WO2018171165A1 - 预测风机的故障的方法和设备 - Google Patents

预测风机的故障的方法和设备 Download PDF

Info

Publication number
WO2018171165A1
WO2018171165A1 PCT/CN2017/105440 CN2017105440W WO2018171165A1 WO 2018171165 A1 WO2018171165 A1 WO 2018171165A1 CN 2017105440 W CN2017105440 W CN 2017105440W WO 2018171165 A1 WO2018171165 A1 WO 2018171165A1
Authority
WO
WIPO (PCT)
Prior art keywords
fan
model
parameters
parameter
fan model
Prior art date
Application number
PCT/CN2017/105440
Other languages
English (en)
French (fr)
Inventor
张光磊
刘源
周杰
Original Assignee
新疆金风科技股份有限公司
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 新疆金风科技股份有限公司 filed Critical 新疆金风科技股份有限公司
Priority to AU2017363221A priority Critical patent/AU2017363221B2/en
Priority to EP17875063.4A priority patent/EP3407271A4/en
Priority to US15/777,065 priority patent/US11035346B2/en
Publication of WO2018171165A1 publication Critical patent/WO2018171165A1/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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • F03D7/028—Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor controlling wind motor output power
    • F03D7/0292—Controlling wind motors  the wind motors having rotation axis substantially parallel to the air flow entering the rotor controlling wind motor output power to reduce fatigue
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/34—Testing dynamo-electric machines
    • G01R31/343—Testing dynamo-electric machines in operation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • 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
    • 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

