线确定模块确定出的所述对风误差与出力性能的关系曲线确定固有对风误差;
    校正模块,用于通过所述固有对风误差确定模块确定的所述固有对风 误差校正所述目标机组的偏航系统中的零位参数。
  8. 根据权利要求7所述的自动校准装置,其特征在于,所述数据剔除模块包括:
    聚类模块,用于对所述历史运行数据进行聚类处理,获得聚类结果;
    剔除模块,用于基于所述聚类处理的聚类结果剔除所述异常数据。
  9. 根据权利要求7所述的装置,其特征在于,所述降维处理模块包括:
    所述概率统计子模块,用于对所述目标数据中的对风误差进行概率统计,并基于概率统计结果确定目标对风误差范围;
    所述数据剔除子模块,用于从所述目标数据中剔除对风误差不在所述目标对风误差范围内的数据,获得第一目标数据;
    所述曲线确定模块包括:第一曲线确定子模块和第二曲线确定子模块;
    所述第一曲线确定子模块,用于利用所述第一目标数据确定与各个对风误差对应的风功率曲线,风功率曲线为风速与有功功率的关系曲线;
    所述第二曲线确定子模块,用于将所述风功率曲线处理成对风误差与出力性能的关系曲线。
  10. 根据权利要求7-9中任意一项所述的装置,其特征在于,所述装置还包括:
    评估模块,用于基于运行数据深度净化技术评估校正前后所述目标机组的出力性能的变化情况;
    所述评估模块包括:
    数据获取子模块,用于获取目标机组的校正前运行数据以及校正后运行数据;
    数据分组模子块,用于以额定风速为界分别将所述校正前运行数据和所述校正后运行数据分成两组;
    异常数据剔除子模块,用于从两组校正前运行数据中剔除异常数据,并从两组校正后运行数据中剔除异常数据;
    数据合并子模块,用于将剔除异常数据后的两组校正前运行数据合并得到校正前目标运行数据,并将剔除异常数据后的两组校正后运行数据合并得到校正后目标运行数据;
    数据拟合子模块,用于分别将所述校正前目标运行数据和所述校正后目标运行数据拟合到所述风功率曲线;
    数据插值子模块,用于分别对所述校正前目标运行数据和所述校正后目标运行数据中的风速以所述风功率曲线为节点进行线性插值,获得与每个风速对应的理论有功功率;
    功率偏差计算子模块,用于分别计算所述校正前目标运行数据和所述校正后目标运行数据中每个数据点的实际有功功率与理论有功功率间的功率偏差,获得与各个数据点对应的功率偏差;
    异常数据剔除子模块,用于对所述与各个数据点对应的功率偏差进行概率统计,并根据概率统计结果将概率密度不在预设范围内的数据点剔除,获得新的校正前的运行数据和新的校正后的运行数据,然后触发数据分组模子块所述以额定风速为界将所述校正前运行数据分成两组,直至拟合得到的风功率曲线不再变化;
    风功率曲线处理子模块,用于分别对校正前后的风功率曲线进行积分处理,将所述校正前后的风功率曲线转换为校正前后的出力性能指标量化 值;
    对比子模块,用于对比校正前后的出力性能指标量化值的大小,获得校正前后所述目标机组出力性能的变化情况。
PCT/CN2017/093303 2016-07-21 2017-07-18 一种风电机组对风误差自动校准方法及装置 WO2018014825A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
DK17830459.8T DK3489508T3 (da) 2016-07-21 2017-07-18 Fremgangsmåde og indretning til automatisk kalibrering af vindjusteringsfejl i vindkraftgenereringsenhed
EP17830459.8A EP3489508B1 (en) 2016-07-21 2017-07-18 Method and device for automatically calibrating wind alignment error of wind power generation unit
ES17830459T ES2926639T3 (es) 2016-07-21 2017-07-18 Método y dispositivo para calibrar automáticamente un error de alineación del viento de una unidad de generación de energía eólica

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201610579606.4A CN106014858B (zh) 2016-07-21 2016-07-21 一种风电机组对风误差自动校准方法及装置
CN201610579606.4 2016-07-21

Publications (1)

