A wind power blade icing state intelligent identification and early warning method

Abstract

The application provides a kind of wind power blade icing state intelligent identification and early warning method, by extracting wind turbine recent set time period historical operation data, and after sufficient cleaning to data, wind turbine current operation wind speed-power curve and wind speed-pitch angle curve are obtained by fitting;At the same time, after the environment reaches icing condition, the wind turbine operation data in the recent preset time period is extracted at a shorter time interval, and after data cleaning, the remaining data is compared with the wind speed-power curve or wind speed-pitch angle curve obtained by fitting, the blade icing degree is determined and the warning is given;The application can more timely identify the blade icing and icing degree, and give timely warning, which can effectively reduce the operation risk of wind turbine under the blade icing state.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wind power generation, and particularly relates to a wind power blade icing state intelligent identification and early warning method. BACKGROUND

[0002] Due to actual needs, some wind turbines are installed in cold mountainous areas. When the temperature is low in winter, there is a great possibility of icing on the blades. The operation of the blades in the icing state will damage the original blade aerodynamic shape and increase the blade mass, which will reduce the power generation efficiency of the unit and cause the unbalanced rotation of the wind wheel. When icing is serious, it may even cause blade overload fracture and unit collapse accidents. Therefore, it is of great significance to identify the blade icing state as soon as possible and take corresponding measures.

[0003] In the early stage, it is difficult to detect whether the blade is iced through effective means. Generally, maintenance personnel need to be dispatched to observe the wind turbine in the field, but the iced wind turbine is mostly located in the winter mountainous area at this time, and the cold climate and snow will bring great safety hazards to the maintenance personnel when they go up the mountain. Therefore, new icing detection methods have been continuously produced in recent years. The detection methods mainly include two types:

[0004] One type is to install additional sensors or detection equipment on the wind turbine, and to detect the icing of the blade through image recognition or based on the change of the blade natural frequency measured by the acceleration sensor. This will undoubtedly increase the detection cost.

[0005] Another type of detection method is to compare the actual wind speed, speed, power and theoretical value of the wind turbine during operation. When the difference exceeds a certain threshold value, it is considered that the blade is iced. This method does not require the installation of additional hardware equipment and has a lower cost, but this method cannot determine the specific degree of blade icing. Moreover, this method ignores the influence of different icing degrees of the blade and different operating intervals of the unit on the operating parameters of the unit, and also ignores the change of the parameter theoretical value caused by the performance degradation of the wind turbine itself. Therefore, the diagnostic accuracy will be reduced. SUMMARY

[0006] The purpose of the present application is to provide a wind power blade icing state intelligent identification and early warning method to make up for the shortcomings of the existing wind power blade icing state identification and early warning method.

[0007] In order to achieve the above purpose, the technical scheme adopted by the present application is:

[0008] The wind power blade icing state intelligent identification and early warning method provided by the present application comprises the following steps:

[0009] Step 1, obtaining wind turbine design parameters and historical operation data in a set time period, wherein the design parameters include cut-in wind speed, cut-out wind speed, rated wind speed, rated power, rated rotor speed and blade pitch angle at full pitch, and the historical operation data include unit operation state, wind speed, active power, rotor speed, blade pitch angle, etc.

[0010] Step 2, obtaining wind speed-power curve and wind speed-pitch angle curve respectively according to the obtained historical operation data;

[0011] Step 3, setting icing environment temperature threshold, total number of data points in the recent preset time period, maximum wind speed threshold in low wind speed area, icing early warning proportion threshold, icing three-level early warning parameter loss percentage threshold, icing two-level early warning parameter loss percentage threshold and icing one-level early warning parameter loss percentage threshold;

[0012] Step 4, obtaining current environment temperature, wherein if the current environment temperature is less than the icing environment temperature threshold, obtaining current operation data of the wind turbine in a preset time, and the current operation data includes unit operation state, wind speed, active power, rotor speed and blade pitch angle;

[0013] Step 5, comparing the obtained current operation data with the wind speed-power curve and wind speed-pitch angle curve obtained in step 2 respectively, identifying the icing state of the wind turbine blade according to the comparison result and pushing the early warning information.