  • This application relates to the field of wind energy. More particularly, it relates to a method and apparatus for predicting a fault in a wind turbine.
  • wind energy As a clean and renewable energy source, wind energy is receiving more and more attention, and the installed capacity of wind energy equipment (ie, wind turbines) is also increasing.
  • the wind turbine may be various devices that are driven by wind energy, for example, wind power generation equipment, windmills, and the like, which are driven by wind energy.
  • Fans are usually operated in natural conditions such as the field, and the cost of maintenance is high. Therefore, predicting the failure of the fan in advance can effectively know the failure of the fan in advance, so that measures can be taken to avoid the occurrence of the failure.
  • the present application provides a method and apparatus that can more accurately predict the failure of a wind turbine.
  • a method for predicting a fault of a wind turbine comprising: initializing a wind turbine model of the wind turbine; periodically updating the wind turbine model of the wind turbine model according to current environmental conditions and current actual wind turbine state parameters of the wind turbine Parameter; based on the historical fan model parameters and corresponding historical environmental conditions, establish a fan model parameter variation model; based on future environmental conditions in the future, use the fan model parameter variation model to predict the fan model parameters at the future time; The fan model parameters and the corresponding historical fan failures establish a fan failure model; predict wind turbine failures in the future based on predicted fan model parameters and fan failure models.
  • an apparatus for predicting a failure of a wind turbine comprising: an initialization unit that initializes a fan model of the wind turbine; and a parameter update unit that periodically according to current environmental conditions and current actual fan state of the wind turbine The parameters are used to update the fan model parameters of the fan model; the parameter model modeling unit establishes a fan model parameter variation model according to historical fan model parameters and corresponding historical environmental conditions; the parameter prediction unit is based on future environmental conditions in the future time, The fan model parameter variation model is used to predict the fan model parameters at the future time; the fault model modeling unit establishes a fan fault model according to the historical fan model parameters and the corresponding historical fan fault; the fault prediction unit, according to the predicted fan Model parameters and fan failure models predict wind turbine failures in the future.
  • a system for predicting a failure of a wind turbine comprising: a processor; a memory storing computer readable code, when the computer readable code is executed by a processor, performing the above method.
  • a computer readable storage medium having stored therein computer readable code, the method being performed when the computer readable code is executed.
  • the influence of environmental factors on the operating state of the fan is fully considered in the fault prediction of the fan, and the specific relationship between the fan and the environmental factor is considered, and the fan model is established based on the influence of environmental factors.
  • the parameters of the fan model are continuously updated to obtain the historical fan model parameters, and the fan model parameter variation model is established.
  • the model can predict the state change of the fan under any environmental conditions.
  • a fan fault model related to the fan model parameters is further established for fault prediction, thereby more accurately predicting the fault of the fan.
  • FIG. 1 shows a flow chart of a method of predicting a failure of a wind turbine, in accordance with an embodiment of the present application
  • FIG. 2 shows a block diagram of an apparatus for predicting a failure of a wind turbine, in accordance with an embodiment of the present application.
  • FIG. 1 illustrates a flow chart of a method of predicting a failure of a wind turbine, in accordance with an embodiment of the present application.
  • step S110 the fan model of the fan is initialized.
  • the fan model embodies the relationship between the environmental conditions in which the fan is located and the state parameters of the fan (hereinafter referred to as the fan state parameters).
  • the fan model can estimate and output the corresponding fan state parameter by inputting the environmental condition to the fan model.
  • the fan module can be established by obtaining the measured environmental conditions and the measured fan state parameters. type.
  • empirical models can also be used as fan models.
  • Various modeling techniques can be utilized to establish the fan model. This application is not limited, as long as the established fan model can reflect the relationship between environmental conditions and fan state parameters.
  • the environmental conditions of the fan are environmental factors that affect the operating state of the fan.
  • Environmental conditions may include, but are not limited to, at least one of the following parameters: wind speed, temperature, humidity, air density, and the like. Environmental conditions can be measured using various sensors for detecting environmental conditions.
  • the fan status parameter may include, but is not limited to, at least one of the following parameters: current speed, yaw direction, pitch angle, output power, and the like.
  • the wind turbine state parameter may also include electrical parameters such as power generation.
  • the initial fan model parameters can be determined based on initial environmental conditions and initial actual fan state parameters.
  • the current environmental condition is acquired as the initial environmental condition
  • the actual fan state parameter is detected as the initial actual fan state parameter.
  • the fan model parameters are initialized based on the initial environmental conditions and the initial actual fan state parameters to determine the initial fan model parameters.
  • the parameters of the fan model are determined in the case where the initial input and the desired output of the fan model are determined.
  • the initial fan model parameters are such that the fan state parameter output by the fan model input with the initial environmental conditions is as close as possible to the initial actual fan state parameter.
  • the initial fan model parameters themselves are further considered in determining the initial fan model parameters.
  • the initial fan model parameters are such that the fan state parameter output by the fan model input with the initial environmental conditions is as close as possible to the initial actual fan state parameter, and the total size of the initial fan model parameters is weighted as small as possible.
  • the total size of the initial fan model parameters can be expressed by the sum of the absolute values of the individual initial fan model parameters, or by the sum of the even powers of the individual initial fan model parameters (it should be understood that there is only one fan model)
  • the "sum" referred to herein is the absolute or even power of the 1 initial fan model parameter.
  • the weights used for weighting can be obtained empirically or experimentally. For example, by obtaining a result of "the change of the fan state parameter outputted by the fan model in which the initial environmental condition is input with respect to the initial actual fan state parameter" and "the total size of the initial fan model parameter are weighted” And minimized fan model parameters as initial fan model parameters.
  • the initial actual fan state parameters are obtained by equation (1) below:
  • x 0 represents the initial environmental condition
  • y 0 represents the initial actual fan state parameter
  • f 0 (x 0 ) represents the fan state parameter output by the fan model to which the initial environmental conditions are input
  • ⁇ 0 represents the initial fan model
  • the parameter, ⁇ represents the weighting factor of the canonical item, Represents obtaining ⁇ 0 that minimizes the formula in parentheses.
  • the gauge weight coefficient ⁇ can be obtained empirically or experimentally. It should be understood that Represents the total size of the initial fan model parameters mentioned above.
  • step S120 the fan model parameters of the fan model are periodically updated according to the current environmental conditions and the current actual fan state parameters of the fan.
  • the updated fan model parameter may be periodically obtained according to the current environmental condition and the current actual fan state parameter of the fan as the fan model parameter of the current cycle. In this way, the fan model can be kept accurate while collecting historical fan model parameters and corresponding historical environmental conditions.
  • the current environmental conditions and the current actual fan state parameters of the fan are obtained to determine the fan model parameters for each cycle with the initial input and desired output of the fan model determined.
  • the updated fan model parameters cause the fan status parameter output by the fan model that inputs the current environmental conditions to be as close as possible to the current actual fan status parameter.
  • the initial fan model parameters are such that the difference between the fan state parameter output by the fan model inputting the current environmental condition and the current actual fan state parameter is minimized.
  • the variation of the fan model parameters is further considered when updating the fan model parameters for each cycle.
  • the updated fan model parameters are such that the fan state parameter output by the fan model inputting the current environmental condition is as close as possible to the current actual fan state parameter, and the updated fan model parameter is compared with the last update.
  • the results of the subsequent fan model parameters ie, the fan model parameters updated in the previous cycle
  • the updated fan model parameter is such that the difference between the fan state parameter output by the fan model inputting the current environmental condition and the current actual fan state parameter is minimized, and the updated fan model parameter and The difference between the fan model parameters after the last update is minimized.
  • the fan model parameters are used as updated fan model parameters.
  • the sum of the absolute values or the even powers of the difference between the different types of fan model parameters between the respective updated fan model parameters and each of the last updated fan model parameters can be calculated as the change (should be understood)
  • the "sum” mentioned here is the absolute value or the even power of the difference between the parameters of the one type of fan model.
  • the weights used for weighting can be obtained empirically or experimentally.
  • the updated fan model parameters are obtained by equation (2) below:
  • x t represents the current environmental condition
  • y t represents the current actual fan state parameter
  • f t (x t ) represents the fan state parameter output by the fan model to which the current environmental condition is input
  • ⁇ t represents the updated fan model
  • the parameter, ⁇ t-1 represents the fan model parameter after the last update
  • ⁇ represents the weight coefficient of the norm item. Represents obtaining ⁇ t that minimizes the formula in parentheses.
  • step S130 a fan model parameter variation model is established according to the historical fan model parameters and the corresponding historical environmental conditions.
  • the fan model parameter variation model embodies the relationship between environmental conditions and fan model parameters.
  • the fan model parameter variation model may estimate and output the corresponding fan model parameter.
  • various modeling techniques can be used to establish the fan model parameter variation model. This application does not limit, as long as the established fan model parameter variation model can reflect the environmental conditions and the fan model. The relationship between the parameters.
  • the wind turbine model parameter variation model is established by fitting using historical fan model parameters and corresponding historical environmental conditions.
  • the fan model parameter variation model can be fitted by various fitting methods.
  • the structure of the fan model parameter variation model can be obtained by fitting (at this time, the structure and parameters of the fan model parameter variation model are determined by fitting) or can be determined in advance (at this time, only the fan model parameter variation is determined by fitting) The parameters of the model).
  • the fan model parameter variation model is expressed as equation (3) below:
  • ⁇ T A ⁇ T-1 +B ⁇ x T (3)
  • ⁇ T represents the fan model parameter at time T
  • ⁇ T-1 represents the fan model parameter at time T-1
  • A is the state transition matrix of the fan model parameter
  • B is the environmental condition pair fan model parameter influence coefficient matrix
  • x T is the environmental condition at time T. It should be understood that the time T-1 represents the fan model parameter at a time after the time T.
  • time T represents a certain moment in the future
  • time T-1 represents when performing fan model parameter prediction or failure prediction.
  • a current time in another embodiment, an environmental condition of one or more times between the current time and a certain future time (hereinafter referred to as an intermediate time) may be acquired, starting from the current time, based on the formula (3) Predicting the fan model parameters at a later time using the environmental conditions at a later one of the two adjacent moments and the fan model parameters at the previous moment until the fan model at a certain moment in the future is predicted parameter.
  • the environmental conditions of the first intermediate moment and the fan model parameters of the current moment to predict the fan model parameters of the first moment using the environmental conditions of the second intermediate moment and The predicted fan model parameter at the first moment predicts the fan model parameter at the second moment, and uses the environmental condition at a certain moment in the future and the predicted fan model parameter at the second moment to predict the fan model parameter at a certain moment in the future.
  • the structure of the fan model parameter variation model is not limited to the function structure expressed by equation (3), and other structures may be determined or adopted by the fitting process.
  • Historical fan model parameters and corresponding historical environmental conditions may be obtained based on step S120.
  • the historical fan model parameters and the corresponding historical environmental conditions may represent the fan model parameters and corresponding environmental conditions of each period in the past period of time obtained through step S120; or all of the fan model parameters that have been obtained through steps S110 and S120 and Corresponding environmental conditions.
  • step S140 based on future environmental conditions at a future time, the fan model parameter change model is used to predict the fan model parameters at the future time.
  • future environmental conditions environmental conditions at the future time
  • future environmental conditions environmental conditions at the future time
  • the model is varied to predict the fan model parameters at the future time.
  • Environmental conditions at this future time can be obtained in various ways, for example, through weather forecasts, environmental models, and the like.
  • the present application does not limit the manner in which environmental conditions at the future time are obtained.
  • a fan failure model is established based on the historical fan model parameters and the corresponding historical fan failure.
  • the fan failure model embodies the relationship between fan model parameters and fan failure.
  • the fan failure model can estimate and output the corresponding fan failure.
  • various modeling techniques can be used to establish a fan fault model. This application is not limited, as long as the established fan fault model can reflect the relationship between the fan model parameters and the fan fault. relationship.
  • the fan model parameters of the current fan model can be obtained, thereby obtaining historical fan model parameters and corresponding historical fan faults.
