WO2019114160A1 - 结冰预测方法、装置、模型生成方法及装置 - Google Patents

结冰预测方法、装置、模型生成方法及装置 Download PDF

Info

Publication number
WO2019114160A1
WO2019114160A1 PCT/CN2018/082514 CN2018082514W WO2019114160A1 WO 2019114160 A1 WO2019114160 A1 WO 2019114160A1 CN 2018082514 W CN2018082514 W CN 2018082514W WO 2019114160 A1 WO2019114160 A1 WO 2019114160A1
Authority
WO
WIPO (PCT)
Prior art keywords
icing
historical
unit
data
information
Prior art date
Application number
PCT/CN2018/082514
Other languages
English (en)
French (fr)
Inventor
周杰
敖娟
王青天
Original Assignee
北京金风科创风电设备有限公司
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 北京金风科创风电设备有限公司 filed Critical 北京金风科创风电设备有限公司
Priority to US16/338,808 priority Critical patent/US20210332796A1/en
Priority to EP18863806.8A priority patent/EP3524813B1/en
Publication of WO2019114160A1 publication Critical patent/WO2019114160A1/zh

Links

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
    • F03D17/00—Monitoring or testing of wind motors, e.g. diagnostics
    • 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
    • F03D80/00—Details, components or accessories not provided for in groups F03D1/00 - F03D17/00
    • F03D80/40—Ice detection; De-icing means
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N5/00—Computing arrangements using knowledge-based models
    • G06N5/04—Inference or reasoning models
    • 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
    • F03D7/046—Automatic control; Regulation by means of an electrical or electronic controller characterised by the type of control logic with learning or adaptive control, e.g. self-tuning, fuzzy logic or neural network
    • 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
    • F05B2260/00—Function
    • F05B2260/82—Forecasts
    • 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
    • F05B2260/00—Function
    • F05B2260/82—Forecasts
    • F05B2260/821—Parameter estimation or prediction
    • F05B2260/8211—Parameter estimation or prediction of the weather
    • 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
    • F05B2260/00—Function
    • F05B2260/83—Testing, e.g. methods, components or tools therefor
    • 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
    • F05B2260/00—Function
    • F05B2260/84—Modelling or simulation
    • 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
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A30/00—Adapting or protecting infrastructure or their operation
    • 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

