A wind turbine anemometer data anomaly identification method and system
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIAN THERMAL POWER RES INST CO LTD
- Filing Date
- 2022-09-08
- Publication Date
- 2026-06-26
AI Technical Summary
In existing technologies, the identification of anomalies in wind speed meter data is not accurate enough, leading to false alarms or missed alarms in various algorithms and models of wind turbines, which affects the accuracy of power generation performance assessment and early warning.
By acquiring the operating data of the wind turbine, cleaning and power compartment processing are performed, the average wind speed is calculated, anomaly level thresholds are set, anomaly data anomalies are identified, and the cause of the anomaly is determined.
It improves the accuracy of anomaly detection in anemometer data, reduces false alarms and missed alarms in wind turbine early warning systems, and is applicable to any megawatt-class wind turbine, with broad applicability.
Smart Images
Figure CN115616248B_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of wind power generation technology, specifically relating to a method and system for identifying abnormal data from wind turbine anemometers. Background Technology
[0002] Anomaly in wind turbine anemometer data is a typical fault of wind speed and direction instruments, usually manifested as a sudden increase or decrease in the wind speed measured by the anemometer. When anemometer data is abnormal, all algorithms or models related to the wind turbine that are measured by the anemometer will malfunction. For example, many algorithms and models for wind turbine power curve fitting, power generation performance evaluation, icing warning, power anomaly warning, and yaw error anomaly warning are developed based on the wind speed measured by the anemometer. If there is an anomaly in the wind speed, it will directly cause the fitted power curve to deviate from the actual power generation performance, and will also cause false alarms or missed alarms when the wind turbine issues warnings for icing, power anomalies, and yaw anomalies.
[0003] In summary, sufficiently accurate anemometer data is fundamental to ensuring the precision of various models and algorithms for wind turbines. Therefore, accurately identifying anemometer data anomalies plays a crucial role in improving the accuracy of other algorithms. However, industry-standard anemometer data anomaly early warning models are not yet mature enough, and their identification of anemometer data anomalies is not precise enough. Summary of the Invention
[0004] The purpose of this invention is to provide a method and system for identifying anomalies in wind turbine anemometer data, thereby effectively improving the accuracy of anomaly identification and reducing false alarms and missed alarms in wind turbine early warning systems.
[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:
[0006] The present invention provides a method for identifying anomalies in wind turbine anemometer data, comprising the following steps:
[0007] Step 1: Obtain the operating data of the wind turbine unit to be analyzed within a set time period;
[0008] Step 2: Clean the acquired runtime data to obtain cleaned runtime data;
[0009] Step 3: Perform power segmentation on the obtained cleaned operating data to obtain multiple power segments;
[0010] Step 4: Calculate the average wind speed of each power compartment to obtain the anemometer data anomaly direction warning value;
[0011] Step 5: Identify anomalies in the wind turbine anemometer data based on the obtained anomaly direction warning values.
[0012] Preferably, in step 1, the operating data of the wind turbine to be analyzed within a set time period is obtained, specifically by:
[0013] Data Q1 is generated by acquiring the operating data of the unit to be analyzed from the previous 60 to the previous 30 days.
[0014] Data Q2 is generated by acquiring the operating data of the unit to be analyzed over the past 30 days.
[0015] Preferably, in step 3, the obtained cleaned operating data is divided into power compartments to obtain multiple power compartments. The specific method is as follows:
[0016] The two sets of cleaned operating data are divided into power compartments, with each power interval being n2*P / n3. Each set of operating data yields multiple corresponding power compartments, where n2 is the set value of the maximum power compartment value, n3 is the set value of the number of power compartments, and P is the rated power of the unit to be analyzed.
[0017] Preferably, in step 4, the average wind speed of each power compartment is calculated to obtain the anemometer data anomaly direction warning value. The specific method is as follows:
[0018] Calculate the average wind speed of each power compartment corresponding to each set of operating data, and form two corresponding matrices;
[0019] Calculate the difference between the two matrices and form a new matrix;
[0020] Calculate the average wind speed corresponding to this matrix to obtain the anemometer data anomaly direction warning value.
[0021] Preferably, in step 5, the abnormal anemometer data of the wind turbine is identified based on the obtained anemometer data abnormality direction warning value. The specific method is as follows:
[0022] Multiple anomaly levels are set for anomaly data, and a threshold is set for each anomaly level. The multiple anomaly levels are Level 1, Level 2, and Level 3.