  • "historical fan failure" herein indicates whether a failure has occurred.
  • obtaining the fan failure model may output information indicating whether a failure has occurred based on the input fan model parameter.
  • the fan model parameters can be used as a feature, whether a fault occurs as a classification label, and a historical fan model parameter and a corresponding historical fan fault are used as training samples to train the classifier as a fan fault model.
  • the fan failure model is a classifier characterized by fan model parameters and classified by whether or not a failure occurs.
  • the fault category can be further obtained.
  • the "historical fan failure" herein can indicate the fault category.
  • the fan failure model may output information indicating the type of the failure that occurred.
  • the fan model parameter can be used as a feature
  • the fault category is used as a classification label
  • the historical fan model parameter and the corresponding historical fan fault are used as training samples to train the classifier as a fan fault model.
  • the fan failure model is a classifier characterized by fan model parameters and classified by fault category.
  • the historical fan model parameters and corresponding historical fan faults used in establishing the wind turbine fault model may be historical wind turbine model parameters and corresponding historical wind turbines that have been collected so far when the wind turbine fault is to be predicted. malfunction.
  • a fan failure at a future time is predicted based on the predicted fan model parameters and the fan failure model.
  • the obtained fan model parameters are input into the established fan failure model, so that the fan failure model predicts the fan failure at a future time.
  • the predicted fault may only indicate whether a fault has occurred or may indicate the type of fault.
  • step S130 and step S150 may be performed simultaneously or interchanged. Further, step S130 and step S150 may be performed to determine that the fan failure is to be predicted to perform modeling based on the latest data. Further, step S130 and step S150 may be performed in advance to be pre-modeled.
  • FIG. 2 shows a block diagram of an apparatus for predicting a failure of a wind turbine, in accordance with an embodiment of the present application.
  • the apparatus 200 for predicting a failure of a wind turbine includes an initialization unit 210, a parameter update unit 220, a parameter model modeling unit 230, a parameter prediction unit 240, a failure model modeling unit 250, and a failure. Prediction unit 260.
  • the initialization unit 210 initializes the fan model of the fan.
  • the fan model embodies the relationship between the environmental conditions in which the fan is located and the state parameters of the fan (hereinafter referred to as the fan state parameters).
  • the fan model when the environmental condition is input to the fan model, the fan model can estimate and output the corresponding fan state parameter.
  • the fan model can be established by taking the measured environmental conditions and the measured fan state parameters.
  • empirical models can also be used as fan models.
  • Various modeling techniques can be utilized to establish the fan model. This application is not limited, as long as the established fan model can reflect the relationship between environmental conditions and fan state parameters.
  • the environmental conditions of the fan are environmental factors that affect the operating state of the fan.
  • Environmental conditions may include, but are not limited to, at least one of the following parameters: wind speed, temperature, humidity, air density, and the like. Environmental conditions can be measured using various sensors for detecting environmental conditions.
  • the fan status parameter may include, but is not limited to, at least one of the following parameters: current speed, yaw direction, pitch angle, output power, and the like.
  • the wind turbine state parameter may also include electrical parameters such as power generation.
  • the initialization unit 210 can determine initial fan model parameters based on initial environmental conditions and initial actual fan state parameters.
  • the initialization unit 210 when the initialization of the fan model is performed, the current environmental condition is acquired as the initial environmental condition, and the actual fan state parameter is detected as the initial actual fan state parameter.
  • the initialization unit 210 initializes the fan model parameters based on the initial environmental conditions and the initial actual fan state parameters to determine the initial fan model parameters.
  • the parameters of the fan model are determined in the case where the initial input and the desired output of the fan model are determined.
  • the initial fan model parameters are such that the fan state parameter output by the fan model input with the initial environmental conditions is as close as possible to the initial actual fan state parameter.
  • the initial fan model parameters are such that the difference between the fan state parameter output by the fan model inputting the current environmental condition and the current actual fan state parameter is minimized.
  • the initial consideration is further considered in determining the initial fan model parameters.
  • the fan model parameters themselves.
  • the initial fan model parameters are such that the fan state parameter output by the fan model input with the initial environmental conditions is as close as possible to the initial actual fan state parameter, and the total size of the initial fan model parameters is weighted as small as possible.
  • the updated fan model parameter is such that the difference between the fan state parameter output by the fan model inputting the current environmental condition and the current actual fan state parameter is minimized, and the updated fan model parameter and The difference between the fan model parameters after the last update is minimized.
  • the fan model parameter with the sum of the results of the weighted changes is used as the updated fan model parameter.
  • the total size of the initial fan model parameters can be expressed by the sum of the absolute values of the individual initial fan model parameters, or by the sum of the even powers of the individual initial fan model parameters (it should be understood that there is only one fan model)
  • the "sum” referred to herein is the absolute value or the even power of the difference between the parameters of the fan type of the one type.
  • the weights used for weighting can be obtained empirically or experimentally.
  • the parameter update unit 220 periodically updates the fan model parameters of the fan model based on the current environmental conditions and the current actual fan state parameters of the wind turbine.
  • the parameter updating unit 220 may periodically obtain the updated fan model parameter as the current model fan model parameter according to the current environmental condition and the current actual fan state parameter of the fan. In this way, the fan model can be kept accurate while collecting historical fan model parameters and corresponding historical environmental conditions.
  • the parameter update unit 220 obtains the current environmental conditions and the current actual fan state parameters of the wind turbine, thereby determining the fan model parameters for each cycle in the event that the initial input and desired output of the wind turbine model are determined.
  • the updated fan model parameters cause the fan status parameter output by the fan model that inputs the current environmental conditions to be as close as possible to the current actual fan status parameter.
  • the variation of the fan model parameters is further considered when updating the fan model parameters for each cycle.
  • the updated fan model parameters are such that the fan state parameter output by the fan model inputting the current environmental condition is as close as possible to the current actual fan state parameter, and the updated fan model parameter is compared with the last update.
  • the changes in the fan model parameters ie, the fan model parameters updated in the previous cycle) are weighted as small as possible.
  • the change is represented by summing the sum of the absolute values or the even powers of the differences between the fan model parameters of the respective types of fan model parameters between the respective updated fan model parameters (should be understood, In the case where the fan model has only one (i.e., one type) fan model parameter, the "sum” referred to herein is the absolute value or the even power of the difference between the parameters of the fan type of the one type.
  • the weights used for weighting can be obtained empirically or experimentally. For example, by obtaining the change of the fan state parameter outputted by the fan model inputting the current environmental condition with respect to the current actual fan state parameter and the updated fan model parameter with respect to the last updated fan model parameter. The fan model parameter with the sum of the results of the weighted changes is used as the updated fan model parameter.
  • the updated fan model parameters are obtained by equation (2) above.
  • the parametric model modeling unit 230 establishes a fan model parameter variation model based on historical fan model parameters and corresponding historical environmental conditions.
  • the fan model parameter variation model embodies the relationship between environmental conditions and fan model parameters.
  • the fan model parameter variation model may estimate and output the corresponding fan model parameter.
  • various modeling techniques can be used to establish the fan model parameter variation model. This application does not limit, as long as the established fan model parameter variation model can reflect the environmental conditions and the fan model. The relationship between the parameters.
  • the wind turbine model parameter variation model is established by fitting using historical fan model parameters and corresponding historical environmental conditions.
  • the fan model parameter variation model can be fitted by various fitting methods.
  • the structure of the fan model parameter variation model can be obtained by fitting (at this time, the structure and parameters of the fan model parameter variation model are determined by fitting) or can be determined in advance (at this time, only the fan model parameter variation is determined by fitting) The parameters of the model).
  • the fan model parameter variation model is represented as equation (3) above.
  • Historical fan model parameters and corresponding historical environmental conditions may be obtained based on parameter update unit 220.
  • the historical fan model parameters and corresponding historical environmental conditions may represent the fan model parameters and corresponding environmental conditions over a past period of time obtained by the parameter update unit 220; or all of the values that have been obtained by the initialization unit 210 and the parameter update unit 220 Fan model parameters and corresponding environmental conditions.
  • the parameter prediction unit 240 predicts the fan model parameters at the future time using the fan model parameter variation model based on future environmental conditions at a future time.
  • the parameter prediction unit 240 may acquire an environmental condition at the future time (hereinafter, referred to as a future environmental condition), and input the acquired future environmental condition.
  • the wind turbine model parameter variation model is used to predict the fan model parameters at the future time.
  • Environmental conditions at this future time can be obtained in various ways, for example, through weather forecasts, environmental models, and the like.
  • the present application does not limit the manner in which environmental conditions at the future time are obtained.
  • the fault model modeling unit 250 establishes a wind turbine fault model based on historical fan model parameters and corresponding historical fan faults.
  • the fan failure model embodies the relationship between fan model parameters and fan failure.
  • the fan failure model can estimate and output the corresponding fan failure.
  • various modeling techniques can be used to establish a fan fault model. This application is not limited, as long as the established fan fault model can reflect the relationship between the fan model parameters and the fan fault. relationship.
  • the fan model parameters of the current fan model are obtained, resulting in historical fan model parameters and corresponding historical fan failures.
  • "historical fan failure" herein indicates whether a failure has occurred.
  • obtaining the fan failure model may output information indicating whether a failure has occurred based on the input fan model parameter.
  • the fan model parameters can be used as a feature, whether a fault occurs as a classification label, and a historical fan model parameter and a corresponding historical fan fault are used as training samples to train the classifier as a fan fault model.
  • the fan failure model is a classifier characterized by fan model parameters and classified by whether or not a failure occurs.
  • the fault category can be further obtained.
  • the "historical fan failure" herein can indicate the fault category.
  • the fan failure model may output information indicating the type of the failure that occurred.
  • the fan model parameter can be used as a feature
  • the fault category is used as a classification label
  • the historical fan model parameter and the corresponding historical fan fault are used as training samples to train the classifier as a fan fault model.
  • the fan failure model is a classifier characterized by fan model parameters and classified by fault category.
  • the historical fan model parameters and corresponding historical fan faults used in establishing the wind turbine fault model may be historical wind turbine model parameters and corresponding historical wind turbines that have been collected so far when the wind turbine fault is to be predicted. malfunction.
  • the fault prediction unit 260 predicts the future based on the predicted fan model parameters and the fan fault model. The wind turbine at the moment is faulty.
  • the obtained fan model parameters are input into the established fan failure model, so that the fan failure model predicts the fan failure at a future time.
  • the predicted fault may only indicate whether a fault has occurred or may indicate the type of fault.
  • the specific relationship between the fan and the environmental factor is taken into consideration when predicting the failure of the fan, and the fan model is established based on the influence of the environmental factor, and the parameters of the fan model are continuously updated to obtain the history.
  • the fan model parameters are used to establish a fan fault model related to the fan model parameters for fault prediction, which can more accurately predict the fault of the fan.
  • the present application also provides a system for predicting a fault of a wind turbine.
  • the system includes a processor and a memory, wherein the memory stores computer readable code that, when executed by the processor, performs the method of predicting a failure of the fan.
  • the above method according to an exemplary embodiment of the present application may be implemented as a computer program on a computer readable medium, thereby implementing the above-described method of predicting a malfunction of a fan when the program is run.
  • each of the above-described devices may be implemented as a hardware component or a software module.
  • those skilled in the art can implement various hardware components by using a field programmable gate array (FPGA), an application specific integrated circuit (ASIC) or a processor according to the processing performed by the defined respective units, which can be implemented by programming techniques.
  • FPGA field programmable gate array
  • ASIC application specific integrated circuit
  • Various software modules are also possible.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Theoretical Computer Science (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Quality & Reliability (AREA)
  • Chemical & Material Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Sustainable Energy (AREA)
  • Combustion & Propulsion (AREA)
  • Sustainable Development (AREA)
  • Mechanical Engineering (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Educational Administration (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Wind Motors (AREA)
  • Structures Of Non-Positive Displacement Pumps (AREA)