Definitions

  • the present application relates to the field of wind power generation technologies, and in particular, to an icing prediction method and apparatus, and an icing prediction model generation method and apparatus.
  • the embodiment of the present application provides a method, a device, a storage medium, and a method and a device for generating an icing prediction model of a wind power generator.
  • an embodiment of the present application provides a method for predicting icing of a wind power generator, the method comprising:
  • the icing prediction model In response to the input, the icing prediction model outputs an icing prediction result.
  • an embodiment of the present application provides a device for predicting icing of a wind power generator, the device comprising:
  • An information extraction unit configured to extract, according to geographic information of the target unit, an effective meteorological prediction data feature of the target unit
  • An information input unit configured to input an effective weather prediction data feature into an icing prediction model for predicting icing information
  • An information output unit for responding to the input, and the icing prediction model outputs an icing prediction result.
  • an embodiment of the present application provides an apparatus for predicting icing of a wind power generator, the apparatus comprising: at least one processor, at least one memory, and a computer program stored in the memory, when the computer program is executed by the processor A method as in the first aspect of the above embodiment is implemented.
  • an embodiment of the present application provides a computer readable storage medium having stored thereon a computer program, and when the computer program is executed by the processor, implements the method of the first aspect of the foregoing embodiment.
  • an embodiment of the present application provides a method for generating an icing prediction model of a wind power generator, the method comprising:
  • the historical weather prediction data feature and the historical icing flag are used as input information to establish an icing prediction model for outputting predicted icing information.
  • the embodiment of the present application provides a device for generating an icing prediction model of a wind power generator set, where the device includes:
  • An information acquiring unit configured to acquire geographic information and historical icing flag information of the target unit in the target wind field
  • a feature acquisition unit configured to acquire historical meteorological prediction data features corresponding to geographic information of the target unit
  • a model establishing unit is configured to use the historical weather prediction data feature and the historical icing flag as input information to establish an icing prediction model for outputting predicted icing information.
  • the embodiment of the present application can accurately extract the effective weather prediction data characteristics of the target unit at different geographical locations based on the geographic information of the target unit, and input the accurate effective weather prediction data features of the target unit into the icing prediction model, and the accurate output is different for different The icing prediction result of the target unit at the geographic location. Therefore, the embodiment of the present application can accurately predict the icing information of the target unit, and obtain valuable time for preventing the operation and maintenance of the icing of the unit, thereby avoiding the loss caused by the passive treatment of icing.
  • FIG. 1 is a schematic diagram of a method for generating an icing prediction model of a wind power generator set according to an embodiment of the present application
  • FIG. 2 is a schematic diagram of acquiring historical weather prediction data features according to an embodiment of the present application.
  • FIG. 3 is a schematic diagram of a method for generating an icing prediction model of a wind power generator set according to another embodiment of the present application;
  • FIG. 4 is a schematic diagram of a method for predicting icing of a wind power generator set according to an embodiment of the present application
  • FIG. 5 is a schematic structural diagram of an apparatus for generating an icing prediction model of a wind power generator according to an embodiment of the present application
  • FIG. 6 is a schematic structural diagram of an apparatus for predicting icing of a wind power generator set according to an embodiment of the present application
  • FIG. 7 is a schematic structural diagram of an apparatus for predicting icing of a wind power generator set according to an embodiment of the present application.
  • FIG. 1 is a schematic diagram of a method of generating an icing prediction model of a wind power generator set according to an embodiment of the present application.
  • the method for generating an icing prediction model of a wind power generator may include the following steps: S110: acquiring historical weather prediction data features; S120, acquiring historical icing flag information; S130, based on historical weather prediction data characteristics and The historical icing flag information is used for model training; S140, the training is obtained by the icing prediction model.
  • historical weather prediction data features may be extracted from global weather data.
  • the acquisition manner of the unit historical icing flag information may include the following two methods:
  • the first way obtained by the icing sensor hardware feedback provided by the unit.
  • the second way based on the icing mechanism analysis unit operation data obtained.
  • the blade icing can change the wing shape, the blade resistance increases, the lift decreases, and the blade enters the stall state. At this time, the generating capacity of the unit decreases, directly in the wind speed and The power relationship is abnormal.
  • the basic discriminant formula for the relationship anomaly can be as shown in the following formula 1:
  • the product of 1/2 and ⁇ can be the coefficient threshold, and its value can be 0. ⁇ 1/2 between.
  • the coefficient threshold can be obtained after a large amount of experimental data.
  • Equation 1 when the actual power generation is low, Equation 1 is established and maintained for a preset time (eg, 2 min), if the ambient temperature meets the icing condition and there is no other power abnormality warning, the unit icing flag position may be 1, that is, the state of the unit is icing.
  • the trigger condition for obtaining the unit icing flag may be adjusted according to the degree of tolerance to blade icing or other factors.
  • the icing flag information can be represented by a Boolean value of 0 or 1. For example, when the icing flag is 1, it indicates that the wind turbine has frozen. When the icing flag is 0, it means that the wind turbine is not frozen.
  • the icing flag information (abbreviated as a flag bit or a tag) may also be shown by an analog quantity (eg, the flag bit is 0.1 when ⁇ is 0.65 in Equation 1, and 0.2 when it is 0.60, when 0.55 is 0.55). The flag is 0.3, and so on.)
  • the size of the analog can represent the difference in thickness of the icing.
  • the second method of generating the flag bit may be: when the ice is caused to cause the blade to stall, the relationship between the wind speed and the impeller speed is abnormal to generate the flag position; or using machine learning
  • the method uses multivariate as the marker bit to generate the input of the model, and then trains to obtain the icing detection model to output the icing flag.
  • the condition of the model training may include: acquiring weather forecast data features of each unit in a certain time period of the target wind field history and icing flag information of the unit in the same time period.
  • Model training based on historical meteorological prediction data characteristics and historical icing flag information (referred to as sample data) can be a supervised learning. Sample data can be divided into training data sets and test data sets. The training data set can be the historical meteorological forecast data characteristics of half of the target wind farms, and the corresponding historical icing flag information of these units.
  • the model tree can be used in the R language decision tree method, as shown in the following formula 2:
  • model training can employ a regression approach.
  • the model training may first perform feature screening and then perform multi-feature dimensionality reduction processing.
  • model training may select units with a higher number of freezes to improve sample imbalance during training.
  • model training can be trained using random forest or other machine learning algorithms.
  • the purpose of model training may be to obtain an icing prediction model based on features (such as historical weather prediction data features) and flag bits.
  • the model is trained in the Python language or the Matlab language or other modeling language, and the expression of the above formula 2 will also change accordingly.
  • the model testing method may perform model testing through a randomly selected training data set, a test data set, or may perform model testing by using a k-fold cross-validation method.
  • an icing prediction model can be obtained by the above formula 2.
  • the sample used for model training may include a plurality of data sets, such as a data set M composed of original data (which may include weather data original features) and a processed data set N processed by the data set M.
  • the processing may include: data deletion processing and data unified format processing.
  • each data set may include one or more subsets of data.
  • data set M may include subset M'
  • data set N may include subset N' or the like.
  • the establishment of the icing prediction model can be completed by searching for the test set data of the other half of the unit.
  • the input of the icing prediction model may include: a subset M' including the original features (raw data) of the meteorological data, and a subset N' obtained by processing the original variables. Since the training data set and the test data set in this embodiment are targeted to the entire wind field, the icing prediction model can output the icing flag information of each unit of the entire wind field.
  • 2 is a schematic diagram of acquiring historical weather prediction data features according to an embodiment of the present application.
  • acquiring the historical weather prediction data feature may include the following steps: S201: acquiring historical global meteorological data; S202, acquiring geographic information of each unit (wind turbine) of the target wind farm; S203, from history Extracting data from global meteorological data; S204, acquiring historical meteorological prediction data features.
  • the historical global meteorological data may be simulated from a weather forecast for the target wind field using a weather mode during a certain period of time (last year of the twelfth month).
  • the weather condition of the wind farm can be numerically predicted according to the topography and climatic characteristics of the area where the wind power is located, an appropriate parameterization scheme, an appropriate simulation range, and a nesting manner.
  • a suitable parameterization scheme for a certain area it is necessary to consider the local weather characteristics to make corresponding selections. For example, because convection is relatively strong in high-altitude terrain and high altitude areas, the micro-physical and cumulus convective parameterization process plays an important role in simulation accuracy.
  • the planetary boundary layer parameterization scheme is more important.
  • the surface parameterization scheme is more important; in addition, microphysics, turbulence, diffusion, long-wave radiation, short-wave radiation, etc. are also important parametric options. Since there are hundreds of parameterization options within the meteorological mode, there is no more detail here.
  • parameterization scheme combinations can be used for forecasting. These combined schemes are combined with local weather characteristics, and are summarized by multiple simulations and compared with measured data.
  • the above describes several main principles for the selection of parameterization schemes, which can be selected based on this principle and combined with the experience of historical simulation data in the region.
  • the range of simulation and the way of nesting are mainly determined by factors such as the size of the wind farm and the resolution of the final data required.
  • the final target wind field numerical prediction variable can reach more than one hundred kinds, for example, it can include latitude and longitude, horizontal and vertical wind speed, disturbed dry air quality, disturbance air pressure, surface heat flux, temperature, water and gas mixing rate, and the like.
  • the modes employed may employ commonly used mesoscale numerical models, such as the Weather Research and Forecasting Model (WRF), mesoscale non-hydrostatic forces. Mode (Mesoscale Model version 5, MM5), Regional Atmospheric Modeling System (RAMS), etc.
  • WRF Weather Research and Forecasting Model
  • MRS Mesoscale non-hydrostatic forces.
  • RAMS Regional Atmospheric Modeling System
  • other numerical models with simulation or forecasting capabilities such as various climate models, ocean modes, ocean-air coupled models, etc., or various predictive models such as linear regression, multiple regression, artificial neural networks, and support vectors can be used. Machine, Bayes, etc., to simulate or forecast global meteorological data.
  • the geographic information may be from the wind farm archive information, and the geographic information of the crew may be acquired according to any other means such as satellite image recognition, drone technology, calibration technology of relative positions between the units, and the like. Geographic information can be refined to the latitude and longitude and altitude of each unit.
  • the typical meteorological characteristics associated with the target wind farm icing are related to the topographic and climatic characteristics of the target wind farm. Therefore, it is first necessary to obtain typical meteorological features related to the icing of the target wind farm based on the topographic and climatic characteristics of the target wind farm, and then extract the global meteorological data of the wind farm based on typical meteorological characteristics.
  • the frost-related weather forecast data ie, historical meteorological forecast data characteristics
  • the method of extracting data can adopt inverse distance weighted interpolation method, improved Cheyder method, bilinear interpolation method, natural neighbor interpolation method, moving average method and the like.
  • the historical weather prediction data features may include: raw data and processing data.
  • the original data may be the original feature of the meteorological data, and may be specifically defined as a set M, including features M1, M2, ..., Mm.
  • the processed data may be a feature set N processed by the original data, including features N1, N2, ..., Nn.
  • the method for acquiring meteorological prediction data features may be related to the knowledge of the icing phenomenon by the technician, or may be related to objective conditions such as different models, weather conditions, and terrain conditions. Differences in these factors can result in set M, set N, and even subset M', and subset N' may be inconsistent in different target wind fields.
  • acquiring historical meteorological prediction data characteristics of the target unit may include: acquiring terrain characteristics and climate characteristics of the target wind field where the target unit is located; acquiring historical global meteorological data of the target wind field; and based on terrain characteristics and climate characteristics, Historical meteorological prediction data features are extracted from historical global meteorological data.
  • the above embodiments can predict the icing information of the target unit actively and accurately by the effective weather prediction data characteristics of the units at different geographical locations obtained through the simulation training, and obtain valuable for preventing the operation and maintenance of the unit icing. Time to avoid the loss of passive handling of icing.
  • Fig. 3 is a schematic diagram showing a method of generating an icing prediction model of a wind power generator set according to another embodiment of the present application.
  • the method for generating an icing prediction model of a wind power generator may include the following steps: S310, acquiring geographic information of a target unit in the target wind field and historical icing flag information; S320, acquiring target unit and geography The historical weather prediction data feature corresponding to the information; S330, using the historical weather prediction data feature and the historical icing flag as the input information, and establishing an icing prediction model for outputting the predicted icing information.
  • acquiring the historical weather prediction data feature corresponding to the geographic information of the target unit in the preset time period may include: acquiring terrain feature data and climate characteristic data of the target unit based on the geographic information; Obtain historical global meteorological data of the target wind field within a preset time period; extract historical meteorological prediction data features from historical global meteorological data based on terrain characteristic data and climate characteristic data.
  • obtaining historical icing flag information of the target unit may include: determining a target unit based on a coefficient threshold and a wind energy utilization coefficient, an air density, an airflow area of the impeller, and a wind speed corresponding to the target unit within a preset time period.
  • the historical reference power generation obtain the historical actual power generation of the target unit monitored in the preset time period; compare the historical actual power generation with the historical reference power generation, and determine the historical icing flag information of the target unit according to the comparison result.
  • the comparison result may include: the historical actual power generation power is less than the historical reference power generation power, and the historical actual power generation power is greater than or equal to the historical reference power generation power.
  • the historical actual power generation When the historical actual power generation is less than the historical reference power generation, it indicates that the target unit power is degraded due to icing. At this time, the historical icing flag is marked to indicate that the icing has been frozen.
  • the historical reference power generation can be determined as a product of a coefficient threshold, a wind energy utilization factor, an air density, an impeller swept area, and a wind speed.