[0023] When the anomaly warning value of the wind speed meter data is greater than or equal to the threshold corresponding to the first level of anomaly, the anomaly meter data of the wind turbine unit to be analyzed is judged to be severely abnormal.
[0024] When the anomaly warning value of the wind turbine data is less than the threshold corresponding to the first level of anomaly and greater than or equal to the threshold corresponding to the second level of anomaly, the anomaly data of the wind turbine to be analyzed is judged to be a relatively severe anomaly.
[0025] When the anomaly warning value of the wind turbine data is less than the threshold corresponding to the second level of anomaly and greater than or equal to the threshold corresponding to the third level of anomaly, the anomaly data of the wind turbine to be analyzed is judged to be slightly abnormal.
[0026] When the anemometer data anomaly warning value is less than the threshold corresponding to the third level of anomaly, the anemometer data of the wind turbine unit to be analyzed is judged to be normal.
[0027] Preferably, step 5 is followed by a method for determining the cause of anomaly in the anemometer data, specifically:
[0028] The anemometer data anomaly direction warning value is obtained by calculating the average absolute value of the wind speed in each power compartment.
[0029] Determine the cause of the anemometer data anomaly based on the abnormal direction warning value obtained from the anemometer data.
[0030] Preferably, if the anemometer data anomaly warning value is greater than 0, the cause of the anemometer data anomaly of the wind turbine unit to be analyzed is determined to be that the wind speed is too low; otherwise, the cause of the anemometer data anomaly of the wind turbine unit to be analyzed is determined to be that the wind speed is too high.
[0031] A wind turbine anemometer data anomaly identification system includes:
[0032] The data acquisition unit is used to acquire the operating data of the wind turbine to be analyzed within a set time period, and divide the obtained operating data into two sets of data.
[0033] The data cleaning unit is used to clean the two sets of data separately to obtain two sets of cleaned operational data.
[0034] The data warehousing unit is used to perform power warehousing on each set of cleaned operating data, and each set of cleaned operating data will result in multiple corresponding power warehousing units.
[0035] The calculation unit is used to calculate the average wind speed of each power compartment corresponding to the operating data after each cleaning, and to obtain the anemometer data abnormality direction warning value;
[0036] The identification unit is used to identify anomalies in the wind turbine anemometer data based on the obtained anomaly direction warning value.
[0037] A wind turbine anemometer data anomaly identification device includes a processor and a computer program capable of running on the processor, wherein the processor executes the computer program to implement the steps of the method described above.
[0038] Compared with the prior art, the beneficial effects of the present invention are:
[0039] This invention provides a method for identifying anomalies in wind turbine anemometer data. It extracts historical operating data from the wind turbine within a set time period at fixed time intervals, divides this data into two equal groups, and cleans each group. By analyzing and comparing the wind speed magnitude and difference corresponding to the same power range in the two cleaned operating data groups, the method identifies anomalies in the wind turbine anemometer data. The proposed method only requires analysis of the basic operating parameters of the wind turbine, and can accurately identify anomalies in anemometer data and provide timely warnings with a small amount of data. It is applicable to wind turbines of any megawatt level equipped with anemometers, effectively reducing false alarms and missed alarms in wind speed-related diagnostic and warning methods, and has broad applicability. Attached Figure Description
[0040] Figure 1 This is a flowchart of the method of the present invention;
[0041] Figure 2 This is a case study of wind speed and power data under normal anemometer conditions;
[0042] Figure 3 This is a case study of wind speed and power data obtained using this method when anemometer data is abnormal. Detailed Implementation
[0043] The present invention will now be further described with reference to the accompanying drawings.
[0044] The present invention provides a method for identifying anomalies in wind turbine anemometer data, comprising the following steps:
[0045] Step 1: Obtain the operating data of the wind turbine unit to be analyzed within a set time period, and divide the obtained operating data into two sets of data;
[0046] Step 2: Clean the two sets of data separately to obtain two sets of cleaned running data;
[0047] Step 3: Perform power compartmentation on each set of cleaned operating data, resulting in multiple corresponding power compartments for each set of cleaned operating data.