Abstract

提供一种预测风机的故障的方法和设备。所述方法包括：初始化风机的风机模型；周期性地根据当前环境条件和风机的当前实际风机状态参数来更新风机模型的风机模型参数；根据历史的风机模型参数和对应的历史的环境条件，建立风机模型参数变化模型；基于未来时刻的未来环境条件，使用风机模型参数变化模型预测在所述未来时刻的风机模型参数；根据历史的风机模型参数和对应的历史的风机故障，建立风机故障模型；根据预测的风机模型参数和风机故障模型，预测未来时刻的风机故障。

Description

预测风机的故障的方法和设备 技术领域
本申请涉及风能领域。更具体地讲，涉及一种预测风机的故障的方法和设备。
背景技术
风能作为一种清洁的可再生能源，越来越受到重视，风能设备(即，风机)的装机量也不断增加。风机可以是利用风能进行驱动的各种设备，例如，风力发电设备、风车等各种通过风能驱动的设备。
风机通常运行在野外等自然条件下，进行维护的成本较高。因此，对风机的故障提前进行预测可以有效地提前获知风机可能发生的故障，从而可以提取采取措施来避免故障的发生。
因此，需要一种能够准确地对风机的故障进行预测的技术。
发明内容
本申请提供一种能够更准确地预测风机的故障的方法和设备。
根据本申请的一方面，提供一种预测风机的故障的方法，所述方法包括：初始化风机的风机模型；周期性地根据当前环境条件和风机的当前实际风机状态参数来更新风机模型的风机模型参数；根据历史的风机模型参数和对应的历史的环境条件，建立风机模型参数变化模型；基于未来时刻的未来环境条件，使用风机模型参数变化模型预测在所述未来时刻的风机模型参数；根据历史的风机模型参数和对应的历史的风机故障，建立风机故障模型；根据预测的风机模型参数和风机故障模型，预测未来时刻的风机故障。
根据本申请的另一方面，提供一种预测风机的故障的设备，所述设备包括：初始化单元，初始化风机的风机模型；参数更新单元，周期性地根据当前环境条件和风机的当前实际风机状态参数来更新风机模型的风机模型参数；参数模型建模单元，根据历史的风机模型参数和对应的历史的环境条件，建立风机模型参数变化模型；参数预测单元，基于未来时刻的未来环境条件， 使用风机模型参数变化模型预测在所述未来时刻的风机模型参数；故障模型建模单元，根据历史的风机模型参数和对应的历史的风机故障，建立风机故障模型；故障预测单元，根据预测的风机模型参数和风机故障模型，预测未来时刻的风机故障。
根据本申请的另一方面，提供一种预测风机的故障的系统，所述系统包括：处理器；存储器，存储有计算机可读代码，当所述计算机可读代码被处理器执行时，执行上述方法。
根据本申请的另一方面，提供一种其中存储有计算机可读代码的计算机可读存储介质，当所述计算机可读代码被执行时执行上述方法。
根据本申请的预测风机的故障的方法和设备，在对风机进行故障预测时充分考虑了环境因素对风机运行状态的影响，考虑到风机与环境因素的特定关系，基于环境因素的影响建立风机模型。同时，不断地更新风机模型的参数来获得历史的风机模型参数，并建立风机模型参数变化模型，通过该模型可以预测在任意环境条件下风机的状态变化。基于风机模型参数变化模型进一步建立与风机模型参数有关的风机故障模型用于故障预测，从而更精确地进行风机的故障预测。
附图说明
通过下面结合附图进行的详细描述，本申请的上述和其它目的、特点和优点将会变得更加清楚，其中：
图1示出根据本申请的实施例的预测风机的故障的方法的流程图；
图2示出根据本申请的实施例的预测风机的故障的设备的框图。
具体实施方式
现在，将参照附图更充分地描述不同的示例实施例。
图1示出根据本申请的实施例的预测风机的故障的方法的流程图。
如图1所示，在步骤S110，初始化风机的风机模型。
风机模型体现了风机所处的环境条件与风机的状态参数(以下称为，风机状态参数)之间的关系。在本申请实施例中通过将环境条件输入到风机模型，风机模型可估计并输出对应的风机状态参数。
可以通过获取实测的环境条件以及实测的风机状态参数来建立风机模 型。此外，也可利用经验模型来作为风机模型。可利用各种建模技术来建立风机模型，本申请不做限制，只要建立的风机模型能够体现环境条件与风机状态参数之间的相互关系。
风机的环境条件是对风机的运行状态产生影响的环境因素。
环境条件可包括(但不限于)下面参数中的至少一个：风速、温度、湿度、空气密度等。可以利用各种用于检测环境状况的传感器来实测环境条件。
风机状态参数可包括(但不限于)下面参数中的至少一个：当前转速、偏航方向、变桨角度、输出功率等。例如，当风机为风力发电机时，风机状态参数还可包括发电功率等电气参数。
可根据初始的环境条件和初始的实际风机状态参数，来确定初始的风机模型参数。在本申请实施例中，当进行风机模型的初始化时，获取当时的环境条件作为初始的环境条件，并检测实际的风机状态参数作为初始的实际风机状态参数。在此情况下，根据初始的环境条件和初始的实际风机状态参数来对风机模型参数进行初始化，从而确定初始的风机模型参数。在本申请实施例中，在确定了风机模型的初始的输入和期望输出的情况下，确定风机模型的参数。
这时，初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数。
在一个优选实施例中，在确定初始的风机模型参数时还进一步考虑初始的风机模型参数本身。此时，初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数，同时初始的风机模型参数的总大小被加权后的结果尽量小。这里，初始的风机模型参数的总大小可通过各个初始的风机模型参数的绝对值之和表示，或者通过各个初始的风机模型参数的偶数次幂之和表示(应该理解，在风机模型只有1个(即，1种类型)风机模型参数的情况下，这里提及的“之和”为该1个初始的风机模型参数的绝对值或偶数次幂)。用于加权的权值可通过经验或实验获得。例如，可通过获取使得“输入了初始的环境条件的风机模型所输出的风机状态参数相对于初始的实际风机状态参数的变化”与“初始的风机模型参数的总大小被加权后的结果”之和最小化的风机模型参数作为初始的风机模型参数。
在该优选实施例中，通过下面的式(1)来获得初始的实际风机状态参数：
Figure PCTCN2017105440-appb-000001
其中，x0表示初始的环境条件，y0表示初始的实际风机状态参数，f0(x0)表示输入了初始的环境条件的风机模型所输出的风机状态参数，θ0表示初始的风机模型参数，α表示规范项权重系数，
Figure PCTCN2017105440-appb-000002
表示获得使得括号内的算式最小化的θ0。这里，规范项权重系数α可通过经验或实验获得。应该理解，
Figure PCTCN2017105440-appb-000003
表示上面提到的初始的风机模型参数的总大小。
在步骤S120，周期性地根据当前环境条件和风机的当前实际风机状态参数来更新风机模型的风机模型参数。
在本申请实施例中，在对风机模型进行初始化之后，可周期性地根据当前环境条件和风机的当前实际风机状态参数获得更新后的风机模型参数作为当前周期的风机模型参数。这样，可以使得风机模型保持准确性，同时可收集历史的风机模型参数和对应的历史的环境条件。
在每个周期，获取当前环境条件和风机的当前实际风机状态参数，从而在确定了风机模型的初始的输入和期望输出的情况下，确定在每个周期的风机模型参数。
更新后的风机模型参数使得输入了当前环境条件的风机模型所输出的风机状态参数尽量接近当前实际风机状态参数。在本申请实施例中，初始的风机模型参数使得：输入了当前环境条件的风机模型所输出的风机状态参数与当前实际风机状态参数之间的差异达到最小。
在一个优选实施例中，在每个周期更新风机模型参数时还进一步考虑风机模型参数的变化。此时，在每个周期，更新后的风机模型参数使得：输入了当前环境条件的风机模型所输出的风机状态参数尽量接近当前实际风机状态参数，同时更新后的风机模型参数相对于上一次更新后的风机模型参数(即，上一周期更新的风机模型参数)的变化被加权后的结果尽量小。在本申请实施例中，更新后的风机模型参数使得：输入了当前环境条件的风机模型所输出的风机状态参数与当前实际风机状态参数之间的差异达到最小，同时更新后的风机模型参数与上一次更新后的风机模型参数的之间的差异达到最小。例如，可通过获取使得“输入了当前环境条件的风机模型所输出的风机状态参数相对于当前实际风机状态参数的变化”与“更新后的风机模型参数相对于上一次更新后的风机模型参数的变化被加权后的结果”之和最小化 的风机模型参数作为更新后的风机模型参数。