  • the coefficient threshold can be between 0 and 1/2.
  • obtaining historical icing flag information of the target unit may include: acquiring icing sensing data collected by a sensor disposed on the target unit; and obtaining a historical icing flag of the target unit based on the icing sensing data Bit information.
  • the historical weather prediction data feature may include: raw data in the historical global weather data, and/or processing data obtained by processing the original data. Therefore, the embodiment of the present application can use the original data and the processing data as the historical weather prediction data features, can make up for the problem of poorly targeted only by using the original data, and can also make up for the problem of large deviation caused by only using the processing data. Predictive accuracy provides a data foundation that ensures accurate prediction of icing information later.
  • the historical weather prediction data feature includes one or more of the following parameters: microphysical parameters, cumulus convection parameters, planetary boundary layer parameters, surface parameters, turbulence parameters, diffusion parameters, radio wave radiation parameters.
  • the historical weather prediction data feature may be all weather data that is finally output or intermediately output in the numerical mode.
  • Historical meteorological prediction data features may also select parameters other than microphysical parameters, cumulus convective parameters, planetary boundary layer parameters, surface parameters, turbulence parameters, diffusion parameters, and radio wave radiation parameters.
  • generating the icing prediction model may include: determining historical weather prediction data features and historical icing flag information as sample data; dividing the sample data into a training data set and a test data set; and training training based on the machine supervised learning method
  • the data set obtains the basic meteorological preset data characteristics; the test data set is used to test the basic meteorological preset data features, and the icing prediction model based on the effective weather forecast data feature prediction icing information is obtained. Therefore, the embodiment of the present application can perfectly combine the prediction and the test through the training data set and the test data set, and can further improve the accuracy of the icing prediction.
  • the method for generating a model may further include: acquiring terrain characteristic data and/or climate characteristic data of each target unit in the target wind field; and characterizing the historical weather prediction data based on the terrain characteristic data and/or the climate characteristic data.
  • the clustering is performed; according to the historical icing flag information and the historical weather forecast data characteristics of the cluster, the icing prediction model based on the effective weather forecasting data feature prediction icing information is trained. Therefore, the embodiment of the present application can make the training more specific through the method of data clustering, not only can improve the precision of the training, thereby improving the accuracy of the icing prediction, and can reduce the amount of data calculation, reduce the calculation overhead, and save training. time.
  • clustering should be performed first, and then different categories are modeled in the manner of the above embodiment.
  • part of the time period data should be taken as the training set data from the time dimension, and another part of the time period data is modeled as the test set data.
  • the simulated area can be All the units in the cluster are clustered.
  • the principles of clustering are terrain conditions, local meteorological conditions, terrain conditions such as altitude, roughness, ruggedness index, etc. Local meteorological conditions such as convective intensity, atmospheric stability, turbulence intensity, wind speed , wind direction, etc., and then according to different categories, modeling in the manner of the above embodiment.
  • the icing prediction model established in this case can be provided for use by multiple target wind farms within the scope of the numerical weather prediction.
  • the method for generating a model may further include: acquiring an operating parameter of the target unit; and based on the operating parameter, the historical weather prediction data feature, and the historical icing flag information, training to obtain icing information based on the feature of the effective weather prediction data.
  • the icing prediction model Therefore, the embodiment of the present application can obtain a predictive model that is highly matched with the target unit through the targeted training of the operating parameters of the target unit, historical weather prediction data characteristics, and historical icing target information, thereby greatly improving the icing prediction. Precision.
  • the operating parameters may include: blade speed, pitch angle, power generation.
  • FIG. 4 is a schematic diagram of a method for predicting icing of a wind power generator set according to an embodiment of the present application.
  • the wind turbine icing prediction method of the wind turbine may include the following steps: S410, extracting effective meteorological prediction data characteristics of the target unit based on the geographic information of the target unit; S420, the characteristics of the effective weather prediction data An icing prediction model for predicting icing information is input; S430, response input, and icing prediction model output icing prediction results.
  • extracting the effective meteorological prediction data feature of the target unit based on the geographic information of the target unit may include: acquiring global meteorological data of the target wind farm where the target unit is located; and acquiring each unit of the target wind farm. Geographic information; according to the geographic information of each unit, the characteristics of effective meteorological prediction data related to icing of each unit are extracted from the global meteorological data.
  • the valid weather prediction data features may include: meteorological data original features and/or processed features processed by weather data original features.
  • the method for extracting the effective meteorological prediction data features of the target unit includes one or more of the following methods: inverse distance weighted interpolation, improved Shepherd method, bilinear interpolation, natural neighbor interpolation Moving average method.
  • acquiring global meteorological data of the target wind farm where the target unit is located includes: acquiring terrain characteristics and climatic characteristics of the target wind farm; and determining global meteorological data including the parameterized scheme according to the topographical features and the climatic characteristics, Numerical weather prediction for the target wind farm.
  • the method for wind turbine icing prediction may further include: acquiring historical meteorological prediction data features of the target unit and historical icing flags in advance; and using historical weather prediction data features and historical icing flags as input information. Establish an icing prediction model for outputting predicted icing information.
  • establishing an icing prediction model for outputting predicted icing information may include dividing the input information into a training data set and test data. Set; training training data set based on machine supervised learning method, and testing with test data set to obtain test results; establishing icing prediction model based on test results.
  • dividing the input information into the training data set and the test data set may include: when the target unit is multiple, clustering the target units based on the target unit's terrain conditions and/or meteorological conditions to generate the first The class unit and the second type unit; the input information of the first type unit is determined as the training data set; and the input information of the second type unit is determined as the test data set.
  • dividing the input information into the training data set and the test data set may include: when the target unit is one, clustering the input information of the target unit of the first time period into a training data set, and The input information of the target unit is clustered into a test data set.
  • the data set is more targeted and more accurate, and can prepare for accurate icing prediction of any number of fans in any subsequent wind farm.
  • acquiring historical meteorological prediction data characteristics of the target unit may include: acquiring terrain characteristics and climate characteristics of the target wind field where the target unit is located; acquiring historical global meteorological data of the target wind field; and based on terrain characteristics and climate characteristics, Historical meteorological prediction data features are extracted from historical global meteorological data.
  • obtaining a historical icing flag of the target unit includes: determining whether the temperature and humidity meet the icing condition; determining whether the operating parameter of the target unit is abnormal when the temperature and the humidity meet the icing condition; When the operating parameters of the unit are abnormal, the historical icing flag of the target unit is obtained.
  • the icing prediction model of the target unit is used to predict the icing information of the target unit.
  • the application scenario of predicted icing information (such as icing flags) can be used to display in the wind farm owner monitoring system.
  • the predicted icing information may be an indication of whether the target unit is frozen after a preset time (eg, after 6 hours) (eg, the icing flag is 1, the icing flag is 0, and the ice sign is very thin)
  • the position is 0.1, the thick ice mark is 0.9), and it can also be the icing index curve of the target unit.
  • the predicted icing information can be output to more systems and incubate corresponding services, such as core variables of the icing prediction service of the regional multi-wind farm scheduling system; and as an input to the dynamic operation and maintenance decision system Application scenarios such as variables.
  • the execution body of the foregoing method may be a processor, a controller, or the like.
  • those skilled in the art can flexibly adjust the sequence of the above-mentioned operation steps according to actual needs, or perform flexible operations such as the above steps.
  • various implementations will not be described again. Additionally, the contents of the various embodiments may be referenced to each other.
  • FIG. 5 is a schematic structural diagram of an apparatus for generating an icing prediction model of a wind power generator according to an embodiment of the present application.
  • the apparatus 500 may include an information acquisition unit 501, a feature acquisition unit 502, and a model establishment unit 503.
  • the information acquiring unit 501 may be configured to acquire geographic information and historical icing flag information of the target unit in the target wind field;
  • the feature acquiring unit 502 may be configured to acquire historical weather forecast data features corresponding to the geographic information of the target unit;
  • the establishing unit 503 can be configured to use the historical weather prediction data feature and the historical icing flag as input information to establish an icing prediction model for outputting the predicted icing information.
  • the information acquiring unit 501 may be configured to: acquire terrain feature data and climate characteristic data of the target unit based on the geographic information; acquire historical global meteorological data of the target wind field within a preset time period; And climate characteristic data, extracting historical meteorological prediction data features from historical global meteorological data.
  • the historical weather prediction data features may include raw data in historical global meteorological data, and/or processed data obtained by processing the raw data.
  • the historical weather prediction data feature includes one or more of the following parameters: microphysical parameters, cumulus convective parameters, planetary boundary layer parameters, surface parameters, turbulence parameters, diffusion parameters, radio wave radiation parameters.
  • the information acquiring unit 501 is further configured to: determine a historical reference of the target unit based on the coefficient threshold and the wind energy utilization coefficient, the air density, the airflow area of the impeller, and the wind speed corresponding to the target unit within the preset time period. Power generation power; obtain the historical actual power generation of the target unit monitored during the preset time period; compare the historical actual power generation with the historical reference power generation, and determine the historical icing flag information of the target unit based on the comparison result.
  • the information obtaining unit 501 is further configured to: determine the historical reference power generation power as a product of a coefficient threshold, a wind energy utilization coefficient, an air density, an impeller swept area, and a wind speed.
  • the coefficient threshold can be between 0 and 1/2.
  • the information obtaining unit 501 is further configured to: acquire icing sensing data collected by a sensor disposed on the target unit; and obtain historical icing flag information of the target unit based on the icing sensing data.
  • the model establishing unit 503 is further configured to: determine historical weather prediction data features and historical icing flag information as sample data; divide the sample data into a training data set and a test data set; The method trains the training data set to obtain the basic meteorological preset data features; uses the test data set to test the basic meteorological preset data features, and obtains the icing prediction model based on the effective weather prediction data feature prediction icing information.
  • FIG. 6 is a schematic structural diagram of an apparatus for predicting icing of a wind power generator set according to an embodiment of the present application.
  • the apparatus 600 may include an information extracting unit 601, an information input unit 602, and an information output unit 603.
  • the information extracting unit 601 may be configured to extract valid weather forecast data features of the target unit based on geographic information of the target unit; the information input unit 602 may be configured to input valid weather forecast data features for predicting icing prediction of the icing information.
  • the model; the information output unit 603 can be used to respond to the input, and the icing prediction model outputs the icing prediction result.
  • the real-time weather prediction data features are calculated using the icing prediction model of FIG. 6 to generate predicted icing prediction flag bits.
  • the computing environment of the icing prediction model can be the central monitoring system of the wind farm owner, or the remote monitoring center computer, or other machines with computing power.
  • the icing prediction model can output an ice icing index curve after 6 hours in the wind farm owner monitoring system.
  • the apparatus 600 can further include: a feature clustering unit.
  • the feature clustering unit may be configured to: acquire terrain feature data and/or climate characteristic data of each target unit in the target wind field; and cluster historical weather prediction data features based on terrain characteristic data and/or climate characteristic data;
  • the icing target information and the historical weather forecast data characteristics of the clustering are trained to obtain an icing prediction model based on the effective weather forecast data to predict the icing information.
  • the apparatus 600 can also include an optimization training unit.
  • the optimization training unit can be used to: obtain the operating parameters of the target unit; and based on the operating parameters, historical weather prediction data features, historical icing flag information, and train an icing prediction model based on the effective weather prediction data feature prediction icing information.
  • the operating parameters may include: blade speed, pitch angle, power generation.
  • Each unit in each of the above embodiments 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 above integrated unit can be implemented in the form of hardware or in the form of a software program unit.
  • FIG. 7 is a schematic structural diagram of an apparatus for predicting icing of a wind power generator set according to an embodiment of the present application.
  • the framework may include a central processing unit (CPU) 701 that may be loaded into a program in a random access memory (RAM) 703 according to a program stored in a read only memory (ROM) 702 or from a storage portion 708.
  • CPU central processing unit
  • RAM random access memory
  • ROM read only memory
  • the framework may include a central processing unit (CPU) 701 that may be loaded into a program in a random access memory (RAM) 703 according to a program stored in a read only memory (ROM) 702 or from a storage portion 708.
  • RAM random access memory
  • ROM read only memory
  • the various operations performed by the embodiments of Figures 1, 2, 3, and 4 are performed.
  • RAM 703 various programs and data required for system architecture operations are also stored.
  • the CPU 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704.
  • An input/output (I/O) interface 705 is also coupled to bus 704.
  • the following components are connected to the I/O interface 705: an input portion 706 including a keyboard, a mouse, etc.; an output portion 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), and the like, and a speaker; a storage portion 708 including a hard disk or the like And a communication portion 709 including a network interface card such as a LAN card, a modem, or the like.
  • the communication section 709 performs communication processing via a network such as the Internet.
  • Driver 710 is also connected to I/O interface 705 as needed.
  • a removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory or the like, is mounted on the drive 710 as needed so that a computer program read therefrom is installed into the storage portion 708 as needed.
  • the processes described above with reference to the respective figures may be implemented as a computer program.
  • the computer program can be downloaded and installed from the network via communication portion 709, and/or installed from removable media 711.
  • the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, ie may be located A place, or it can be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the embodiment. Those of ordinary skill in the art can understand and implement without deliberate labor.