Publication Number Publication Date
WO2018014825A1 true WO2018014825A1 (zh) 2018-01-25

Family

ID=57117269

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/093303 WO2018014825A1 (zh) 2016-07-21 2017-07-18 一种风电机组对风误差自动校准方法及装置

Country Status (5)

Country Link
EP (1) EP3489508B1 (zh)
CN (1) CN106014858B (zh)
DK (1) DK3489508T3 (zh)
ES (1) ES2926639T3 (zh)
WO (1) WO2018014825A1 (zh)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109681380A (zh) * 2018-12-29 2019-04-26 新疆金风科技股份有限公司 用于调整最小桨距角位置的方法及装置
CN111145109A (zh) * 2019-12-09 2020-05-12 深圳先进技术研究院 基于图像的风力发电功率曲线异常数据识别与清洗方法
CN113236507A (zh) * 2021-05-28 2021-08-10 中南大学 一种风电机组偏航静态误差诊断方法和系统

Families Citing this family (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106014858B (zh) * 2016-07-21 2019-11-22 浙江运达风电股份有限公司 一种风电机组对风误差自动校准方法及装置
CN108071562B (zh) * 2016-11-17 2021-01-15 中国电力科学研究院 一种基于能量流的风电机组能效状态诊断方法
CN106870298B (zh) * 2017-03-28 2020-04-07 中车株洲电力机车研究所有限公司 基于机器学习的叶根螺栓断裂故障检测方法
CN106991508A (zh) * 2017-05-25 2017-07-28 华北电力大学 一种基于dbscan的风电机组运行状态识别方法
CN109086793B (zh) * 2018-06-27 2021-11-16 东北大学 一种风力发电机的异常识别方法
CN109345113B (zh) * 2018-09-29 2021-03-02 北京拾易技术有限公司 风电机组性能评估方法及介质
CN109667727B (zh) * 2018-11-18 2020-01-10 浙江大学 一种基于功率曲线分析的风电机组偏航误差固有偏差辨识及补偿方法
CN109488528B (zh) * 2018-11-23 2020-03-17 华润电力技术研究院有限公司 一种风机偏航系统的调整方法及装置
CN109581269A (zh) * 2018-12-25 2019-04-05 中国电力科学研究院有限公司 一种电子式互感器误差特性快速评价方法及系统
CN110094300B (zh) * 2019-01-24 2020-04-21 上海电力学院 风电机组的对风偏差矫正方法、装置、设备及介质
CN110094310B (zh) * 2019-05-13 2020-05-19 北京天泽智云科技有限公司 一种识别风力发电机偏航对风不正的方法
CN111322205B (zh) * 2019-12-13 2021-11-23 风脉能源(武汉)股份有限公司 风电机组风向标零位误差辨识方法及校正方法
CN113048016A (zh) * 2019-12-27 2021-06-29 新疆金风科技股份有限公司 在线修正风力发电机组的对风偏差的方法和装置
CN111522808B (zh) * 2020-04-29 2023-07-28 贵州电网有限责任公司 一种风电机组异常运行数据处理方法
CN111828253B (zh) * 2020-07-29 2022-02-11 湘电风能有限公司 一种基于风向方位角的风力发电机组保护控制方法
CN112032003B (zh) * 2020-09-01 2021-08-17 浙江运达风电股份有限公司 一种大型风电机组运行性能监测方法
CN112031997B (zh) * 2020-09-15 2021-08-13 华北电力科学研究院有限责任公司 一种风电机组偏航静态偏差确定方法、装置
CN112145373B (zh) * 2020-09-15 2021-09-10 东北大学 一种风力发电机测风仪风速的校正方法
US11231012B1 (en) * 2020-09-22 2022-01-25 General Electric Renovables Espana, S.L. Systems and methods for controlling a wind turbine
CN112443461B (zh) * 2020-10-10 2022-02-08 东北电力大学 基于对称声压传感器的风力发电机组偏航误差分析方法
CN112648139B (zh) * 2020-11-12 2022-03-04 北京金风慧能技术有限公司 风力发电机组对风不正校正方法、装置及控制器
CN112699913B (zh) * 2020-11-25 2023-08-29 国网湖南省电力有限公司 一种台区户变关系异常诊断方法及装置
CN112632773B (zh) * 2020-12-21 2024-04-05 北京华能新锐控制技术有限公司 一种风电机组可靠性预测方法
CN112381185B (zh) * 2021-01-15 2022-03-01 北京工业大数据创新中心有限公司 一种工业设备特性曲线相似度获取方法及装置
CN113309661B (zh) * 2021-03-09 2022-04-26 明阳智慧能源集团股份公司 一种提升风力发电机组发电量的方法
CN113256179B (zh) * 2021-07-06 2021-09-24 大唐环境产业集团股份有限公司 脱硫系统浆液品质在线分类评价方法、系统及装置
CN113757041A (zh) * 2021-08-12 2021-12-07 太原重工股份有限公司 一种风力发电机组风向数据的智能校准方法及系统
CN113653609B (zh) * 2021-09-17 2022-04-19 中节能风力发电股份有限公司 风电机组风向标故障识别方法、系统、设备和存储介质
CN114033631B (zh) * 2021-11-08 2023-08-01 浙江运达风电股份有限公司 一种风电机组风能利用系数在线辨识方法