[0014] Preferably, the wind speed-power curve and the wind speed-pitch angle curve are obtained respectively according to the obtained historical operation data, and the specific method is as follows:

[0015] S21, binning the obtained historical operation data at a set wind speed interval to obtain multiple wind speed interval data;

[0016] S22, cleaning the obtained wind speed interval data to obtain each sub wind speed interval;

[0017] S23, data fitting the data of each sub wind speed interval to obtain wind speed-power curve and wind speed-pitch angle curve respectively.

[0018] Preferably, before binning the historical operation data, the data of non-normal power generation such as shutdown, idling, fault, maintenance and start-up in the obtained historical operation data are removed to obtain historical data of normal power generation of the wind turbine.

[0019] Preferably, in S22, the obtained multiple wind speed interval data are cleaned to obtain each sub wind speed interval, and the specific method is as follows:

[0020] calculating the power average μ and the standard deviation σ in each wind speed interval;

[0021] From each wind speed interval, data with output power less than μ-3σ or output power greater than μ+3σ is removed to obtain a corresponding sub-wind speed interval.

[0022] Preferably, S23, data fitting is performed on data of each sub-wind speed interval to obtain wind speed-power curve and wind speed-pitch angle curve respectively, and the specific method is:

[0023] The wind speed mean value, active power mean value and blade pitch angle mean value corresponding to each sub-wind speed interval are calculated to obtain multiple wind speed mean values, active power mean values and blade pitch angle mean values.

[0024] The wind speed-power curve and wind speed-pitch angle curve are respectively fitted according to the obtained multiple wind speed mean values, active power mean values and blade pitch angle mean values.

[0025] Preferably, in step 5, the obtained current operation data is compared with the wind speed-power curve and wind speed-pitch angle curve obtained in step 2 respectively, and the icing state of the wind turbine blade is identified according to the comparison result, and the specific method is:

[0026] S51, the obtained current operation data is compared with the wind speed-power curve and wind speed-pitch angle curve obtained in step 2 respectively, wherein if the proportion of data points below the wind speed-power curve or wind speed-pitch angle curve in the current operation data is less than the icing early warning proportion threshold, it is determined that the unit is normally generating electricity; otherwise, S52 is entered;

[0027] S52, the wind speed mean value in the current operation data is calculated, wherein if the wind speed mean value is less than the maximum wind speed threshold of the low wind speed interval, it is determined that the blade is icing at low wind speed, and the icing level three early warning information is pushed; otherwise, S53 is entered;

[0028] S53, when the wind speed mean value is greater than the rated wind speed, the loss percentage of each data point in the current operation data relative to the corresponding pitch angle of the pitch angle curve at the same wind speed is calculated, and the mean value of the calculated loss percentages is taken; otherwise, the loss percentage of each data point in the current operation data relative to the corresponding power of the power curve at the same wind speed is calculated, and the mean value of the calculated loss percentages is taken;

[0029] S54, the icing degree of the blade is determined according to the obtained loss percentage mean value.

[0030] Preferably, before comparing the obtained current operation data with the wind speed-power curve and wind speed-pitch angle curve obtained in step 2 respectively, the data of limited power operation and the data of abnormal power generation such as shutdown, idling, failure, maintenance and start-up in the current operation data are removed to obtain normal current operation data of the wind turbine unit.

[0031] Preferably, when the total number of data points in the obtained normal current operation data of the wind turbine is less than half of the total number of data points in the recent specified time period, no determination is made, and the program ends; otherwise, it proceeds to S51.

[0032] Preferably, in S54, the icing degree of the blade is determined according to the obtained average loss percentage, and the specific method is as follows:

[0033] When the average loss percentage is greater than the loss percentage threshold of the icing level one early warning parameter, it is determined that the blade has a severe icing phenomenon, and a level one early warning information is pushed;

[0034] When the average loss percentage is greater than the loss percentage threshold of the icing level two early warning parameter, it is determined that the blade has a moderate icing phenomenon, and a level two early warning information is pushed;

[0035] When the average loss percentage is greater than the loss percentage threshold of the icing level three early warning parameter, it is determined that the blade has a mild icing phenomenon, and a level three early warning information is pushed.