Landscapes

  • Engineering & Computer Science (AREA)
  • General 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)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computational Linguistics (AREA)
  • Wind Motors (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

公开了一种结冰预测方法、装置、结冰预测模型生成方法及装置。其中，结冰预测方法包括：基于目标机组的地理信息，提取目标机组的有效气象预测数据特征；将有效气象预测数据特征输入用于预测结冰信息的结冰预测模型；响应输入，结冰预测模型输出结冰预测结果。

Description

结冰预测方法、装置、模型生成方法及装置 技术领域
本申请涉及风力发电技术领域，尤其涉及一种结冰预测方法、装置、结冰预测模型生成方法及装置。
背景技术
机组结冰的问题一直是风电行业的传统问题，其成为了影响风电场收益的重要因素。结冰严重时，风机叶片承受着很大的冰载，叶片寿命、机组安全运行受到了很大威胁。
因此，如何对各个风力发电机组的结冰情况进行准确预测，成为目前亟待解决的技术问题。
发明内容
本申请实施例提供了一种风力发电机组结冰预测的方法、装置、存储介质、风力发电机组的结冰预测模型的生成方法及装置。
第一方面，本申请实施例提供了一种风力发电机组结冰预测的方法，该方法包括：
基于目标机组的地理信息，提取目标机组的有效气象预测数据特征；
将有效气象预测数据特征输入用于预测结冰信息的结冰预测模型；
响应输入，结冰预测模型输出结冰预测结果。
第二方面，本申请实施例提供了一种风力发电机组结冰预测的装置，该装置包括：
信息提取单元，用于基于目标机组的地理信息，提取目标机组的有效气象预测数据特征；
信息输入单元，用于将有效气象预测数据特征输入用于预测结冰信息的结冰预测模型；
信息输出单元，用于响应输入，结冰预测模型输出结冰预测结果。
第三方面，本申请实施例提供了一种风力发电机组结冰预测的装置，该装置包括：至少一个处理器、至少一个存储器以及存储在存储器中的计算机程序，当计算机程序被处理器执行时实现如上述实施方式中第一方面的方法。
第四方面，本申请实施例提供了一种计算机可读存储介质，其上存储有计算机程序，当计算机程序被处理器执行时实现如上述实施方式中第一方面的方法。
第五方面，本申请实施例提供了一种风力发电机组的结冰预测模型的生成方法，该方法包括：
获取目标风场中目标机组的地理信息和历史结冰标志位信息；
获取目标机组的与地理信息对应的历史气象预测数据特征；
将历史气象预测数据特征以及历史结冰标志位作为输入信息，建立用于输出预测结冰信息的结冰预测模型。
第六方面，本申请实施例提供了一种风力发电机组的结冰预测模型的生成装置，该装置包括：
信息获取单元，用于获取目标风场中目标机组的地理信息和历史结冰标志位信息；
特征获取单元，用于获取目标机组的与地理信息对应的历史气象预测数据特征；
模型建立单元，用于将历史气象预测数据特征以及历史结冰标志位作为输入信息，建立用于输出预测结冰信息的结冰预测模型。
本申请实施例可以基于目标机组的地理信息，精确提取不同地理位置处的目标机组的有效气象预测数据特征，将针对目标机组的精确的有效气象预测数据特征输入结冰预测模型，精确输出针对不同地理位置处的目标机组的结冰预测结果。由此，本申请实施例可以高精度地预测目标机组的结冰信息，为预防机组结冰的运维工作争取到宝贵的时间，从而可以避免被动处理结冰所造成的损失。
附图说明
为了更清楚地说明本申请实施例的技术方案，下面将对本申请实施例中所需要使用的附图作简单地介绍，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，还可以根据这些附图获得其他的附图。
图1是本申请一实施例的风力发电机组的结冰预测模型的生成方法的示意图；
图2是本申请一实施例的获取历史气象预测数据特征的示意图；
图3是本申请另一实施例的风力发电机组的结冰预测模型的生成方法的示意图；
图4是本申请一实施例的风力发电机组结冰预测的方法的示意图；
图5是本申请一实施例的风力发电机组的结冰预测模型的生成装置的结构示意图；
图6是本申请一实施例的风力发电机组结冰预测的装置的结构示意图；
图7是本申请一实施例的风力发电机组结冰预测的装置的结构示意图。
具体实施方式
图1是本申请一实施例的风力发电机组的结冰预测模型的生成方法的示意图。
如图1所示，风力发电机组的结冰预测模型的生成方法可以包括以下步骤：S110，获取历史气象预测数据特征；S120，获取历史结冰标志位信息；S130，基于历史气象预测数据特征和历史结冰标志位信息进行模型训练；S140，训练得到结冰预测模型。
在S110中，可以从全局气象数据中提取历史气象预测数据特征。
在S120中，机组历史结冰标志位信息的获取方式可以包括以下两种方式：
第一种方式：由机组配备的结冰传感器硬件反馈获得。
第二种方式：基于结冰机理分析机组运行数据获得。
针对第二种方式，根据机组结冰的机理可知：机组叶片覆冰会使翼形 发生改变，叶片的阻力上升，升力下降，叶片进入失速状态，此时机组发电能力下降，直接表现在风速与功率的关系异常。关系异常的基本判别公式可以如下面公式1所示：
Figure PCTCN2018082514-appb-000001
在上述公式1中，Pr为实际发电功率(kW)；Cp为风能利用系数；ρ为空气密度，取1.23kg/m 3或由海拔高度及环境温度进行估算；A为根据叶轮直径计算的扫风面积；V为风速；1/2为系数；α为阈值，α取值在0～1之间，典型值可以如0.65。1/2与α的乘积可以为系数阈值，其数值可以在0～1/2之间。该系数阈值可以是经过大量的实验数据得到。
在一些实施例中，当实际发电功率较低，公式1成立并保持预设时间(如2min)时，如果环境温度满足结冰条件且无其他功率异常的警告，则可将机组结冰标志位置1，即机组的状态是结冰状态。
在一些实施例中，利用以上公式1获得机组结冰标志位时，获得机组结冰标志位的触发条件可以根据对叶片结冰的容忍程度或其他因素调整。
在一些实施例中，结冰标志位信息可以用布尔量0或者1表示。例如，当结冰标志位为1时，表示风力发电机组已经结冰。当结冰标志位为0时，表示风力发电机组没有结冰。
在一些实施例中，结冰标志位信息(简称标志位或者标签)也可以用模拟量示出(如式1中α为0.65时标志位为0.1，为0.60时标志位为0.2，为0.55时标志位为0.3，以此类推)。模拟量的大小可以表示结冰的薄厚差异。
在一些实施例中，除了上述风速与功率比较的方法以外，标志位生成的第二种方法可以为：结冰导致叶片失速时，风速与叶轮转速关系存在异常来生成标志位；或者采用机器学习方法将多变量作为标志位生成模型的输入，再经训练得到结冰检测模型以输出结冰标志位。
在S130中，在本实施例中，模型训练的条件可以包括：获取目标风场历史上某一时间段各机组的气象预测数据特征以及相同时间段机组的结冰标志位信息。基于历史气象预测数据特征和历史结冰标志位信息(称为样本数据)进行模型训练可以是一种有监督学习。样本数据可以分为训练数据集和测试数据集。训练数据集可以为目标风场一半机组的历史气象预 测数据特征，以及对应的这些机组的历史结冰标志位信息。模型训练时可以采用R语言中的决策树方法，具体可以如下面公式2所示：
band_Clas＜-rpart(state～.，Train_Ice，method＝″class″，minsplit＝8)(公式2)
在一些实施例中，模型训练可以采用回归的方法。
在一些实施例中，模型训练可以先进行特征筛选，然后进行多特征降维处理。
在一些实施例中，模型训练可以选择结冰次数较多的机组以改善训练过程中样本不均衡问题。
在一些实施例中，模型训练可以采用随机森林或其他机器学习算法进行训练。
在一些实施例中，模型训练的目的可以是获得基于特征(如历史气象预测数据特征)及标志位的结冰预测模型。
在一些实施例中，模型训练时采用Python语言或Matlab语言或其他建模语言，此时上述公式2的表述也会有相应的变化。
在一些实施例中，模型测试方法可以通过随机选取的训练数据集、测试数据集进行模型测试，还可以采用k折交叉验证的方式进行模型测试。
在S140中，由上述公式2可得到结冰预测模型。在本实施例中，模型训练所用样本可以包括多个数据集，如：由原始数据组成的数据集M(可以包括气象数据原始特征)和由数据集M加工处理得到的处理数据集N。其中，加工处理可以包括：数据删选处理和数据统一格式处理等。其中，各个数据集可以包括一个或者多个数据子集。如，数据集M可以包括子集M’，数据集N可以包括子集N’等。
在本实施例中，如果模型训练所用样本是充足的，可以经寻优及另一半机组的测试集数据完成结冰预测模型的建立。结冰预测模型的输入可以包含：包括气象数据原始特征(原始数据)的子集M’，以及经原始变量加工得到的子集N’等。由于本实施例中的训练数据集和测试数据集以整个风场为对象，所以结冰预测模型可以输出整个风场的各个机组的结冰标志位信息。图2是本申请一实施例的获取历史气象预测数据特征的示意图。
如图2所示，获取历史气象预测数据特征(即S110)可以包括以下步 骤：S201，获取历史全局气象数据；S202，获取目标风场各机组(风力发电机组)的地理信息；S203，从历史全局气象数据中提取数据；S204，获取历史气象预测数据特征。