[0048] Step 4: Calculate the average wind speed of each power compartment corresponding to the operating data after each cleaning, and obtain the anemometer data abnormality direction warning value;
[0049] Step 5: Identify anomalies in the wind turbine anemometer data based on the obtained anomaly direction warning values.
[0050] This invention discloses a wind turbine anemometer data anomaly identification system, comprising:
[0051] The data acquisition unit is used to acquire the operating data of the wind turbine to be analyzed within a set time period, and divide the obtained operating data into two sets of data.
[0052] The data cleaning unit is used to clean the two sets of data separately to obtain two sets of cleaned operational data.
[0053] The data warehousing unit is used to perform power warehousing on each set of cleaned operating data, and each set of cleaned operating data will result in multiple corresponding power warehousing units.
[0054] The calculation unit is used to calculate the average wind speed of each power compartment corresponding to the operating data after each cleaning, and to obtain the anemometer data abnormality direction warning value;
[0055] The identification unit is used to identify anomalies in the wind turbine anemometer data based on the obtained anomaly direction warning value.
[0056] The aforementioned wind turbine anemometer data anomaly identification device can be a desktop computer, laptop, handheld computer, or cloud server, etc. This device may include, but is not limited to, a processor and memory. ...
[0057] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. …
[0058] Example
[0059] like Figure 1 As shown, the present invention provides a method for identifying anomalies in wind turbine anemometer data, which specifically includes the following steps:
[0060] Step 1: Identify the rated power P and cut-out wind speed V of the unit to be analyzed; set the minimum data volume threshold n1, the ratio of the maximum power compartment value to the rated power n2, the number of power compartments n3, the first-level anomaly threshold n4, the second-level anomaly threshold n5, and the third-level anomaly threshold n6 involved in the automatic identification method of anomaly data anomaly.
[0061] Step 2: At fixed intervals, acquire the operating data of the unit under analysis, including wind speed, active power, rotor speed, and blade pitch angle at the 10-minute level for 30 days from the 60 days prior to the 30 days prior to the analysis, and define it as Q1; acquire the operating data of the unit under analysis, including wind speed, active power, rotor speed, and blade pitch angle at the 10-minute level for 30 days prior to the analysis, and define it as Q2; record the date of the 30th day prior to the analysis of the unit under analysis as d;
[0062] Here, the interval for the early warning method analysis should be in days, with intervals ranging from 1 to 10 days; less than 1 day will result in too much computation, and more than 10 days will reduce the accuracy of the early warning.
[0063] Step 3: Perform data cleaning on Q1 and Q2 respectively, and remove invalid data points, data points where the wind turbine is not in normal power generation state, data points where the wind speed is less than or equal to 0, data points where the wind speed is greater than V, data points where the power is less than or equal to 0, data points where power limiting control is achieved by adjusting the pitch or reducing the speed in advance, and data points where the wind speed-power data shows a large range of outliers.
[0064] Step 4: Save the wind speed and active power data of Q1 and Q2 after data cleaning in Step 3, and define them as Q3 and Q4 respectively; determine the relative size of the data volume of Q3 and Q4 with n1. If the data volume of both Q3 and Q4 is greater than or equal to n1, proceed to Step 5. If the data volume of any one or both of Q3 and Q4 is less than n1, push "Insufficient data volume" to the wind turbine monitoring system and terminate the calculation.
[0065] Step 5: Divide the Q3 and Q4 data into compartments with each power interval as n2*P / n3. Calculate the average wind speed in each power compartment in Q3 and form a matrix q1. Calculate the average wind speed in each power compartment in Q4 and form a matrix q2.
[0066] Here, the maximum power compartment value is the product of the power interval n2*P / n3 and the number of power compartments n3, which is n2*P. Under normal circumstances, wind turbines have a large amount of data in the low wind speed and low power range, while the amount of data that can reach a certain proportion of rated power is relatively small. Therefore, the maximum power compartment value relative to the rated power n2 can be set according to the actual operating conditions of the wind turbine. For example, n2 can be set to 50% in areas with poor wind conditions and 80% in areas with good wind conditions, thereby ensuring sufficient data in each power compartment.
[0067] Step 6: Calculate the difference between matrix q1 and matrix q2, form a matrix and define it as q3. Take the average value of each wind speed data in q3 to obtain the anemometer data abnormality direction warning value a; take the absolute value of each wind speed data in q3 and then average it to obtain the anemometer data abnormality warning value b.