[0036] An intelligent identification and early warning system for icing state of wind power blades, comprising:

[0037] A data acquisition unit is configured to acquire design parameters of a wind turbine and historical operation data in a set time period, wherein the design parameters include cut-in wind speed, cut-out wind speed, rated wind speed, rated power, rated wind rotor speed, and blade pitch angle at full pitch, and the historical operation data includes unit operation state, wind speed, active power, wind rotor speed, blade pitch angle, etc.

[0038] A curve fitting unit is configured to obtain wind speed-power curve and wind speed-pitch angle curve respectively according to the obtained historical operation data.

[0039] A threshold preset unit is configured to set icing environment temperature threshold, total number of data points in a recent preset time period, maximum wind speed threshold in low wind speed area, icing early warning proportion threshold, loss percentage threshold of icing level three early warning parameter, loss percentage threshold of icing level two early warning parameter, and loss percentage threshold of icing level one early warning parameter.

[0040] A data acquisition unit is configured to acquire current environment temperature, wherein if the current environment temperature is less than the icing environment temperature threshold, current operation data of the wind turbine in a preset time period is acquired, and the current operation data includes unit operation state, wind speed, active power, wind rotor speed, and blade pitch angle.

[0041] A state identification and early warning unit is configured to compare the obtained current operation data with the obtained wind speed-power curve and wind speed-pitch angle curve respectively, identify the icing state of the wind power blade according to the comparison result, and push early warning information.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention provides an intelligent identification and early warning method for wind turbine blade icing. This method extracts historical operating data from a wind turbine within a recently set time period, cleans the data, and then fits the wind turbine's current operating wind speed-power curve and wind speed-pitch angle curve. Furthermore, after the environment reaches icing conditions, the wind turbine's operating data from the recently set time period is extracted at short intervals. After cleaning the data, the remaining data is compared with the fitted wind speed-power curve or wind speed-pitch angle curve to determine the blade icing level and issue an early warning. The proposed method only requires analysis of the wind turbine's existing operating data, eliminating the need for additional detection equipment or sensors and incurring additional costs. Furthermore, the method considers the impact of different wind turbine operating ranges and blade icing levels on various wind turbine operating parameters. Compared to the wind turbine's own Supervisory Control and Data Acquisition (SCADA) system, it can identify blade icing and its level earlier and issue timely early warnings, effectively reducing the operational risk of wind turbines in the presence of blade icing. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 is a flow chart of the method of the present invention;

[0045] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0046] The present invention will be further described below with reference to the accompanying drawings.

[0047] like Figure 1 As shown, the present invention provides a method for intelligently identifying and warning the icing status of wind turbine blades, which specifically includes the following steps:

[0048] Step 1: Read the wind turbine's design parameters, including the cut-in wind speed (line 1), cut-out wind speed (line 2), rated wind speed (line 3), rated power (line 4), rated rotor speed (line 5), and blade pitch angle when the blades are fully opened (line 7). At regular intervals, extract the wind turbine's 10-minute operating data, including the unit's operating status, wind speed, active power, rotor speed, and blade pitch angle, from the SCADA system over the past three months.

[0049] Extract 10 min level data is for subsequent power curve fitting, as wind turbine power curve usually does not change significantly in a short time, so the data extraction time interval is set to half a month or a month. Each time the data should be able to guarantee that the wind turbine output power is less than the rated power, and the data points in each 0.5 m / s wind speed interval are not less than 30, so as to ensure that the power curve obtained by fitting is accurate enough. Usually, the specified time is set to 3 months to meet this requirement.

[0050] Step 2, data cleaning, removing the wind turbine operation data extracted in step 1, such as shutdown, idling, failure, maintenance, start-up and other abnormal power generation data to obtain wind turbine normal power generation state data;

[0051] During data cleaning, shutdown, idling, failure, maintenance, start-up and other states can be directly identified through SCADA data unit operation state.

[0052] Step 3, divide the data into bins with each 0.5 m / s as a wind speed interval, calculate the power average (μ) and standard deviation (σ) in each wind speed interval, remove large range outliers with output power less than μ-3σ and output power greater than μ+3σ in each wind speed interval, and obtain each sub wind speed interval;

[0053] Calculate the average wind speed, active power average and blade pitch angle average of each data point in each sub wind speed interval, save wind speed-power and wind speed-pitch angle data and fit the wind speed-power curve and wind speed-pitch angle curve of the unit in the last 3 months through difference;

[0054] Icing on the blades will affect the aerodynamic shape of the blades, resulting in a decrease in the output power of the unit at the same wind speed. The unit needs to reach the rated output power at a higher wind speed, resulting in a decrease in the blade pitch angle corresponding to the rated power data point. Therefore, the degree of icing on the blades can be determined by comparing the difference between the current short-term wind turbine output power, pitch angle and the theoretical output power and pitch angle at the same wind speed without icing.

[0055] There is no identification and control strategy for blade icing state in the wind turbine SCADA system. When the output power is far less than the initial design or initial installation and operation of the wind turbine under the same wind speed, the unit is directly judged as failure and shutdown. At this time, the blade icing condition is usually very serious, and the unit cannot automatically identify the early icing of the blade and the icing degree, so it cannot take measures to prevent the wind turbine from running below the dangerous working condition. In addition, with the increase of the running time of the wind turbine, the power generation performance of the wind turbine will degenerate to a certain extent, which also leads to the change of the wind speed-power curve. At the same time, the actual operation wind speed-power curve of the wind turbine in different seasons, different environments and different air densities also has certain differences. Therefore, in order to accurately judge whether the blade is iced, steps 1, 2 and 3 need to be analyzed for each wind turbine to obtain the latest fitted wind speed-power curve and wind speed-pitch angle curve of each unit, and to take them as the benchmark for blade icing diagnosis and early warning, so as to effectively improve the diagnosis accuracy.

[0056] Step 4, setting icing environment temperature threshold n1, total number of data points n2 in the last 2 hours, maximum wind speed threshold n3 in the low wind speed area, icing early warning proportion threshold n4, icing three-level early warning parameter loss percentage threshold n5, icing two-level early warning parameter loss percentage threshold n6, and icing one-level early warning parameter loss percentage threshold n7;

[0057] Here, the icing one-level early warning, two-level early warning and three-level early warning respectively refer to heavy icing, moderate icing and light icing of the blade.

[0058] Step 5, when the environment temperature is less than the icing environment temperature threshold n1, starting the icing early warning model, and extracting 1min-level data of the unit operation state, wind speed, active power, wind wheel speed, blade pitch angle and other parameters of the wind turbine in the last 2 hours from the SCADA system every fixed time;

[0059] Here, the interval of data extraction can be selected according to the icing degree and the icing speed, for example, when the blade icing degree is more serious or the icing speed is faster, a shorter interval time is selected and corresponding measures are taken to prevent the influence of serious blade icing on the operation stability of the unit.

[0060] Step 6, removing the data of shutdown, idling, failure, start-up and other abnormal power generation in the wind turbine operation data extracted in step 5 to obtain the normal power generation state data x of the wind turbine in the last 2 hours; when the data point number of x is less than n2 / 2, it is determined that the "data amount is insufficient", and the determination is temporarily not performed; otherwise, step 7) is entered;

[0061] Step 7, compare the data x with the wind speed-power curve and the wind speed-pitch angle curve obtained in step 3 within the last 3 months, if the proportion of data points below the wind speed-power curve or the wind speed-pitch angle curve in the data x is less than the icing early warning proportion threshold n4, it is determined that the unit is normally generating electricity, otherwise it is determined that the unit is icing and step 8 is entered;

[0062] Step 8, calculate the average wind speed in the data x, when the average wind speed is less than the maximum wind speed threshold n3 in the low wind speed region, it is determined that the blade is icing at low wind speed, and the monitoring system is pushed: "icing level three warning: blade icing at low wind speed"; otherwise, step 9 is entered;

[0063] The blade icing will reduce the wind turbine torque and further reduce the wind turbine speed. Therefore, in the low wind speed region, severe icing of the blade will cause the wind turbine speed to fail to reach the unit grid-connected speed, that is, the unit will be in an idle state without power output. If there is active power output in the low wind speed region, the icing degree is usually relatively light, and at the same time the load and vibration of the low wind speed region wind turbine are small, and the safety risk is relatively small, so the third level icing warning can be directly given here.