在S201中，历史全局气象数据可以是在历史上某一时间段内(去年腊月)采用气象模式从对目标风场的天气预报中模拟得到的。具体地，可以根据该风电场所处地区的地形及气候特征，选择合适的参数化方案、恰当的模拟范围以及嵌套方式对风电场的天气状况进行数值预报。针对某一地区的合适的参数化方案，需要考虑局地天气特征进行相应的选择。例如，因为在地形复杂、海拔较高的高原地区对流比较旺盛，所以微物理以及积云对流参数化过程对于模拟精度有重要的作用；在天气多变的山区，行星边界层参数化方案比较重要；在沿海或沿湖等水陆交界的地区，地表参数化方案比较重要；除此之外，微物理、湍流、扩散、长波辐射、短波辐射等也是比较重要的参数化方案选项。由于气象模式内部的参数化选项多达上百种，在此不多做赘述。
针对不同地区的风电场，可以采取不同的参数化方案组合进行预报。这些组合方案是结合当地的天气特征，经过多次模拟试验并与实测数据对比效果总结出来的。上述描述的是挑选参数化方案的几个主要原则，可依此原则，并结合该地区的历史模拟数据的经验进行选择。而模拟的范围以及嵌套的方式主要根据风电场范围的大小以及最终所需数据的分辨率等因素来决定。最终得到的目标风场数值预报变量可以达到一百余种，例如，可以包含经纬度、水平及垂直风速、扰动干空气质量、扰动气压、地表热通量、温度、水气混合率等。
在一些实施例中，进行全局气象数据的数值模拟时，所采用的模式可以采用常用的中尺度数值模式，如中尺度天气预报模式(Weather Research and Forecasting Model，WRF)、中尺度非流体静力模式(Mesoscale Model version5，MM5)、区域大气模拟系统(Regional Atmospheric Modeling System，RAMS)等。另外，还可以采用其他具有模拟或预报能力的数值模式，如各种气候模式、海洋模式、海气耦合模式等；或利用各种预测模型，如线性回归、多元回归、人工神经网络、支持向量机、贝叶 斯等，进行全局气象数据的模拟或预报。
在S202中，地理信息可以来自风场存档信息，还可以根据卫星图片识别、无人机技术、机组间相对位置的标定技术等其他任何手段来获取机组的地理信息。地理信息可以细化至每个机组机位点的经纬度及海拔高度。
在S203中，目标风电场结冰相关的典型气象特征与目标风电场的地形与气候特点有关。因此，首先需要根据目标风电场的地形和气候特点得到与该目标风电场结冰相关的典型气象特征，然后根据典型气象特征对风电场的全局气象数据进行提取。
根据目标风场各机组地理经纬度信息可从历史全局气象数据中提取目标风场中各机组的与冰冻相关的气象预报数据(即历史气象预测数据特征)。提取数据的方法可以采用反距离加权插值法、改进谢别德法、双线性插值法、自然邻点插值法、移动平均法等。
在S204中，历史气象预测数据特征可以包括：原始数据和处理数据。其中，原始数据可以是气象数据原始特征，具体可以定义为集M，含特征M1、M2……Mm。处理数据可以是经原始数据加工得到的特征集N，含特征N1、N2……Nn。
在一些实施例中，涉及的气象预测数据特征获取的方法，可以与技术人员对结冰现象的认识相关，也可以与不同机型、气象状况、地形条件等客观条件相关。这些因素的差异可以导致集M、集N乃至子集M’、子集N’在不同目标风场可能不一致。
在一些实施例中，获取目标机组的历史气象预测数据特征可以包括：获取目标机组所在目标风场的地形特性和气候特征；获取目标风场的历史全局气象数据；基于地形特性和气候特征，从历史全局气象数据中提取历史气象预测数据特征。
由此，上述实施例可以通过模拟训练得到的针对不同地理位置处的机组的有效气象预测数据特征主动且高精度地预测目标机组的结冰信息，为预防机组结冰的运维工作争取到了宝贵的时间，从而可以避免被动处理结冰所造成的损失。
图3是本申请另一实施例的风力发电机组的结冰预测模型的生成方法 的示意图。
如图3所示，风力发电机组的结冰预测模型的生成方法可以包括以下步骤：S310，获取目标风场中目标机组的地理信息和历史结冰标志位信息；S320，获取目标机组的与地理信息对应的历史气象预测数据特征；S330，将历史气象预测数据特征以及历史结冰标志位作为输入信息，建立用于输出预测结冰信息的结冰预测模型。
在一些实施例中，获取目标机组在预设时间段内的与地理信息对应的历史气象预测数据特征(即步骤S320)，可以包括：基于地理信息获取目标机组的地形特点数据和气候特点数据；获取在预设时间段内的目标风场的历史全局气象数据；基于地形特点数据和气候特点数据，从历史全局气象数据中提取历史气象预测数据特征。
在S310中，获取目标机组的历史结冰标志位信息，可以包括：基于系数阈值以及预设时间段内的与目标机组对应的风能利用系数、空气密度、叶轮扫风面积、风速，确定目标机组的历史参考发电功率；获取预设时间段内监测的目标机组的历史实际发电功率；比较历史实际发电功率与历史参考发电功率的大小，并根据比较结果确定目标机组的历史结冰标志位信息。在本实施例中，比较结果可以包括：历史实际发电功率小于历史参考发电功率，历史实际发电功率大于等于历史参考发电功率。当历史实际发电功率小于历史参考发电功率时，说明由于结冰导致目标机组功率下降，此时，标记历史结冰标志位，用于说明已经结冰。在一些实施例中，可以将历史参考发电功率确定为：系数阈值、风能利用系数、空气密度、叶轮扫风面积和风速的乘积。
在一些实施例中，系数阈值可以在0至1/2之间。
在一些实施例中，获取目标机组的历史结冰标志位信息可以包括：获取设置于目标机组上的传感器采集的结冰传感数据；基于结冰传感数据，得到目标机组的历史结冰标志位信息。
在S320中，历史气象预测数据特征可以包括：历史全局气象数据中的原始数据，和/或，将原始数据加工得到的处理数据。由此，本申请实施例可以采用原始数据和加工数据作为历史气象预测数据特征，可以弥补仅 采用原始数据导致的针对性差的问题，也可以弥补仅采用加工数据导致的偏差大的问题，为后期预测精度提供了数据基础，保证了后期能够精确预测结冰信息。
在一些实施例中，历史气象预测数据特征包括以下参数中的一种或者多种：微物理参数、积云对流参数、行星边界层参数、地表参数、湍流参数、扩散参数、电波辐射参数。其中，历史气象预测数据特征可以是在数值模式下最终输出或者中间输出的所有气象数据。历史气象预测数据特征还可以选择微物理参数、积云对流参数、行星边界层参数、地表参数、湍流参数、扩散参数、电波辐射参数之外的参数。
在S330中，生成结冰预测模型可以包括：将历史气象预测数据特征和历史结冰标志位信息确定为样本数据；将样本数据分为训练数据集和测试数据集；基于机器监督学习方法训练训练数据集，得到基础气象预设数据特征；利用测试数据集测试基础气象预设数据特征，得到基于有效气象预测数据特征预测结冰信息的结冰预测模型。由此，本申请实施例可以通过训练数据集和测试数据集将预测与测试完美结合，可以进一步提升结冰预测的精度。
在一些实施例中，模型生成的方法还可以包括：获取目标风场中各个目标机组的地形特点数据和/或气候特点数据；基于地形特点数据和/或气候特点数据，将历史气象预测数据特征进行聚类；根据历史结冰标志位信息和聚类的历史气象预测数据特征，训练得到基于有效气象预测数据特征预测结冰信息的结冰预测模型。由此，本申请实施例可以通过数据聚类的方法，使得训练更有针对性，不仅可以提高训练的精度，从而提升结冰预测的精度，而且可以减少数据运算量，减少计算开销，节约训练时间。
在一些实施例中，如果目标风场中机组数目较大且分布较为分散或不同机组地形差别较大，则应先聚类再对不同的类别按上述实施例的方式进行建模。
在一些实施例中，如果目标风场仅有一台机组，则应从时间的维度取一部分时间段数据作为训练集数据，另一部分时间段数据作为测试集数据进行建模。
特别地，由于数值天气预报涵盖的区域通常可达上百公里，若在数值预报的模拟区域内有多个目标风场，每个目标风场内有一台或多台机组，则可对模拟区域内的所有机组进行聚类，聚类的原则为地形条件、局地气象条件等，地形条件如海拔、粗糙度、崎岖指数等，局地气象条件如对流强度、大气稳定度、湍流强度、风速、风向等，再按照不同类别，依照上述实施例的方式进行建模。在此情况下建立的结冰预测模型可以提供给数值天气预报所涵盖范围内的多个目标风场使用。
在一些实施例中，模型生成的方法还可以包括：获取目标机组的运行参数；基于运行参数、历史气象预测数据特征、历史结冰标志位信息，训练得到基于有效气象预测数据特征预测结冰信息的结冰预测模型。由此，本申请实施例可以通过目标机组的运行参数、历史气象预测数据特征、历史结冰标志位信息有针对性的训练，得到与目标机组高度匹配的预测模型，从而大幅度提高结冰预测的精度。
在一些实施例中，运行参数可以包括：叶片转速、桨距角、发电功率。
图4是本申请一实施例的风力发电机组结冰预测的方法的示意图。
如图4所示，风力发电机组的风力发电机组结冰预测的方法可以包括以下步骤：S410，基于目标机组的地理信息，提取目标机组的有效气象预测数据特征；S420，将有效气象预测数据特征输入用于预测结冰信息的结冰预测模型；S430，响应输入，结冰预测模型输出结冰预测结果。
在一些实施例中，基于目标机组的地理信息，提取目标机组的有效气象预测数据特征(S410)，可以包括：获取目标机组所在的目标风电场的全局气象数据；获取目标风电场中各个机组的地理信息；根据各个机组的地理信息，从全局气象数据中提取各个机组的与结冰相关的有效气象预测数据特征。
在一些实施例中，有效气象预测数据特征可以包括：气象数据原始特征和/或经气象数据原始特征加工得到的加工特征。