[0068] The typical manifestation of anemometer data anomalies is a sudden increase or decrease in the wind speed measured by the anemometer compared to the actual wind speed. When the wind speed measured by the anemometer suddenly decreases at a certain moment, the wind speed measured after that moment at the same power will be significantly lower than the wind speed measured before that moment. Similarly, when the wind speed measured by the anemometer suddenly increases at a certain moment, the wind speed measured after that moment at the same power will be significantly higher than the wind speed measured before that moment. Analyzing the relative magnitudes of the wind speeds measured by the anemometer within the same power range over different time periods can determine whether the anemometer data is abnormal. Here, the differences in wind speeds measured by the anemometer within different power ranges are averaged to avoid errors.
[0069] Step 7: Determine the relative magnitude of the anemometer data anomaly warning value b with the anemometer data anomaly warning thresholds at each level. When the anemometer data anomaly warning value b is greater than or equal to the first-level anomaly threshold n4, the anemometer data is determined to be severely anomaly and proceeds to step 8. When the anemometer data anomaly warning value b is less than the first-level anomaly threshold n4 but greater than or equal to the second-level anomaly threshold n5, the anemometer data is determined to be relatively severely anomaly and proceeds to step 9. When the anemometer data anomaly warning value b is less than the second-level anomaly threshold n5 but greater than or equal to the third-level anomaly threshold n6, the anemometer data is determined to be slightly anomaly and proceeds to step 10. When the anemometer data anomaly warning value b is less than the third-level anomaly threshold n6, the anemometer data is determined to be normal and a message "The wind speed data measured by the anemometer on date d is normal" is pushed to the wind turbine monitoring system.
[0070] Step 8: When the anomaly direction judgment value 'a' of the anemometer data is greater than 0, it is determined that the wind speed measured by the anemometer is too low, and a "Level 1 warning: the wind speed data measured by the anemometer on date d is severely too low" is pushed to the wind turbine monitoring system; when the anomaly direction judgment value 'a' of the anemometer data is less than 0, it is determined that the wind speed measured by the anemometer is too high, and a "Level 1 warning: the wind speed data measured by the anemometer on date d is severely too high" is pushed to the wind turbine monitoring system.
[0071] Step 9: When the anomaly direction judgment value 'a' of the anemometer data is greater than 0, it is determined that the wind speed measured by the anemometer is too low, and a "Level 2 warning: the wind speed data measured by the anemometer on date d is moderately low" is pushed to the wind turbine monitoring system; when the anomaly direction judgment value 'a' of the anemometer data is less than 0, it is determined that the wind speed measured by the anemometer is too high, and a "Level 2 warning: the wind speed data measured by the anemometer on date d is moderately high" is pushed to the wind turbine monitoring system.
[0072] Step 10: When the anomaly direction judgment value 'a' of the anemometer data is greater than 0, it is determined that the anemometer wind speed is too low, and a "Level 3 warning: the wind speed data measured by the anemometer on date d is slightly too low" is pushed to the wind turbine monitoring system; when the anomaly direction judgment value 'a' of the anemometer data is less than 0, it is determined that the anemometer wind speed is too high, and a "Level 3 warning: the wind speed data measured by the anemometer on date d is slightly too high" is pushed to the wind turbine monitoring system.
[0073] Figure 2 This example demonstrates how this method determines wind speed and power data in Q3 and Q4 when the anemometer data is normal. From... Figure 2 It can be seen that when the anemometer data is normal, the wind speed power data in Q3 and Q4 obtained after calculations in steps 1 to 4 show good consistency. The anemometer data abnormality warning value b calculated in steps 5 to 7 is much smaller than the anemometer data abnormality level three warning threshold n6.
[0074] Figure 3 This example demonstrates how this method identifies wind speed and power data in Q3 and Q4 when anemometer data exhibits severe anomalies. From... Figure 3 It can be seen that when the anemometer data is abnormal, the wind speed power data in Q3 and Q4 obtained after calculations in steps 1 to 4 show significant deviations. The anemometer data abnormality warning value b calculated in steps 5 to 10 is significantly greater than the first-level warning threshold n4 for anemometer data abnormality, and the anemometer data abnormality direction judgment value a is greater than 0. Therefore, the warning system will issue a warning for the corresponding date: "First-level warning: The wind speed data measured by the anemometer is severely undervalued".