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103675356A (zh) * 2013-11-19 2014-03-26 中国大唐集团科学技术研究院有限公司 基于粒子群优化的风速仪故障检测方法和系统
US20140199156A1 (en) * 2013-01-17 2014-07-17 Alstom Renovables Espana, S.L. Method of operating a wind turbine
US20150152846A1 (en) * 2013-11-29 2015-06-04 Alstom Renewable Technologies Methods of operating a wind turbine, wind turbines and wind parks
CN104747368A (zh) * 2015-01-27 2015-07-01 风脉(武汉)可再生能源技术有限责任公司 一种风电机组功率优化的方法和系统
CN104989592A (zh) * 2015-07-07 2015-10-21 中能电力科技开发有限公司 风力发电机组机舱风向校正方法
CN105259374A (zh) * 2015-11-25 2016-01-20 江苏天赋新能源工程技术有限公司 风向标零位校正装置
CN105257470A (zh) * 2015-09-25 2016-01-20 南车株洲电力机车研究所有限公司 一种风电机组风向补偿优化方法及装置
CN106014858A (zh) * 2016-07-21 2016-10-12 浙江运达风电股份有限公司 一种风电机组对风误差自动校准方法及装置

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5022102B2 (ja) * 2007-05-25 2012-09-12 三菱重工業株式会社 風力発電装置、風力発電システムおよび風力発電装置の発電制御方法
JP4995209B2 (ja) * 2009-01-05 2012-08-08 三菱重工業株式会社 風力発電装置及び風力発電装置の風向推定方法
CN102213182B (zh) * 2011-05-12 2013-09-04 北京金风科创风电设备有限公司 偏航误差角获得方法及偏航控制方法/装置和风力发电机组
US20150086357A1 (en) * 2013-09-24 2015-03-26 General Electric Company Wind turbine and method for adjusting yaw bias in wind turbine
WO2015048972A1 (en) * 2013-10-01 2015-04-09 Vestas Wind Systems A/S Safe mode operation at high yaw error
US10393093B2 (en) * 2013-11-21 2019-08-27 General Electric Company System and method for assessing the performance impact of wind turbine upgrades
CN104481804B (zh) * 2014-12-05 2017-02-22 北京金风科创风电设备有限公司 风力发电机组对风矫正控制方法、装置和系统
CN107208607A (zh) * 2015-01-28 2017-09-26 Kk风能解决方案公司 校准风力机的风向标
CN105484939A (zh) * 2015-12-16 2016-04-13 大连尚能科技发展有限公司 一种风速风向仪的角度测量误差曲线的替代学习方法
CN105569922A (zh) * 2015-12-16 2016-05-11 大连尚能科技发展有限公司 一种基于风速影响的风速风向仪的角度测量误差补偿方法