在一些实施例中，提取目标机组的有效气象预测数据特征的提取方法包括以下方法中的一种或者多种：反距离加权插值法、改进谢别德法、双线性插值法、自然邻点插值法、移动平均法。
在一些实施例中，获取目标机组所在的目标风电场的全局气象数据，包括：获取目标风电场的地形特性和气候特征；根据地形特征和气候特征，确定包括参数化方案的全局气象数据，以针对目标风电场进行数值天气预报。
在一些实施例中，风力发电机组结冰预测的方法还可以包括：预先获取目标机组的历史气象预测数据特征以及历史结冰标志位；将历史气象预测数据特征以及历史结冰标志位作为输入信息，建立用于输出预测结冰信息的结冰预测模型。
在一些实施例中，将历史气象预测数据特征以及历史结冰标志位作为输入信息，建立用于输出预测结冰信息的结冰预测模型，可以包括：将输入信息分为训练数据集和测试数据集；基于机器监督学习方法训练训练数据集，并用测试数据集进行测试，得到测试结果；根据测试结果建立结冰预测模型。
在一些实施例中，将输入信息分为训练数据集和测试数据集可以包括：当目标机组为多个时，基于目标机组的地形条件和/或气象条件将目标机组进行聚类，生成第一类机组和第二类机组；将第一类机组的输入信息确定为训练数据集；将第二类机组的输入信息确定为测试数据集。
在一些实施例中，将输入信息分为训练数据集和测试数据集可以包括：目标机组为1个时，将第一时段的目标机组的输入信息聚类为训练数据集，将第二时段的目标机组的输入信息聚类为测试数据集。
由于上述实施例通过聚类处理，数据集更加有针对性、更加精确，可以为后续任意风电场任意数目的风机的精确结冰预测做准备。
在一些实施例中，获取目标机组的历史气象预测数据特征可以包括：获取目标机组所在目标风场的地形特性和气候特征；获取目标风场的历史全局气象数据；基于地形特性和气候特征，从历史全局气象数据中提取历史气象预测数据特征。
在一些实施例中，获取目标机组的历史结冰标志位，包括：判断温度和湿度是否符合结冰条件；当温度和湿度符合结冰条件时，判断目标机组的运行参数是否异常；当判定目标机组的运行参数异常时，得到目标机组 的历史结冰标志位。本申请实施例是利用上述结冰预测模型，预测目标机组的结冰信息。预测的结冰信息(如结冰标志位)的应用场景可以用于显示于风电场业主监控系统中。预测的结冰信息可以是预设时间后(如6小时后)的关于目标机组是否结冰的标识(如，结冰标志位为1，不结冰标志位为0，结很薄的冰标志位为0.1，结很厚的冰标志位为0.9)，也可以是目标机组的结冰指数曲线。
在一些实施例中，预测的结冰信息可以输出至更多系统并孵化相应服务，如作为区域多风场调度系统的结冰预测服务的核心变量；又如可作为动态运维决策系统的输入变量等应用场景。
需要说明的是，上述方法的执行主体可以是处理器，也可以是控制器等。在不冲突的情况下，本领域的技术人员可以按实际需要将上述的操作步骤的顺序进行灵活调整，或者将上述步骤进行灵活组合等操作。为了简明，不再赘述各种实现方式。另外，各实施例的内容可以相互参考引用。
图5是本申请一实施例的风力发电机组的结冰预测模型的生成装置的结构示意图。
如图5所示，该装置500可以包括：信息获取单元501、特征获取单元502和模型建立单元503。其中，信息获取单元501可以用于获取目标风场中目标机组的地理信息和历史结冰标志位信息；特征获取单元502可以用于获取目标机组的与地理信息对应的历史气象预测数据特征；模型建立单元503可以用于将历史气象预测数据特征以及历史结冰标志位作为输入信息，建立用于输出预测结冰信息的结冰预测模型。
在一些实施例中，信息获取单元501可以用于：基于地理信息获取目标机组的地形特点数据和气候特点数据；获取在预设时间段内的目标风场的历史全局气象数据；基于地形特点数据和气候特点数据，从历史全局气象数据中提取历史气象预测数据特征。
在一些实施例中，历史气象预测数据特征可以包括：历史全局气象数据中的原始数据，和/或，将原始数据加工得到的处理数据。
在一些实施例中，历史气象预测数据特征包括以下参数中的一种或者多种：微物理参数、积云对流参数、行星边界层参数、地表参数、湍流参 数、扩散参数、电波辐射参数。
在一些实施例中，信息获取单元501还可以用于：基于系数阈值以及预设时间段内的与目标机组对应的风能利用系数、空气密度、叶轮扫风面积、风速，确定目标机组的历史参考发电功率；获取预设时间段内监测的目标机组的历史实际发电功率；比较历史实际发电功率与历史参考发电功率的大小，并根据比较结果确定目标机组的历史结冰标志位信息。
在一些实施例中，信息获取单元501还可以用于：将历史参考发电功率确定为：系数阈值、风能利用系数、空气密度、叶轮扫风面积和风速的乘积。
在一些实施例中，系数阈值可以在0至1/2之间。
在一些实施例中，信息获取单元501还可以用于：获取设置于目标机组上的传感器采集的结冰传感数据；基于结冰传感数据，得到目标机组的历史结冰标志位信息。
在一些实施例中，模型建立单元503还可以用于：将历史气象预测数据特征和历史结冰标志位信息确定为样本数据；将样本数据分为训练数据集和测试数据集；基于机器监督学习方法训练训练数据集，得到基础气象预设数据特征；利用测试数据集测试基础气象预设数据特征，得到基于有效气象预测数据特征预测结冰信息的结冰预测模型。
图6是本申请一实施例的风力发电机组结冰预测的装置的结构示意图。
如图6所示，该装置600可以包括：信息提取单元601、信息输入单元602和信息输出单元603。其中，信息提取单元601可以用于基于目标机组的地理信息，提取目标机组的有效气象预测数据特征；信息输入单元602可以用于将有效气象预测数据特征输入用于预测结冰信息的结冰预测模型；信息输出单元603可以用于响应输入，结冰预测模型输出结冰预测结果。
在本实施例中，利用图6中的结冰预测模型对实时气象预测数据特征进行运算，生成预测的结冰预测标志位。结冰预测模型的运算环境可以是风场业主的中央监控系统，也可是远程监控中心计算机，或其他具备运算能力的机器。
在一些实施例中，结冰预测模型可以输出风电场业主监控系统中的6小时后的结冰指数曲线。
在一些实施例中，该装置600还可以包括：特征聚类单元。特征聚类单元可以用于：获取目标风场中各个目标机组的地形特点数据和/或气候特点数据；基于地形特点数据和/或气候特点数据，将历史气象预测数据特征进行聚类；根据历史结冰标志位信息和聚类的历史气象预测数据特征，训练得到基于有效气象预测数据特征预测结冰信息的结冰预测模型。
在一些实施例中，该装置600还可以包括：优化训练单元。优化训练单元可以用于：获取目标机组的运行参数；基于运行参数、历史气象预测数据特征、历史结冰标志位信息，训练得到基于有效气象预测数据特征预测结冰信息的结冰预测模型。
在一些实施例中，运行参数可以包括：叶片转速、桨距角、发电功率。
上述各个实施例中的各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，也可以是两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件程序单元的形式实现。
需要说明的是，上述各实施例的装置可作为上述各实施例的用于各实施例的方法中的执行主体，可以实现各个方法中的相应流程，实现相同的技术效果，为了简洁，此方面内容不再赘述。
图7是本申请一实施例的风力发电机组结冰预测的装置的架构示意图。
如图7所示，该框架可以包括中央处理单元(CPU)701，其可以根据存储在只读存储器(ROM)702中的程序或者从存储部分708加载到随机访问存储器(RAM)703中的程序而执行图1、2、3、4实施例所做的各种操作。在RAM703中，还存储有系统架构操作所需的各种程序和数据。CPU 701、ROM 702以及RAM 703通过总线704彼此相连。输入/输出(I/O)接口705也连接至总线704。
以下部件连接至I/O接口705：包括键盘、鼠标等的输入部分706；包括诸如阴极射线管(CRT)、液晶显示器(LCD)等以及扬声器等的输出部分707；包括硬盘等的存储部分708；以及包括诸如LAN卡、调制解调 器等的网络接口卡的通信部分709。通信部分709经由诸如因特网的网络执行通信处理。驱动器710也根据需要连接至I/O接口705。可拆卸介质711，诸如磁盘、光盘、磁光盘、半导体存储器等等，根据需要安装在驱动器710上，以便于从其上读出的计算机程序根据需要被安装入存储部分708。
在本申请的一实施例中，上文参考相应附图描述的过程可以被实现为计算机程序。。在这样的实施例中，该计算机程序可以通过通信部分709从网络上被下载和安装，和/或从可拆卸介质711被安装。
以上所描述的装置实施例仅仅是示意性的，其中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模块来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下，即可以理解并实施。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件。基于这样的理解，上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，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)执行各个实施例或者实施例的某些部分所述的方法。
最后应说明的是：以上实施例仅用以说明本申请的技术方案，而非对其限制；尽管参照前述实施例对本申请进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。