Claims
1. A method for identifying anomalies in wind turbine anemometer data, characterized in that, Includes the following steps: Step 1: Obtain the operating data of the wind turbine unit to be analyzed within a set time period, and divide the obtained operating data into two sets of data; Step 2: Clean the two sets of data separately to obtain two sets of cleaned running data; Step 3: Perform power compartmentation on each set of cleaned operating data, resulting in multiple corresponding power compartments for each set of cleaned operating data. Step 4: Calculate the average wind speed of each power compartment corresponding to the operating data after each cleaning, and obtain the anemometer data abnormality direction warning value; Step 5: Identify anomalies in the wind turbine anemometer data based on the obtained anomaly direction warning values. In step 4, the average wind speed of each power compartment is calculated to obtain the anemometer data anomaly direction warning value. The specific method is as follows: Calculate the average wind speed of each power compartment corresponding to each set of operating data, and form two corresponding matrices; Calculate the difference between the two matrices and form a new matrix; Calculate the average wind speed corresponding to the matrix to obtain the anemometer data anomaly direction warning value; in step 3, perform power compartmentalization on each group of cleaned operating data to obtain multiple power compartments. The specific method is as follows: With each For each power range, the cleaned operating data is divided into power compartments, resulting in multiple corresponding power compartments for each set of operating data. Here, n2 is the set value for the maximum power compartment value; n3 is the set value for the number of power compartments; and P is the rated power of the unit to be analyzed. In step 5, anomalies in the wind turbine anemometer data are identified based on the obtained anemometer data anomaly direction warning value. The specific method is as follows: Multiple anomaly levels are set for anomaly data, and a threshold is set for each anomaly level. The multiple anomaly levels are Level 1, Level 2, and Level 3. When the anomaly warning value of the wind speed meter data is greater than or equal to the threshold corresponding to the first level of anomaly, the anomaly meter data of the wind turbine unit to be analyzed is judged to be severely abnormal. When the anomaly warning value of the wind turbine data is less than the threshold corresponding to the first level of anomaly and greater than or equal to the threshold corresponding to the second level of anomaly, the anomaly data of the wind turbine to be analyzed is judged to be a relatively severe anomaly. When the anomaly warning value of the wind turbine data is less than the threshold corresponding to the second level of anomaly and greater than or equal to the threshold corresponding to the third level of anomaly, the anomaly data of the wind turbine to be analyzed is judged to be slightly abnormal. When the anemometer data anomaly warning value is less than the threshold corresponding to the third-level anomaly, the anemometer data of the wind turbine being analyzed is determined to be normal. Step 5 is followed by a method for determining the cause of the anemometer data anomaly, specifically: The anemometer data anomaly direction warning value is obtained by calculating the average absolute value of the wind speed in each power compartment. Determine the cause of the anemometer data anomaly based on the abnormal direction warning value obtained from the anemometer data.
2. The method for identifying anomalies in wind turbine anemometer data according to claim 1, characterized in that, If the anemometer data anomaly warning value is greater than 0, the cause of the anemometer data anomaly in the wind turbine unit to be analyzed is determined to be that the wind speed is too low; otherwise, the cause of the anemometer data anomaly in the wind turbine unit to be analyzed is determined to be that the wind speed is too high.
3. A wind turbine anemometer data anomaly identification system, characterized in that, The identification method based on claim 1 includes: The data acquisition unit is used to acquire the operating data of the wind turbine to be analyzed within a set time period, and divide the obtained operating data into two sets of data. The data cleaning unit is used to clean the two sets of data separately to obtain two sets of cleaned operational data. The data warehousing unit is used to perform power warehousing on each set of cleaned operating data, and each set of cleaned operating data will result in multiple corresponding power warehousing units. The calculation unit is used to calculate the average wind speed of each power compartment corresponding to the operating data after each cleaning, and to obtain the anemometer data abnormality direction warning value; The identification unit is used to identify anomalies in the wind turbine anemometer data based on the obtained anomaly direction warning value.
4. A wind turbine anemometer data anomaly identification device, comprising a processor and a computer program capable of running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-2.