这里，可通过计算各个更新后的风机模型参数与各个上一次更新后的风机模型参数之间的相同类型风机模型参数之差的绝对值之和或偶数次幂之和作为所述变化(应该理解，在风机模型只有1个(即，1种类型)风机模型参数的情况下，这里提及的“之和”为该1种类型的风机模型参数之差的绝对值或偶数次幂)。用于加权的权值可通过经验或实验获得。
在该优选实施例中，通过下面的式(2)来获得更新后的风机模型参数：
Figure PCTCN2017105440-appb-000004
其中，xt表示当前环境条件，yt表示当前实际风机状态参数，ft(xt)表示输入了当前环境条件的风机模型所输出的风机状态参数，θt表示所述更新后的风机模型参数，θt-1表示上一次更新后的风机模型参数，α表示规范项权重系数，
Figure PCTCN2017105440-appb-000005
表示获得使得括号内的算式最小化的θt。
应该理解，(θt-θt-1)T(θt-θt-1)表示上面提到的更新后的风机模型参数相对于上一次更新后的风机模型参数的变化。
在步骤S130，根据历史的风机模型参数和对应的历史的环境条件，建立风机模型参数变化模型。
风机模型参数变化模型体现了环境条件与风机模型参数之间的关系。在本申请实施例中，在将环境条件输入到风机模型参数变化模型时，风机模型参数变化模型可估计并输出对应的风机模型参数。可基于历史的风机模型参数和对应的历史的环境条件，利用各种建模技术来建立风机模型参数变化模型，本申请不做限制，只要建立的风机模型参数变化模型能够体现环境条件与风机模型参数之间的相互关系。
在一个实施例中，利用历史的风机模型参数和对应的历史的环境条件，通过拟合来建立风机模型参数变化模型。可通过各种拟合方法来拟合出风机模型参数变化模型。另外，风机模型参数变化模型的结构可通过拟合得到(此时，通过拟合确定风机模型参数变化模型的结构和参数)或者也可预先确定(此时，通过拟合仅确定风机模型参数变化模型的参数)。
在一个优选实施例中，风机模型参数变化模型被表示为下面的式(3)：
θT＝A·θT-1+B·xT               (3)
其中，θT表示在时刻T的风机模型参数，θT-1表示在时刻T-1的风机模 型参数，A为风机模型参数的状态转移矩阵，B为环境条件对风机模型参数影响系数矩阵，xT为在时刻T的环境条件。应该理解，时刻T-1表示时刻T之后的一个时刻的风机模型参数。
应该理解，参数A和B可通过拟合而获得。
在使用式(3)来预测在未来某个时刻的风机模型参数时，在一个实施例中，时刻T表示该未来某个时刻，时刻T-1表示在执行风机模型参数预测或故障预测时的当前时刻；在另一个实施例中，可获取当前时刻与该未来某个时刻之间的一个或多个时刻(以下称为中间时刻)的环境条件，从所述当前时刻开始，基于式(3)使用相邻的两个时刻中的在后的一个时刻的环境条件以及在前的一个时刻的风机模型参数预测在后的一个时刻的风机模型参数，直到预测出该未来某个时刻的风机模型参数。例如，在获取两个中间时刻的环境条件的情况下，使用第一中间时刻的环境条件以及所述当前时刻的风机模型参数预测第一时刻的风机模型参数，使用第二中间时刻的环境条件以及预测的第一时刻的风机模型参数预测第二时刻的风机模型参数，使用该未来某个时刻的环境条件以及预测的第二时刻的风机模型参数预测该未来某个时刻的风机模型参数。
应该理解，风机模型参数变化模型的结构不限于式(3)所表达的函数结构，也可通过拟合过程来确定或者采用其他的结构。
可基于步骤S120来获得历史的风机模型参数和对应的历史的环境条件。历史的风机模型参数和对应的历史的环境条件可表示通过步骤S120获得的过去一段时间内各个周期的风机模型参数和对应的环境条件；或者通过步骤S110和S120已经获得的所有的风机模型参数和对应的环境条件。
在步骤S140，基于未来时刻的未来环境条件，使用风机模型参数变化模型预测在所述未来时刻的风机模型参数。
当希望预测未来时刻(即，未来的某个时间)的风机故障时，可获取在该未来时刻的环境条件(以下，称为未来环境条件)，并将获取的未来环境条件输入到风机模型参数变化模型，来预测在所述未来时刻的风机模型参数。
可通过各种方式来获取在该未来时刻的环境条件，例如，通过天气预报、环境模型等。本申请对获取在该未来时刻的环境条件的方式不进行限定。
在步骤S150，根据历史的风机模型参数和对应的历史的风机故障，建立风机故障模型。风机故障模型体现了风机模型参数与风机故障之间的关系。 在本申请实施例中，在将风机模型参数输入到风机故障模型时，风机故障模型可估计并输出对应的风机故障。
可基于历史的风机模型参数和对应的历史的风机故障，利用各种建模技术来建立风机故障模型，本申请不做限制，只要建立的风机故障模型能够体现风机模型参数与风机故障之间的关系。
可在风机发生故障时，获取当前风机模型的风机模型参数，从而得到历史的风机模型参数和对应的历史的风机故障。在本申请实施例中，这里的“历史的风机故障”指示是否发生故障。此时，获得风机故障模型可基于输入的风机模型参数输出指示是否发生故障的信息。例如，可将风机模型参数作为特征，将是否发生故障作为分类标签，使用历史的风机模型参数和对应的历史的风机故障作为训练样本来训练分类器作为风机故障模型。此时，风机故障模型是以风机模型参数为特征并且以是否发生故障为分类标签的分类器。
此外，在风机发生故障时，还可进一步获取故障类别。在本申请实施例中，这里的“历史的风机故障”能够指示故障类别。此时，当预测出故障时，风机故障模型可输出指示发生的故障的类别的信息。例如，可将风机模型参数作为特征，将故障类别作为分类标签，使用历史的风机模型参数和对应的历史的风机故障作为训练样本来训练分类器作为风机故障模型。此时，风机故障模型是以风机模型参数为特征并且以故障类别为分类标签的分类器。
优选地，建立风机故障模型时使用的历史的风机模型参数和对应的历史的风机故障可以是每次当要预测风机故障时到目前为止已经收集到的历史的风机模型参数和对应的历史的风机故障。
在步骤S160，根据预测的风机模型参数和风机故障模型，预测未来时刻的风机故障。
在已经在步骤S140得到所述未来时刻的风机模型参数后，将得到的风机模型参数输入已经建立的风机故障模型，从而风机故障模型预测未来时刻的风机故障。根据上面提到的风机故障模型的不同，预测的故障可以仅指示是否发生故障或者可指示故障的种类。
应该理解，步骤(S130、S150)的执行顺序以及步骤(S140、S150)的执行顺序可以同时进行或互换。此外，步骤S130和步骤S150可在确定要预测风机故障时被执行以基于最新的数据来执行建模。此外，也可预先执行步骤S130和步骤S150从而预先建模。
下面参照图2描述根据本申请的实施例的预测风机的故障的设备。
图2示出根据本申请的实施例的预测风机的故障的设备的框图。
如图2所示，根据本申请的实施例的预测风机的故障的设备200包括初始化单元210、参数更新单元220、参数模型建模单元230、参数预测单元240、故障模型建模单元250、故障预测单元260。
初始化单元210初始化风机的风机模型。
风机模型体现了风机所处的环境条件与风机的状态参数(以下称为，风机状态参数)之间的关系。在本申请实施例中，在将环境条件输入到风机模型时，风机模型可估计并输出对应的风机状态参数。
可以通过获取实测的环境条件以及实测的风机状态参数来建立风机模型。此外，也可利用经验模型来作为风机模型。可利用各种建模技术来建立风机模型，本申请不做限制，只要建立的风机模型能够体现环境条件与风机状态参数之间的相互关系。
风机的环境条件是对风机的运行状态产生影响的环境因素。
环境条件可包括(但不限于)下面参数中的至少一个：风速、温度、湿度、空气密度等。可以利用各种用于检测环境状况的传感器来实测环境条件。
风机状态参数可包括(但不限于)下面参数中的至少一个：当前转速、偏航方向、变桨角度、输出功率等。例如，当风机为风力发电机时，风机状态参数还可包括发电功率等电气参数。
初始化单元210可根据初始的环境条件和初始的实际风机状态参数，来确定初始的风机模型参数。在本申请实施例中，当进行风机模型的初始化时，获取当时的环境条件作为初始的环境条件，并检测实际的风机状态参数作为初始的实际风机状态参数。在此情况下，初始化单元210根据初始的环境条件和初始的实际风机状态参数来对风机模型参数进行初始化，从而确定初始的风机模型参数。在本申请实施例中，在确定了风机模型的初始的输入和期望输出的情况下，确定风机模型的参数。