[0064] Step 9, when the average wind speed in the data x is greater than the rated wind speed line3, calculate the loss percentage of each data point in x relative to the corresponding pitch angle of the pitch angle curve at the same wind speed, and take the average of the calculated loss percentages; when the average wind speed in the data x is less than or equal to the rated wind speed line3, calculate the loss percentage of each data point in x relative to the corresponding power of the power curve at the same wind speed, and take the average of the calculated loss percentages;

[0065] Under normal circumstances, the wind turbine is in a fully open pitch state when the wind speed is less than the rated wind speed, and the pitch angle remains unchanged. When the wind speed is greater than the rated wind speed, the unit output power reaches the rated output power, and the unit maintains stable power output by changing the blade pitch angle. Therefore, step 9 determines the main operating interval of the wind turbine in the last two hours by determining the relative size of the average wind speed of the data x and the rated wind speed, so as to determine whether to use the pitch angle or the output power for analysis, and ensure that the proposed icing state intelligent identification method is applicable to each operating interval of the wind turbine.

[0066] Step 10, when the loss percentage calculated in step 9 is greater than the icing level one warning parameter loss percentage threshold n7, it is determined that the blade is severely icing, and the monitoring system is pushed: "level one warning: blade severe icing";

[0067] When the loss percentage average is less than the icing first-level early warning parameter loss percentage threshold n7 and greater than the icing second-level early warning parameter loss percentage threshold n6, it is determined that the blade is moderately iced, and a "second-level warning: blade moderate icing" is pushed to the monitoring system;

[0068] When the loss percentage average is less than the icing second-level early warning parameter loss percentage threshold n6 and greater than the icing third-level early warning parameter loss percentage threshold n5, it is determined that the blade is lightly iced, and a "third-level warning: blade light icing" is pushed to the monitoring system.

[0069] As shown in Figure 2 The present application provides a wind turbine blade icing state intelligent identification and early warning system, which comprises:

[0070] A data acquisition unit is configured to acquire wind turbine design parameters and historical operation data in a set time period, wherein the design parameters include cut-in wind speed, cut-out wind speed, rated wind speed, rated power, rated wind rotor speed, and blade pitch angle at full-pitch, and the historical operation data includes unit operation state, wind speed, active power, wind rotor speed, blade pitch angle, etc.

[0071] A curve fitting unit is configured to obtain wind speed-power curve and wind speed-pitch angle curve respectively according to the obtained historical operation data.

[0072] A threshold preset unit is configured to set icing environment temperature threshold, total number of data points in a preset recent time period, maximum wind speed threshold in low wind speed area, icing early warning proportion threshold, icing third-level early warning parameter loss percentage threshold, icing second-level early warning parameter loss percentage threshold, and icing first-level early warning parameter loss percentage threshold.

[0073] A data acquisition unit is configured to acquire current environment temperature, wherein if the current environment temperature is less than the icing environment temperature threshold, current operation data of the wind turbine in a preset time is acquired, and the current operation data includes unit operation state, wind speed, active power, wind rotor speed, and blade pitch angle.

[0074] A state identification and early warning unit is configured to compare the obtained current operation data with the obtained wind speed-power curve and wind speed-pitch angle curve respectively, identify the wind turbine blade icing state according to the comparison result, and push early warning information.