Claims (15)

  1. 一种风力发电机组结冰预测的方法，其特征在于，所述方法包括：
    基于目标机组的地理信息，提取所述目标机组的有效气象预测数据特征；
    将所述有效气象预测数据特征输入用于预测结冰信息的结冰预测模型；
    响应所述输入，所述结冰预测模型输出结冰预测结果。
  2. 根据权利要求1所述的方法，其特征在于，所述基于目标机组的地理信息，提取所述目标机组的有效气象预测数据特征，包括：
    获取所述目标机组所在的目标风电场的全局气象数据；
    获取所述目标风电场中各个机组的地理信息；
    根据各个机组的地理信息，从所述全局气象数据中提取各个机组的与结冰相关的所述有效气象预测数据特征。
  3. 根据权利要求1所述的方法，其特征在于，所述有效气象预测数据特征包括：
    气象数据原始特征和/或经所述气象数据原始特征加工得到的加工特征。
  4. 根据权利要求1所述的方法，其特征在于，提取所述目标机组的有效气象预测数据特征的提取方法包括以下方法中的一种或者多种：
    反距离加权插值法、改进谢别德法、双线性插值法、自然邻点插值法、移动平均法。
  5. 根据权利要求2所述的方法，其特征在于，所述获取所述目标机组所在的目标风电场的全局气象数据，包括：
    获取所述目标风电场的地形特性和气候特征；
    根据所述地形特征和所述气候特征，确定包括参数化方案的所述全局气象数据，以针对所述目标风电场进行数值天气预报。
  6. 根据权利要求1所述的方法，其特征在于，还包括：
    预先获取所述目标机组的历史气象预测数据特征以及历史结冰标志位；
    将所述历史气象预测数据特征以及所述历史结冰标志位作为输入信息，建立用于输出预测结冰信息的所述结冰预测模型。
  7. 根据权利要求6所述的方法，其特征在于，将所述历史气象预测数 据特征以及所述历史结冰标志位作为输入信息，建立用于输出预测结冰信息的所述结冰预测模型，包括：
    将所述输入信息分为训练数据集和测试数据集；
    基于机器监督学习方法训练所述训练数据集，并用所述测试数据集进行测试，得到测试结果；
    根据所述测试结果建立所述结冰预测模型。
  8. 根据权利要求6所述的方法，其特征在于，所述将所述输入信息分为训练数据集和测试数据集，包括：
    当所述目标机组为多个时，基于所述目标机组的地形条件和/或气象条件将所述目标机组进行聚类，生成第一类机组和第二类机组；
    将所述第一类机组的所述输入信息确定为所述训练数据集；
    将所述第二类机组的所述输入信息确定为所述测试数据集；
    或者，
    当所述目标机组为一个时，将第一时段的所述目标机组的所述输入信息聚类为所述训练数据集，将第二时段的所述目标机组的所述输入信息聚类为所述测试数据集。
  9. 根据权利要求6所述的方法，其特征在于，所述获取所述目标机组的历史气象预测数据特征，包括：
    获取所述目标机组所在目标风场的地形特性和气候特征；
    获取所述目标风场的历史全局气象数据；
    基于所述地形特性和所述气候特征，从所述历史全局气象数据中提取所述历史气象预测数据特征。
  10. 根据权利要求6所述的方法，其特征在于，所述获取所述目标机组的历史结冰标志位，包括：
    判断温度和湿度是否符合结冰条件；
    当所述温度和所述湿度符合结冰条件时，判断所述目标机组的运行参数是否异常；
    当判定所述目标机组的运行参数异常时，得到所述目标机组的所述历史结冰标志位。
  11. 一种风力发电机组结冰预测的装置，其特征在于，所述装置包括：
    信息提取单元，用于基于目标机组的地理信息，提取所述目标机组的有效气象预测数据特征；
    信息输入单元，用于将所述有效气象预测数据特征输入用于预测结冰信息的结冰预测模型；
    信息输出单元，用于响应所述输入，所述结冰预测模型输出结冰预测结果。
  12. 一种风力发电机组结冰预测的装置，其特征在于，所述装置包括：
    至少一个处理器、至少一个存储器以及存储在所述存储器中的计算机程序，
    当所述计算机程序被所述处理器执行时实现如权利要求1-10中任一项所述的方法。
  13. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，当所述计算机程序被处理器执行时实现如权利要求1-10中任一项所述的方法。
  14. 一种风力发电机组的结冰预测模型的生成方法，其特征在于，所述方法包括：
    获取目标风场中目标机组的地理信息和历史结冰标志位信息；
    获取所述目标机组的与所述地理信息对应的历史气象预测数据特征；
    将所述历史气象预测数据特征以及所述历史结冰标志位作为输入信息，建立用于输出预测结冰信息的所述结冰预测模型。
  15. 一种风力发电机组的结冰预测模型的生成装置，其特征在于，所述装置包括：
    信息获取单元，用于获取目标风场中目标机组的地理信息和历史结冰标志位信息；
    特征获取单元，用于获取所述目标机组的与所述地理信息对应的历史气象预测数据特征；
    模型建立单元，用于将所述历史气象预测数据特征以及所述历史结冰标志位作为输入信息，建立用于输出预测结冰信息的所述结冰预测模型。
PCT/CN2018/082514 2017-12-14 2018-04-10 结冰预测方法、装置、模型生成方法及装置 WO2019114160A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US16/338,808 US20210332796A1 (en) 2017-12-14 2018-04-10 Method and apparatus for icing prediction, method and apparatus for generatng icing prediction model
EP18863806.8A EP3524813B1 (en) 2017-12-14 2018-04-10 Method and apparatus for predicting ice formation

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201711341540.6 2017-12-14
CN201711341540.6A CN109958588B (zh) 2017-12-14 2017-12-14 结冰预测方法、装置、存储介质、模型生成方法及装置

Publications (1)

Publication Number Publication Date
WO2019114160A1 true WO2019114160A1 (zh) 2019-06-20

Family

ID=66820715

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/082514 WO2019114160A1 (zh) 2017-12-14 2018-04-10 结冰预测方法、装置、模型生成方法及装置

Country Status (4)

Country Link
US (1) US20210332796A1 (zh)
EP (1) EP3524813B1 (zh)
CN (1) CN109958588B (zh)
WO (1) WO2019114160A1 (zh)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111539842A (zh) * 2020-04-08 2020-08-14 成都思晗科技股份有限公司 基于气象和地理环境的架空输电线路覆冰预测方法
CN112684520A (zh) * 2020-12-30 2021-04-20 北京墨迹风云科技股份有限公司 一种气象预报的订正方法、装置、计算机设备及存储介质
CN113569452A (zh) * 2021-07-22 2021-10-29 四川大学 一种结冰水库回水区冰塞反演及预测方法
CN114430816A (zh) * 2019-10-02 2022-05-03 通用电气公司 用于融合风力涡轮机的多种分析以提高效率的系统和方法
CN115438554A (zh) * 2022-11-04 2022-12-06 国网江西省电力有限公司电力科学研究院 一种基于天气预报的风电覆冰预测方法
CN115649455A (zh) * 2022-12-07 2023-01-31 中国空气动力研究与发展中心低速空气动力研究所 一种基于电热除冰信号判别结冰的方法及装置
CN117191334A (zh) * 2023-11-08 2023-12-08 中国空气动力研究与发展中心低速空气动力研究所 一种复杂冰形测量误差评估方法
CN118469553A (zh) * 2024-07-12 2024-08-09 联智科技（天津）有限责任公司 基于多源气象数据的输电线路缺陷隐患预测方法及系统

Families Citing this family (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110502526B (zh) * 2019-08-26 2023-05-09 安徽省气象信息中心 一种适用于结冰现象的资料序列插补的方法
CN111723929A (zh) * 2020-05-08 2020-09-29 广州华工中云信息技术有限公司 一种基于神经网络的数值预报产品订正方法、装置和系统
CN113191099A (zh) * 2021-04-28 2021-07-30 北京航空航天大学 一种考虑结冰影响的无人机动力学建模方法
CN113435645B (zh) * 2021-06-24 2023-10-10 长安大学 一种路面凝冰时间预测方法、系统、装置及存储介质
CN113847216B (zh) * 2021-10-14 2023-09-26 远景智能国际私人投资有限公司 风机叶片的状态预测方法、装置、设备及存储介质
CN114354674A (zh) * 2021-12-06 2022-04-15 西安交通大学 一种水滴撞击过冷表面的实验装置及应用
CN114220284B (zh) * 2021-12-13 2023-08-08 四川路桥建设集团交通工程有限公司 一种高速公路监控方法、系统、计算机设备及计算机可读存储介质
CN114966898B (zh) * 2022-05-09 2024-06-04 广东省韶关市气象局 一种高速路段雨雪冰冻天气预警方法及系统
CN114999117B (zh) * 2022-06-29 2023-09-01 洛阳腾飞建设工程集团有限公司 一种机场道面冰雪状况监测预警方法、系统、终端及介质
CN115857059B (zh) * 2022-11-24 2023-07-07 中国民航科学技术研究院 一种积冰预报管理方法、系统、可读存储介质和设备
CN116006809B (zh) * 2022-12-20 2024-06-21 成都秦川物联网科技股份有限公司 一种基于智慧燃气的管道低温维护方法和物联网系统
CN116383728A (zh) * 2023-06-05 2023-07-04 四川省气象服务中心(四川省专业气象台 四川省气象影视中心) 基于分区域svm的气象预测方法
CN116451594B (zh) * 2023-06-15 2023-08-18 北京东润环能科技股份有限公司 覆冰预测模型的训练方法、预测方法、装置和电子设备
CN116805045B (zh) * 2023-08-17 2024-01-23 北京电科智芯科技有限公司 气象预测模型修正方法、装置、设备及可读存储介质
CN117408053B (zh) * 2023-10-18 2024-05-07 中国空气动力研究与发展中心计算空气动力研究所 强对流条件下低温平板干模态结霜特征曲线的建立方法
CN117407773B (zh) * 2023-12-14 2024-06-18 山东捷瑞数字科技股份有限公司 基于数字孪生的风机叶片结冰状态预测方法、系统及设备
CN117638926B (zh) * 2024-01-25 2024-04-05 国能日新科技股份有限公司 基于覆冰与功率耦合建模的新能源功率预测方法及装置
CN117892213B (zh) * 2024-03-18 2024-06-25 中国水利水电第十四工程局有限公司 一种风力发电机叶片覆冰侦测与预警的诊断方法
CN118094485B (zh) * 2024-04-26 2024-07-23 湖南防灾科技有限公司 一种适用于复杂微地形的架空线路覆冰厚度分析方法

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110089692A1 (en) * 2006-10-18 2011-04-21 Boralex Inc. System and Method for Controlling a Wind Turbine
CN102706389A (zh) * 2012-06-19 2012-10-03 中国民航大学 采用滚动预测的飞机表面积冰短时预报系统及预报方法
EP2592447A1 (en) * 2011-11-08 2013-05-15 Topwind Consultancy B.V. Frost condition detection system and method
KR101520598B1 (ko) * 2015-02-10 2015-05-14 김국빈 풍력발전기 블레이드의 결빙감시시스템
CN204666839U (zh) * 2015-05-22 2015-09-23 中能电力科技开发有限公司 风电场天气检测装置
CN105298761A (zh) * 2015-11-06 2016-02-03 周志宏 一种风力发电机组结冰预警和控制方法及其装置
CN105402093A (zh) * 2015-12-30 2016-03-16 国电联合动力技术有限公司 一种用于风场级别的风机结冰检测方法和装置
CN105464912A (zh) * 2016-01-27 2016-04-06 国电联合动力技术有限公司 一种风力发电机组叶片结冰检测的方法和装置
CN105508152A (zh) * 2015-12-31 2016-04-20 北京金风科创风电设备有限公司 叶片结冰模型的构建方法、结冰状态的监测方法和装置
CN106679719A (zh) * 2016-12-13 2017-05-17 云南能投海装新能源设备有限公司 一种风机结冰检测装置

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
ITTO20060400A1 (it) * 2006-05-31 2007-12-01 Lorenzo Battisti Metodo e sistema per la rilevazione di pericolo di formazione di ghiaccio su superfici aerodinamiche
GB2482038B (en) * 2010-07-14 2014-07-23 Vestas Wind Sys As Ice detection method and system for wind turbine blades
US20120226485A1 (en) * 2011-03-03 2012-09-06 Inventus Holdings, Llc Methods for predicting the formation of wind turbine blade ice
US8606418B1 (en) * 2011-03-18 2013-12-10 Rockwell Collins, Inc. Wind prediction for wind farms through the use of weather radar
EP2969757B1 (en) * 2013-03-15 2017-08-23 Unmanned Aerospace Technologies Ltd Methods and system for deicing a surface
EP2826993B1 (de) * 2013-07-17 2017-04-12 ADIOS Patent GmbH Windenergieanlagenrotorblattenteisungsverfahren sowie Windenergieanlagenrotorblattenteisungssystem
DE102017125457B4 (de) * 2017-10-30 2023-02-23 fos4X GmbH Verfahren zum Bestimmen einer Wahrscheinlichkeit zu einem Drosseln und/oder einem Abschalten zumindest einer Windkraftanlage aufgrund von Eisansatz

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110089692A1 (en) * 2006-10-18 2011-04-21 Boralex Inc. System and Method for Controlling a Wind Turbine
EP2592447A1 (en) * 2011-11-08 2013-05-15 Topwind Consultancy B.V. Frost condition detection system and method
CN102706389A (zh) * 2012-06-19 2012-10-03 中国民航大学 采用滚动预测的飞机表面积冰短时预报系统及预报方法
KR101520598B1 (ko) * 2015-02-10 2015-05-14 김국빈 풍력발전기 블레이드의 결빙감시시스템
CN204666839U (zh) * 2015-05-22 2015-09-23 中能电力科技开发有限公司 风电场天气检测装置
CN105298761A (zh) * 2015-11-06 2016-02-03 周志宏 一种风力发电机组结冰预警和控制方法及其装置
CN105402093A (zh) * 2015-12-30 2016-03-16 国电联合动力技术有限公司 一种用于风场级别的风机结冰检测方法和装置
CN105508152A (zh) * 2015-12-31 2016-04-20 北京金风科创风电设备有限公司 叶片结冰模型的构建方法、结冰状态的监测方法和装置
CN105464912A (zh) * 2016-01-27 2016-04-06 国电联合动力技术有限公司 一种风力发电机组叶片结冰检测的方法和装置
CN106679719A (zh) * 2016-12-13 2017-05-17 云南能投海装新能源设备有限公司 一种风机结冰检测装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3524813A4 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114430816A (zh) * 2019-10-02 2022-05-03 通用电气公司 用于融合风力涡轮机的多种分析以提高效率的系统和方法
CN111539842A (zh) * 2020-04-08 2020-08-14 成都思晗科技股份有限公司 基于气象和地理环境的架空输电线路覆冰预测方法
CN111539842B (zh) * 2020-04-08 2023-05-23 成都思晗科技股份有限公司 基于气象和地理环境的架空输电线路覆冰预测方法
CN112684520A (zh) * 2020-12-30 2021-04-20 北京墨迹风云科技股份有限公司 一种气象预报的订正方法、装置、计算机设备及存储介质
CN113569452A (zh) * 2021-07-22 2021-10-29 四川大学 一种结冰水库回水区冰塞反演及预测方法
CN113569452B (zh) * 2021-07-22 2022-03-15 四川大学 一种结冰水库回水区冰塞反演及预测方法
CN115438554A (zh) * 2022-11-04 2022-12-06 国网江西省电力有限公司电力科学研究院 一种基于天气预报的风电覆冰预测方法
CN115649455A (zh) * 2022-12-07 2023-01-31 中国空气动力研究与发展中心低速空气动力研究所 一种基于电热除冰信号判别结冰的方法及装置
CN117191334A (zh) * 2023-11-08 2023-12-08 中国空气动力研究与发展中心低速空气动力研究所 一种复杂冰形测量误差评估方法
CN118469553A (zh) * 2024-07-12 2024-08-09 联智科技（天津）有限责任公司 基于多源气象数据的输电线路缺陷隐患预测方法及系统

Also Published As

Publication number Publication date
CN109958588A (zh) 2019-07-02
EP3524813B1 (en) 2022-05-25
US20210332796A1 (en) 2021-10-28
EP3524813A4 (en) 2019-12-11
CN109958588B (zh) 2020-08-07
EP3524813A1 (en) 2019-08-14

Similar Documents

Publication Publication Date Title
WO2019114160A1 (zh) 结冰预测方法、装置、模型生成方法及装置
Taszarek et al. Sounding-derived parameters associated with convective hazards in Europe
Adams-Selin et al. Forecasting hail using a one-dimensional hail growth model within WRF
Davy et al. Statistical downscaling of wind variability from meteorological fields
Jergensen et al. Classifying convective storms using machine learning
CN109751206B (zh) 风机叶片结冰故障预测方法、装置及存储介质
CN105182450B (zh) 一种强对流天气短时预警系统
Davis et al. Does increased horizontal resolution improve hurricane wind forecasts?
Kain et al. A feasibility study for probabilistic convection initiation forecasts based on explicit numerical guidance
CN110705115B (zh) 一种基于深度置信网络的气象预报方法和系统
Adams-Selin et al. Evolution of WRF-HAILCAST during the 2014–16 NOAA/hazardous weather testbed spring forecasting experiments
Rajeevan et al. Development of a perfect prognosis probabilistic model for prediction of lightning over south-east India
CN109299208B (zh) 一种台风灾害下输电杆塔智能可视化风险评估方法
CN113033910A (zh) 一种光伏发电功率的预测方法、存储介质及终端设备
Steinkruger et al. An artificially intelligent system for the automated issuance of tornado warnings in simulated convective storms
CN116021981A (zh) 配电网线路覆冰故障预测方法、装置、设备及存储介质
Liou et al. Thermodynamic recovery of the pressure and temperature fields over complex terrain using wind fields derived by multiple-Doppler radar synthesis
Miller et al. A preliminary assessment of using spatiotemporal lightning patterns for a binary classification of thunderstorm mode
CN115047544B (zh) 一种覆冰预警方法及系统
Marchand et al. A bootstrap technique for testing the relationship between local-scale radar observations of cloud occurrence and large-scale atmospheric fields
CN110555552B (zh) 基于天气危险指数的终端区起降容量预测方法
CN112580899A (zh) 一种融合机器学习模型的中长期风力发电预测方法及系统
Lockwood et al. A Machine Learning Approach to Model Over-Ocean Tropical Cyclone Precipitation
CN116756644B (zh) 一种风电机组风速风向仪覆冰故障的预警方法
CN110533241A (zh) 终端区起降容量预测系统

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2018863806

Country of ref document: EP

Effective date: 20190411

ENP Entry into the national phase

Ref document number: 2018863806

Country of ref document: EP

Effective date: 20190411

NENP Non-entry into the national phase

Ref country code: DE