这时，初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数。在本申请实施例中，初始的风机模型参数使得：输入了当前环境条件的风机模型所输出的风机状态参数与当前实际风机状态参数之间的差异达到最小。
在一个优选实施例中，在确定初始的风机模型参数时还进一步考虑初始 的风机模型参数本身。此时，初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数，同时初始的风机模型参数的总大小被加权后的结果尽量小。在本申请实施例中，更新后的风机模型参数使得：输入了当前环境条件的风机模型所输出的风机状态参数与当前实际风机状态参数之间的差异达到最小，同时更新后的风机模型参数与上一次更新后的风机模型参数的之间的差异达到最小。例如，可通过获取使得“输入了当前环境条件的风机模型所输出的风机状态参数相对于当前实际风机状态参数的变化”与“更新后的风机模型参数相对于上一次更新后的风机模型参数的变化被加权后的结果”之和最小化的风机模型参数作为更新后的风机模型参数。
这里，初始的风机模型参数的总大小可通过各个初始的风机模型参数的绝对值之和表示，或者通过各个初始的风机模型参数的偶数次幂之和表示(应该理解，在风机模型只有1个(即，1种类型)风机模型参数的情况下，这里提及的“之和”为该1种类型的风机模型参数之差的绝对值或偶数次幂)。用于加权的权值可通过经验或实验获得。
参数更新单元220周期性地根据当前环境条件和风机的当前实际风机状态参数来更新风机模型的风机模型参数。
在本申请实施例中，在对风机模型进行初始化之后，参数更新单元220可周期性地根据当前环境条件和风机的当前实际风机状态参数获得更新后的风机模型参数作为当前周期的风机模型参数。这样，可以使得风机模型保持准确性，同时可收集历史的风机模型参数和对应的历史的环境条件。
在每个周期，参数更新单元220获取当前环境条件和风机的当前实际风机状态参数，从而在确定了风机模型的初始的输入和期望输出的情况下，确定在每个周期的风机模型参数。
更新后的风机模型参数使得输入了当前环境条件的风机模型所输出的风机状态参数尽量接近当前实际风机状态参数。
在一个优选实施例中，在每个周期更新风机模型参数时还进一步考虑风机模型参数的变化。此时，在每个周期，更新后的风机模型参数使得输入了当前环境条件的风机模型所输出的风机状态参数尽量接近当前实际风机状态参数，同时更新后的风机模型参数相对于上一次更新后的风机模型参数(即，上一周期更新的风机模型参数)的变化被加权后的结果尽量小。这里，可通 过计算各个更新后的风机模型参数与各个上一次更新后的风机模型参数之间的相同类型风机模型参数之差的绝对值之和或偶数次幂之和来表示所述变化(应该理解，在风机模型只有1个(即，1种类型)风机模型参数的情况下，这里提及的“之和”为该1种类型的风机模型参数之差的绝对值或偶数次幂)。用于加权的权值可通过经验或实验获得。例如，可通过获取使得“输入了当前环境条件的风机模型所输出的风机状态参数相对于当前实际风机状态参数的变化”与“更新后的风机模型参数相对于上一次更新后的风机模型参数的变化被加权后的结果”之和最小化的风机模型参数作为更新后的风机模型参数。
在该优选实施例中，通过上面的式(2)来获得更新后的风机模型参数。
参数模型建模单元230根据历史的风机模型参数和对应的历史的环境条件，建立风机模型参数变化模型。
风机模型参数变化模型体现了环境条件与风机模型参数之间的关系。在本申请实施例中，在将环境条件输入到风机模型参数变化模型时，风机模型参数变化模型可估计并输出对应的风机模型参数。可基于历史的风机模型参数和对应的历史的环境条件，利用各种建模技术来建立风机模型参数变化模型，本申请不做限制，只要建立的风机模型参数变化模型能够体现环境条件与风机模型参数之间的相互关系。
在一个实施例中，利用历史的风机模型参数和对应的历史的环境条件，通过拟合来建立风机模型参数变化模型。可通过各种拟合方法来拟合出风机模型参数变化模型。另外，风机模型参数变化模型的结构可通过拟合得到(此时，通过拟合确定风机模型参数变化模型的结构和参数)或者也可预先确定(此时，通过拟合仅确定风机模型参数变化模型的参数)。在一个优选实施例中，风机模型参数变化模型被表示为上面的式(3)。
可基于参数更新单元220来获得历史的风机模型参数和对应的历史的环境条件。历史的风机模型参数和对应的历史的环境条件可表示通过参数更新单元220获得的过去一段时间内的风机模型参数和对应的环境条件；或者通过初始化单元210和参数更新单元220已经获得的所有的风机模型参数和对应的环境条件。
参数预测单元240基于未来时刻的未来环境条件，使用风机模型参数变化模型预测在所述未来时刻的风机模型参数。
当希望预测未来时刻(即，未来的某个时间)的风机故障时，参数预测单元240可获取在该未来时刻的环境条件(以下，称为未来环境条件)，并将获取的未来环境条件输入到风机模型参数变化模型，来预测在所述未来时刻的风机模型参数。
可通过各种方式来获取在该未来时刻的环境条件，例如，通过天气预报、环境模型等。本申请对获取在该未来时刻的环境条件的方式不进行限定。
故障模型建模单元250根据历史的风机模型参数和对应的历史的风机故障，建立风机故障模型。风机故障模型体现了风机模型参数与风机故障之间的关系。在本申请实施例中，在将风机模型参数输入到风机故障模型时，风机故障模型可估计并输出对应的风机故障。
可基于历史的风机模型参数和对应的历史的风机故障，利用各种建模技术来建立风机故障模型，本申请不做限制，只要建立的风机故障模型能够体现风机模型参数与风机故障之间的关系。
可在风机发生故障时，当前风机模型的风机模型参数被获得，从而得到历史的风机模型参数和对应的历史的风机故障。在本申请实施例中，这里的“历史的风机故障”指示是否发生故障。此时，获得风机故障模型可基于输入的风机模型参数输出指示是否发生故障的信息。例如，可将风机模型参数作为特征，将是否发生故障作为分类标签，使用历史的风机模型参数和对应的历史的风机故障作为训练样本来训练分类器作为风机故障模型。此时，风机故障模型是以风机模型参数为特征并且以是否发生故障为分类标签的分类器。
此外，在风机发生故障时，还可进一步获取故障类别。在本申请实施例中，这里的“历史的风机故障”能够指示故障类别。此时，当预测出故障时，风机故障模型可输出指示发生的故障的类别的信息。例如，可将风机模型参数作为特征，将故障类别作为分类标签，使用历史的风机模型参数和对应的历史的风机故障作为训练样本来训练分类器作为风机故障模型。此时，风机故障模型是以风机模型参数为特征并且以故障类别为分类标签的分类器。
优选地，建立风机故障模型时使用的历史的风机模型参数和对应的历史的风机故障可以是每次当要预测风机故障时到目前为止已经收集到的历史的风机模型参数和对应的历史的风机故障。
故障预测单元260根据预测的风机模型参数和风机故障模型，预测未来 时刻的风机故障。
在已经通过参数预测单元240得到所述未来时刻的风机模型参数后，将得到的风机模型参数输入已经建立的风机故障模型，从而风机故障模型预测未来时刻的风机故障。根据上面提到的风机故障模型的不同，预测的故障可以仅指示是否发生故障或者可指示故障的种类。
根据本申请的预测风机的故障的方法和设备，在对风机进行故障预测时考虑到风机与环境因素的特定关系，基于环境因素的影响建立风机模型，同时不断地更新风机模型的参数来获得历史的风机模型参数，从而建立与风机模型参数有关的风机故障模型用于故障预测，可以更精确地进行风机的故障预测。
此外，本申请还提供一种预测风机的故障的系统。所述系统包括：处理器和存储器，其中，存储器存储有计算机可读代码，当所述计算机可读代码被处理器执行时，执行上述预测风机的故障的方法。
此外，根据本申请的示例性实施例的上述方法可以被实现为计算机可读介质上的计算机程序，从而当运行该程序时，实现上述预测风机的故障的方法。
此外，根据本申请的示例性实施例的上述设备中的各个单元可被实现硬件组件或软件模块。此外，本领域技术人员可根据限定的各个单元所执行的处理，通过例如使用现场可编程门阵列(FPGA)、专用集成电路(ASIC)或处理器来实现各个硬件组件，可以通过编程技术来实现各个软件模块。
尽管已经参照其示例性实施例具体显示和描述了本申请，但是本领域的技术人员应该理解，在不脱离权利要求所限定的本申请的精神和范围的情况下，可以对其进行形式和细节上的各种改变。