[0075] The state identification and early warning unit comprises:

[0076] The threshold identification early warning unit is configured to compare the obtained current operation data with the obtained wind speed-power curve and wind speed-pitch angle curve respectively, wherein if the proportion of data points below the wind speed-power curve or wind speed-pitch angle curve in the current operation data is less than the icing early warning proportion threshold, it is determined that the unit is normally generating electricity; otherwise, the wind speed identification early warning unit is entered;

[0077] The wind speed identification early warning unit is configured to calculate the wind speed average in the current operation data, wherein if the wind speed average is less than the maximum wind speed threshold in the low wind speed area, it is determined that the blade is icing at a low wind speed, and the early warning of "icing level three: blade icing at a low wind speed" is pushed; otherwise, the loss percentage calculation unit is entered;

[0078] The loss percentage calculation unit is configured to, when the wind speed average is greater than the rated wind speed, calculate the loss percentage of each data point in the current operation data relative to the corresponding pitch angle of the pitch angle curve at the same wind speed, and take the average of the calculated loss percentages; otherwise, calculate the loss percentage of each data point in the current operation data relative to the corresponding power of the power curve at the same wind speed, and take the average of the calculated loss percentages;

[0079] The loss percentage identification early warning unit is configured to determine the icing degree of the blade according to the obtained loss percentage average.

Claims

1. A method for intelligent identification and early warning of wind turbine blade icing status, characterized in that: The following steps are involved: Step 1: Acquire wind turbine design parameters and historical operating data within a set time period, wherein the design parameters include cut-in wind speed, cut-out wind speed, rated wind speed, rated power, rated rotor speed, and blade pitch angle when the rotor is fully open; and the historical operating data includes turbine operating status, wind speed, active power, rotor speed, and blade pitch angle; Step 2: Obtain a wind speed-power curve and a wind speed-pitch angle curve based on the obtained historical operation data; Step 3: Set the icing environment temperature threshold, the total number of data points in the most recent preset time period, the maximum wind speed threshold in the low wind speed zone, the icing warning ratio threshold, the icing level 3 warning parameter loss percentage threshold, the icing level 2 warning parameter loss percentage threshold, and the icing level 1 warning parameter loss percentage threshold. Step 4: Obtain the current ambient temperature. If the current ambient temperature is less than the icing ambient temperature threshold, obtain the current operating data of the wind turbine within a preset time. The current operating data includes the turbine operating status, wind speed, active power, rotor speed, and blade pitch angle. Step 5: Compare the current operating data obtained with the wind speed-power curve and wind speed-pitch angle curve obtained in step 2, identify the icing status of the wind turbine blades based on the comparison results, and send a warning message; In step 5, the current operating data obtained is compared with the wind speed-power curve and wind speed-pitch angle curve obtained in step 2, and the icing state of the wind turbine blade is identified based on the comparison results. The specific method is: S51: Compare the current operating data obtained with the wind speed-power curve and the wind speed-pitch angle curve obtained in step 2, respectively. If the proportion of data points below the wind speed-power curve or the wind speed-pitch angle curve in the current operating data is less than the icing warning ratio threshold, it is determined that the unit is generating electricity normally; otherwise, proceed to S52. S52, calculating the average wind speed in the current operating data. If the average wind speed is less than the maximum wind speed threshold in the low wind speed zone, it is determined that the blades are icing at low wind speeds and a Level 3 icing warning message is issued; otherwise, the process proceeds to S53; S53: When the mean wind speed is greater than the rated wind speed, the loss percentage of each data point in the current operating data relative to the pitch angle corresponding to the pitch angle curve at the same wind speed is calculated, and the average of the calculated loss percentages is taken; otherwise, the loss percentage of each data point in the current operating data relative to the power corresponding to the power curve at the same wind speed is calculated, and the average of the calculated loss percentages is taken; S54, judging the degree of icing of the blades based on the obtained mean loss percentage; In S54, the degree of ice coverage of the blade is determined based on the obtained average loss percentage. The specific method is: When the average loss percentage is greater than the loss percentage threshold of the first-level icing warning parameter, it is determined that the blades are severely iced and a first-level warning message is sent; When the average loss percentage is greater than the loss percentage threshold of the second-level icing warning parameter, it is determined that the blades are moderately iced and a second-level warning message is sent; When the average loss percentage is greater than the loss percentage threshold of the third-level icing warning parameter, it is determined that the blades are slightly frozen and a third-level warning information is pushed.