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140199156A1 (en) * 2013-01-17 2014-07-17 Alstom Renovables Espana, S.L. Method of operating a wind turbine
CN103675356A (zh) * 2013-11-19 2014-03-26 中国大唐集团科学技术研究院有限公司 基于粒子群优化的风速仪故障检测方法和系统
US20150152846A1 (en) * 2013-11-29 2015-06-04 Alstom Renewable Technologies Methods of operating a wind turbine, wind turbines and wind parks
CN104747368A (zh) * 2015-01-27 2015-07-01 风脉(武汉)可再生能源技术有限责任公司 一种风电机组功率优化的方法和系统
CN104989592A (zh) * 2015-07-07 2015-10-21 中能电力科技开发有限公司 风力发电机组机舱风向校正方法
CN105257470A (zh) * 2015-09-25 2016-01-20 南车株洲电力机车研究所有限公司 一种风电机组风向补偿优化方法及装置
CN105259374A (zh) * 2015-11-25 2016-01-20 江苏天赋新能源工程技术有限公司 风向标零位校正装置
CN106014858A (zh) * 2016-07-21 2016-10-12 浙江运达风电股份有限公司 一种风电机组对风误差自动校准方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3489508A4 *

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109681380A (zh) * 2018-12-29 2019-04-26 新疆金风科技股份有限公司 用于调整最小桨距角位置的方法及装置
CN111145109A (zh) * 2019-12-09 2020-05-12 深圳先进技术研究院 基于图像的风力发电功率曲线异常数据识别与清洗方法
CN113236507A (zh) * 2021-05-28 2021-08-10 中南大学 一种风电机组偏航静态误差诊断方法和系统
CN113236507B (zh) * 2021-05-28 2022-08-19 中南大学 一种风电机组偏航静态误差诊断方法和系统

Also Published As

Publication number Publication date
EP3489508B1 (en) 2022-06-08
CN106014858B (zh) 2019-11-22
ES2926639T3 (es) 2022-10-27
EP3489508A4 (en) 2019-08-07
DK3489508T3 (da) 2022-07-04
CN106014858A (zh) 2016-10-12
EP3489508A1 (en) 2019-05-29

Similar Documents

Publication Publication Date Title
WO2018014825A1 (zh) 一种风电机组对风误差自动校准方法及装置
CN117093879B (zh) 一种数据中心智能化运营管理方法及系统
US10190575B2 (en) Method, device and system for compensating output power of wind turbine generator set
US11644009B2 (en) Method and apparatus for detecting yaw-to-wind abnormality, and device and storage medium thereof
US20120230821A1 (en) Wind power prediction method of single wind turbine generator
CN106570790B (zh) 一种计及风速数据分段特性的风电场出力数据修复方法
CN110795690A (zh) 风电场运行异常数据检测方法
CN116522268B (zh) 一种配电网的线损异常识别方法
CN113626767A (zh) 一种风机限电识别方法、系统、设备和存储介质
CN111159900B (zh) 风机风速的预测方法及装置
CN114187464A (zh) 复杂环境下激光雷达与视觉融合的多周期目标识别方法
CN107357714B (zh) 一种基于监控平台的故障分析方法和装置
CN115408860A (zh) 一种风电场风速-功率数据的异常值监测及修正方法
CN113554079B (zh) 一种基于二次检测法的电力负荷异常数据检测方法及系统
CN116151799A (zh) 一种基于bp神经网络的配电线路多工况故障率快速评估方法
CN112965964B (zh) 一种实测飞参数据的野值检测方法、系统及计算机相关产品
CN110826904B (zh) 风机的数据处理方法、装置、处理设备及可读存储介质
CN112632785A (zh) 一种改进非参数核密度估计光伏出力建模方法
Liu et al. The detection method of wind turbine operation outliers based on mutidimensional cluster
CN117093464B (zh) 一种用户进线实时监控预警方法及系统
CN112800103B (zh) 一种scada数据挖掘的风电机组功率曲线的获取系统及方法
CN117649042A (zh) 火电厂机组agc性能在线评价方法及系统
CN117189509A (zh) 一种海上风机雷击故障检测方法
CN115438310A (zh) 新能源场站的功率确定方法及装置
CN117522220A (zh) 一种基于仪表日志的仪表管理方法及系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17830459

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2017830459

Country of ref document: EP

Effective date: 20190221