Claims (26)

  1. 一种预测风机的故障的方法，其特征在于，所述方法包括：
    初始化风机的风机模型；
    周期性地根据当前环境条件和风机的当前实际风机状态参数来更新风机模型的风机模型参数；
    根据历史的风机模型参数和对应的历史的环境条件，建立风机模型参数变化模型；
    基于未来时刻的未来环境条件，使用风机模型参数变化模型预测在所述未来时刻的风机模型参数；
    根据历史的风机模型参数和对应的历史的风机故障，建立风机故障模型；
    根据预测的风机模型参数和风机故障模型，预测未来时刻的风机故障。
  2. 根据权利要求1所述的方法，其特征在于，风机模型体现了环境条件与风机状态参数之间的关系，风机模型参数变化模型体现了环境条件与风机模型参数之间的关系，风机故障模型体现了风机模型参数与风机故障之间的关系。
  3. 根据权利要求1所述的方法，其特征在于，初始化风机的风机模型的步骤包括：
    根据初始的环境条件和初始的实际风机状态参数，来确定初始的风机模型参数。
  4. 根据权利要求3所述的方法，其特征在于，初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数。
  5. 根据权利要求4所述的方法，其特征在于，初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数，同时初始的风机模型参数的总大小的加权结果尽量小。
  6. 根据权利要求5所述的方法，其特征在于，初始的风机模型参数的总大小为各个初始的风机模型参数的绝对值之和，或者各个初始的风机模型参数的偶数次幂之和。
  7. 根据权利要求1所述的方法，其特征在于，更新风机模型的风机模型参数的步骤包括：
    周期性地根据当前环境条件和风机的当前实际风机状态参数获得更新后的风机模型参数，其中，更新后的风机模型参数使得输入了当前环境条件的风机模型所输出的风机状态参数尽量接近当前实际风机状态参数。
  8. 根据权利要求7所述的方法，其特征在于，更新后的风机模型参数使得输入了当前环境条件的风机模型所输出的风机状态参数尽量接近当前实际风机状态参数，同时更新后的风机模型参数相对于上一次更新后的风机模型参数的变化的加权结果尽量小。
  9. 根据权利要求8所述的方法，其特征在于，通过计算各个更新后的风机模型参数与各个上一次更新后的风机模型参数之间的相同类型风机模型参数之差的绝对值之和或偶数次幂之和作为所述变化。
  10. 根据权利要求1或2所述的方法，其特征在于，建立风机模型参数变化模型的步骤包括：
    利用历史的风机模型参数和对应的历史的环境条件，通过拟合来建立风机模型参数变化模型。
  11. 根据权利要10所述的方法，其特征在于，风机模型参数变化模型被表示如下：
    θT＝A·θT-1+B·xT
    其中，θT表示在时刻T的风机模型参数，θT-1表示在时刻T-1的风机模型参数，A为风机模型参数的状态转移矩阵，B为环境条件对风机模型参数影响系数矩阵，xT为在时刻T的环境条件，时刻T-1表示时刻T之后的一个时刻。
  12. 根据权利要求1所述的方法，其特征在于，风机故障模型是以风机模型参数为特征并且以故障类别或是否发生故障为分类标签的分类器。
  13. 一种预测风机的故障的设备，其特征在于，所述设备包括：
    初始化单元，初始化风机的风机模型；
    参数更新单元，周期性地根据当前环境条件和风机的当前实际风机状态参数来更新风机模型的风机模型参数；
    参数模型建模单元，根据历史的风机模型参数和对应的历史的环境条件，建立风机模型参数变化模型；
    参数预测单元，基于未来时刻的未来环境条件，使用风机模型参数变化模型预测在所述未来时刻的风机模型参数；
    故障模型建模单元，根据历史的风机模型参数和对应的历史的风机故障，建立风机故障模型；
    故障预测单元，根据预测的风机模型参数和风机故障模型，预测未来时刻的风机故障。
  14. 根据权利要求13所述的设备，其特征在于，风机模型体现了环境条件与风机状态参数之间的关系，风机模型参数变化模型体现了环境条件与风机模型参数之间的关系，风机故障模型体现了风机模型参数与风机故障之间的关系。
  15. 根据权利要求13所述的设备，其特征在于，初始化单元根据初始的环境条件和初始的实际风机状态参数，来确定初始的风机模型参数。
  16. 根据权利要求15所述的设备，其特征在于，
    初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数。
  17. 根据权利要求16所述的设备，其特征在于，初始的风机模型参数使得：输入了初始的环境条件的风机模型所输出的风机状态参数尽量接近初始的实际风机状态参数，同时初始的风机模型参数的总大小的加权结果尽量小。
  18. 根据权利要求17所述的设备，其特征在于，初始的风机模型参数的总大小为各个初始的风机模型参数的绝对值之和，或者各个初始的风机模型参数的偶数次幂之和。
  19. 根据权利要求13所述的设备，其特征在于，参数更新单元周期性地根据当前环境条件和风机的当前实际风机状态参数获得更新后的风机模型参数，其中，更新后的风机模型参数使得输入了当前环境条件的风机模型所输出的风机模型状态尽量接近当前实际风机状态参数。
  20. 根据权利要求19所述的设备，其特征在于，更新后的风机模型参数使得输入了当前环境条件的风机模型所输出的风机状态参数尽量接近当前实际风机状态参数，同时更新后的风机模型参数相对于上一次更新后的风机模型参数的变化的加权结果尽量小。
  21. 根据权利要求20所述的设备，其特征在于，通过计算各个更新后的风机模型参数与各个上一次更新后的风机模型参数之间的相同类型风机模型参数之差的绝对值之和或偶数次幂之和作为所述变化。
  22. 根据权利要求13或14所述的设备，其特征在于，参数模型建模单 元利用历史的风机模型参数和对应的历史的环境条件，通过拟合来建立风机模型参数变化模型。
  23. 根据权利要22所述的设备，其特征在于，风机模型参数变化模型被表示如下：
    θT＝A·θT-1+B·xT
    其中，θT表示在时刻T的风机模型参数，θT-1表示在时刻T-1的风机模型参数，A为风机模型参数的状态转移矩阵，B为环境条件对风机模型参数影响系数矩阵，xT为在时刻T的环境条件，时刻T-1表示时刻T之后的一个时刻。
  24. 根据权利要求13所述的设备，其特征在于，风机故障模型是以风机模型参数为特征并且以故障类别或是否发生故障为分类标签的分类器。
  25. 一种预测风机的故障的系统，其特征在于，所述系统包括：
    处理器；
    存储器，存储有计算机可读代码，当所述计算机可读代码被处理器执行时，执行权利要求1至12中的任意一项所述的方法。
  26. 一种其中存储有计算机可读代码的计算机可读存储介质，当所述计算机可读代码被执行时执行权利要求1至12中的任意一项所述的方法。
PCT/CN2017/105440 2017-03-24 2017-10-10 预测风机的故障的方法和设备 WO2018171165A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
AU2017363221A AU2017363221B2 (en) 2017-03-24 2017-10-10 Method and device for predicting wind turbine fault
EP17875063.4A EP3407271A4 (en) 2017-03-24 2017-10-10 FAULT PROGNOSIS AND DEVICE FOR A FAN
US15/777,065 US11035346B2 (en) 2017-03-24 2017-10-10 Method and device for predicting wind turbine fault

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710180903.6 2017-03-24
CN201710180903.6A CN106951997B (zh) 2017-03-24 2017-03-24 预测风机的故障的方法和设备

Publications (1)

Publication Number Publication Date
WO2018171165A1 true WO2018171165A1 (zh) 2018-09-27

Family

ID=59472757

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/105440 WO2018171165A1 (zh) 2017-03-24 2017-10-10 预测风机的故障的方法和设备

Country Status (5)

Country Link
US (1) US11035346B2 (zh)
EP (1) EP3407271A4 (zh)
CN (1) CN106951997B (zh)
AU (1) AU2017363221B2 (zh)
WO (1) WO2018171165A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113240158A (zh) * 2021-04-25 2021-08-10 国网湖北省电力有限公司 一种状态预测方法和电子设备