2. The method for intelligently identifying and warning of ice-covered wind turbine blades according to claim 1, characterized in that: In step 2, the wind speed-power curve and the wind speed-pitch angle curve are obtained according to the historical operation data. The specific method is: S21, dividing the obtained historical operation data into bins according to the set wind speed interval to obtain multiple wind speed interval data; S22, cleaning the obtained wind speed interval data to obtain each sub-wind speed interval; S23 , performing data fitting on the data of each sub-wind speed interval to obtain a wind speed-power curve and a wind speed-pitch angle curve, respectively.

3. The method for intelligently identifying and warning of ice-covered wind turbine blades according to claim 2, characterized in that: Before dividing the historical operation data into bins, the abnormal power generation data of shutdown, idling, failure, maintenance and startup in the historical operation data are eliminated to obtain the historical data of normal power generation operation of the wind turbine.

4. The method for intelligently identifying and warning of ice-covered wind turbine blades according to claim 2, characterized in that: In S22, the obtained multiple wind speed interval data are cleaned to obtain each sub-wind speed interval. The specific method is: Calculate the power mean μ and standard deviation σ in each wind speed interval; From each wind speed interval obtained, data with output power less than μ-3σ or with output power greater than μ+3σ are eliminated to obtain the corresponding sub-wind speed interval.

5. The method for intelligently identifying and warning of ice-covered wind turbine blades according to claim 2, characterized in that: S23, performing data fitting on the data of each sub-wind speed interval to obtain a wind speed-power curve and a wind speed-pitch angle curve, respectively. The specific method is: Calculating the wind speed mean, active power mean, and blade pitch angle mean corresponding to each wind speed sub-interval to obtain multiple wind speed mean values, active power mean values, and blade pitch angle mean values; The wind speed-power curve and the wind speed-pitch angle curve are respectively fitted according to the obtained multiple wind speed average values, active power average values ​​and blade pitch angle average values.

6. The method for intelligently identifying and warning of ice-covered wind turbine blades according to claim 1, characterized in that: Before comparing the current operating data obtained with the wind speed-power curve and wind speed-pitch angle curve obtained in step 2, the data of limited power operation and the data of abnormal power generation such as shutdown, idling, failure, maintenance and startup in the current operating data are eliminated to obtain the normal current operating data of the wind turbine.

7. The method for intelligently identifying and warning of ice-covered wind turbine blades according to claim 6, characterized in that: When the total number of data points in the normal current operation data of the wind turbine generator obtained is less than half of the total number of data points in the latest preset time period, no determination is performed and the program ends; otherwise, the process enters S51.

8. An intelligent identification and early warning system for wind turbine blade icing status, characterized in that: The method according to claim 1, comprising: a data acquisition unit, configured to acquire design parameters of the wind turbine and historical operating data within a set time period, wherein the design parameters include cut-in wind speed, cut-out wind speed, rated wind speed, rated power, rated rotor speed, and blade pitch angle at full propeller speed; and the historical operating data includes turbine operating status, wind speed, active power, rotor speed, and blade pitch angle; A curve fitting unit is used to obtain a wind speed-power curve and a wind speed-pitch angle curve according to the obtained historical operation data; A threshold preset unit is used to set the icing environment temperature threshold, the total number of data points in the most recent preset time period, the maximum wind speed threshold in the low wind speed zone, the icing warning ratio threshold, the icing level 3 warning parameter loss percentage threshold, the icing level 2 warning parameter loss percentage threshold, and the icing level 1 warning parameter loss percentage threshold; a data acquisition unit, configured to acquire the current ambient temperature, wherein if the current ambient temperature is less than an icing ambient temperature threshold, the unit acquires current operating data of the wind turbine within a preset time period, the current operating data including the turbine operating status, wind speed, active power, rotor speed, and blade pitch angle; The status identification and early warning unit is used to compare the current operating data obtained with the wind speed-power curve and wind speed-pitch angle curve respectively, and identify the icing status of the wind turbine blades according to the comparison results and push early warning information.

Images