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106951997B (zh) 2017-03-24 2020-11-10 新疆金风科技股份有限公司 预测风机的故障的方法和设备
CN110968069B (zh) * 2018-09-28 2022-01-25 新疆金风科技股份有限公司 风力发电机组的故障预测方法、相应的装置及电子设备
CN109653961B (zh) * 2018-12-14 2020-05-19 三一重能有限公司 一种风力发电机的监测方法及装置
CN110349289A (zh) * 2019-06-21 2019-10-18 东软集团股份有限公司 故障预测方法、装置、存储介质和电子设备
CN110735769A (zh) * 2019-09-18 2020-01-31 西安察柏科技咨询有限公司 一种预测风机故障的方法及装置、系统
CN110991666B (zh) * 2019-11-25 2023-09-15 远景智能国际私人投资有限公司 故障检测方法、模型的训练方法、装置、设备及存储介质
CN111047121B (zh) * 2020-03-17 2020-06-09 浙江上风高科专风实业有限公司 一种风机故障预测方法及系统
CN113494428B (zh) * 2020-03-20 2022-11-22 新疆金风科技股份有限公司 风力发电机组的故障检测方法和装置
CN113503232A (zh) * 2021-08-20 2021-10-15 西安热工研究院有限公司 一种风机运行健康状态预警方法及系统
CN114077922A (zh) * 2021-10-15 2022-02-22 华能射阳新能源发电有限公司 风电机组的故障预测方法、装置及电子设备
CN115079307B (zh) * 2022-06-09 2023-07-25 中能融合智慧科技有限公司 一种基于多模式最优集成的风电场气象预报方法和系统
CN117662395A (zh) * 2022-08-31 2024-03-08 北京金风慧能技术有限公司 风力发电机组的变桨故障预测方法和装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102721924A (zh) * 2012-06-26 2012-10-10 新疆金风科技股份有限公司 风力发电机组的故障预警方法
CN102778358A (zh) * 2012-06-04 2012-11-14 上海东锐风电技术有限公司 故障预测模型建立方法及系统、风机监测预警系统及方法
US20130073223A1 (en) * 2010-05-13 2013-03-21 University Of Cincinnati Turbine-To-Turbine Prognostics Technique For Wind Farms
CN105508149A (zh) * 2015-12-31 2016-04-20 北京金风科创风电设备有限公司 用于风力发电机组的故障检测方法及装置
CN106951997A (zh) * 2017-03-24 2017-07-14 新疆金风科技股份有限公司 预测风机的故障的方法和设备

Family Cites Families (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2502174B1 (en) * 2009-11-16 2018-06-13 Simmonds Precision Products, Inc. Data acquisition system for condition-based maintenance
US8326577B2 (en) 2011-09-20 2012-12-04 General Electric Company System and method for predicting wind turbine component failures
CN102567636B (zh) 2011-12-27 2015-01-07 华北电力大学 基于相关向量机的高精度风电场功率区间预测方法
US9529057B2 (en) * 2012-07-24 2016-12-27 Binod Shrestha Apparatus and method for out-of-step protection using the analysis of trajectories of electrical measurements in state plane
US10181101B2 (en) * 2012-12-17 2019-01-15 Arizona Board Of Regents On Behalf Of Arizona State University Support vector machine enhanced models for short-term wind farm generation forecasting
US20140324495A1 (en) * 2013-02-22 2014-10-30 Vestas Wind Systems A/S Wind turbine maintenance optimizer
CN103268366B (zh) * 2013-03-06 2017-09-29 辽宁省电力有限公司电力科学研究院 一种适用于分散式风电场的组合风电功率预测方法
CN103940611B (zh) * 2014-04-09 2016-08-17 中国水利水电科学研究院 一种风电机组变工况下滚动轴承自适应异常检测方法
CN104005917B (zh) * 2014-04-30 2016-12-07 叶翔 基于贝叶斯推理方式对风机状态进行预测的方法和系统
US20160010628A1 (en) * 2014-07-10 2016-01-14 General Electric Company System and method for determining life of a wind turbine
KR20160017681A (ko) * 2014-07-31 2016-02-17 두산중공업 주식회사 풍력플랜트 관리 시스템 및 그 방법
CN104200396B (zh) * 2014-08-26 2017-08-11 燕山大学 一种风力发电机部件故障预警方法
US10443577B2 (en) 2015-07-17 2019-10-15 General Electric Company Systems and methods for improved wind power generation
CN105550943A (zh) * 2016-01-18 2016-05-04 重庆大学 一种基于模糊综合评判的风电机组状态参数异常辨识方法
CN105719002A (zh) * 2016-01-18 2016-06-29 重庆大学 一种基于组合预测的风电机组状态参数异常辨识方法
CN105760617A (zh) * 2016-03-07 2016-07-13 华北电力大学（保定） 一种应用于风力发电机组的多参量故障预测判定指标的计算方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130073223A1 (en) * 2010-05-13 2013-03-21 University Of Cincinnati Turbine-To-Turbine Prognostics Technique For Wind Farms
CN102778358A (zh) * 2012-06-04 2012-11-14 上海东锐风电技术有限公司 故障预测模型建立方法及系统、风机监测预警系统及方法
CN102721924A (zh) * 2012-06-26 2012-10-10 新疆金风科技股份有限公司 风力发电机组的故障预警方法
CN105508149A (zh) * 2015-12-31 2016-04-20 北京金风科创风电设备有限公司 用于风力发电机组的故障检测方法及装置
CN106951997A (zh) * 2017-03-24 2017-07-14 新疆金风科技股份有限公司 预测风机的故障的方法和设备

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113240158A (zh) * 2021-04-25 2021-08-10 国网湖北省电力有限公司 一种状态预测方法和电子设备

Also Published As

Publication number Publication date
US11035346B2 (en) 2021-06-15
AU2017363221A1 (en) 2018-10-11
EP3407271A4 (en) 2019-05-01
CN106951997A (zh) 2017-07-14
US20200309099A1 (en) 2020-10-01
EP3407271A1 (en) 2018-11-28
AU2017363221B2 (en) 2019-08-29
CN106951997B (zh) 2020-11-10

Similar Documents

Publication Publication Date Title
WO2018171165A1 (zh) 预测风机的故障的方法和设备
JP7069269B2 (ja) デジタル・ツイン・シミュレーション・データを利用した時系列データに基づく、大規模な産業用監視システム向けの半教師あり深層異常検出のための方法およびシステム
AU2019413696B2 (en) System and method for evaluating models for predictive failure of renewable energy assets
CN104915747B (zh) 一种发电机组的发电性能评估方法及设备
EP3364324B1 (en) Method and device for detecting equivalent load of wind turbine generator system
US20220412318A1 (en) Method and System for Building Prescriptive Analytics to Prevent Wind Turbine Failures
Byon et al. Adaptive learning in time-variant processes with application to wind power systems
JP6543028B2 (ja) 異常診断装置及び異常診断方法
CN110006552B (zh) 一种机组设备温度异常检测方法
CN109242215B (zh) 一种基于粒子群-支持向量机的旋转机械设备运行工况预测方法
CN115616423B (zh) 液冷储能系统以及状态检测方法、装置
US20220178353A1 (en) A method for estimating remaining useful life of components of an operational wind turbine
CN115314235A (zh) 用于风力涡轮控制系统中的网络攻击检测的系统和方法
CN117212077A (zh) 一种风电机的风轮故障监测方法、装置、设备及存储介质
US20180087489A1 (en) Method for windmill farm monitoring
US20230221694A1 (en) Method and computing system for performing a prognostic health analysis for an asset
EP3923213A1 (en) Method and computing system for performing a prognostic health analysis for an asset
Dinh et al. Implementation of Digital Twin-Assisted Condition Monitoring and Fault Diagnosis for Wind Turbines
CN115329639B (zh) 一种风电站维护方法、平台和电子设备
US20190006850A1 (en) Method for forecasting the power daily generable by a solar inverter
US12001180B2 (en) Condition-based method for malfunction prediction
KR102553193B1 (ko) 태양광 발전량 예측 시스템 및 방법
Huang et al. A Similarity-Based Remaining Useful Life Prediction Method for Aero Engines with Small Smples

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 2017875063

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2017875063

Country of ref document: EP

Effective date: 20180522

ENP Entry into the national phase

Ref document number: 2017363221

Country of ref document: AU

Date of ref document: 20171010

Kind code of ref document: A

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

Ref document number: 17875063

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE