Static yaw error discrimination method based on SCADA data analysis of wind turbine generator

Abstract

The invention discloses a static yaw error discrimination method based on SCADA data analysis of a wind turbine generator. The static yaw error discrimination method comprises the following steps: acquiring parameters and SCADA data of the wind turbine generator; identifying and cleaning abnormal data in the SCADA data; eliminating operation data about the influence of the wind turbine generator control strategy in the identified and cleaned SCADA data; and analyzing and judging a static yaw error value. According to the static yaw error discrimination method based on SCADA data analysis of the wind turbine generator set, the static yaw error can be effectively identified, power generation loss and unbalanced load of the wind turbine generator set caused by the static yaw error are avoided, and the problem of high cost caused by calibration of the static yaw error by a laser wind finding radar is also solved.

Description

Technical Field

[0001] The invention belongs to the technical field of wind power generation, and in particular relates to a static yaw error determination method based on SCADA data analysis of a wind turbine generator set. Background Art

[0002] As a key subsystem of a wind turbine, the yaw system is mainly used to enable the nacelle and the wind rotor to automatically adjust their direction according to the change of wind direction, ensuring that the wind rotor is always aligned with the incoming wind direction, thereby maximizing the capture of wind energy and improving power generation efficiency. In actual operation, due to many factors such as the variability of wind direction, the accuracy limitations of measurement equipment, installation and commissioning errors, and control strategies, the occurrence of yaw error is difficult to avoid. The yaw error of a wind turbine is mainly divided into dynamic yaw error caused by control strategy and static yaw error caused by measurement errors of wind direction measurement equipment, such as Figure 1 shown.

[0003] From the perspective of energy capture, accurate measurement and effective control of static yaw error can enable wind turbines to capture wind energy more efficiently and improve wind power efficiency. From the perspective of equipment safety, reducing or eliminating static yaw error can prevent wind turbines from being subjected to long-term non-frontal wind impacts, thereby reducing the unbalanced loads on components such as blades, hubs, and towers, extending the service life of the equipment, reducing the probability of equipment failure, and ensuring stable and reliable operation of wind turbines. Laser wind radar can capture wind direction changes in real time with high precision and can achieve accurate measurement of yaw error, but it is costly and difficult to maintain. Summary of the invention

[0004] The purpose of the present invention is to provide a static yaw error discrimination method based on wind turbine SCADA data analysis, which can effectively identify the static yaw error, avoid the wind turbine power generation loss and unbalanced load caused by the static yaw error, and solve the high cost problem caused by the laser wind radar calibration of the static yaw error.

[0005] The technical solution adopted by the present invention is a static yaw error determination method based on wind turbine SCADA data analysis, which specifically includes the following steps: Step 1: Obtain wind turbine parameters and SCADA data; Step 2: Identify and clean abnormal data in SCADA data; Step 3: Eliminate and filter the operation data of wind turbines under the influence of control strategies in the SCADA data after identification and cleaning; Step 4: Analysis and determination of static yaw error value.

[0006] The present invention is also characterized in that: In step 1, the parameters of the wind turbine, including the rotor diameter D, Wind rotor swept area A , wind energy utilization coefficient C P , wind turbine theoretical power curve, cut-in wind speed v in And the optimal tip speed ratio l opt ;The theoretical power curve of wind turbine is the wind speed-power curve; SCADA data is the historical operation data of wind turbines, including power P , wind speed v , Air density of the operating environment r , wind wheel speed oh , wind direction angle f , Cabin angle c and pitch angle β .

[0007] Step 2 specifically includes the following steps: Step 2.1: Based on the parameters of the wind turbine and the actual operating conditions, remove the abnormal data in the SCADA data; Step 2.2: The data obtained in step 2.1 are further subjected to the quartering method to eliminate the abnormal values ​​that deviate from the theoretical power curve of the wind turbine; Step 2.3: The data obtained in step 2.2 is further clustered using the DBSCAN algorithm to remove outliers.

[0008] In step 2.1, data that meets any of the following conditions is considered abnormal data in SCADA data: Condition 1: v < v in ; Condition 2: P <0; Condition 3: C P-SCADA > C Pmax ; In the formula, , A=πD 2 / 4 , C Pmax is the wind energy utilization coefficient in the parameters of the wind turbine C P The maximum value of .

[0009] Step 2.2 is as follows: The data obtained in step 2.1 are P bin is the step size of the power subinterval, and the power P Divide into several sub-intervals,P bin Set to: (1) In the formula, r is the power step coefficient, r =0.002, P l is the rated power of the wind turbine; For each power data point in the sub-interval P i ,Will P i Corresponding wind speed data points v i Use quartiles to filter out outliers. Specifically, sort all wind speed data in the sub-interval by numerical value and obtain the upper quartile of the wind speed. v UQ and lower quartile v LQ , then the interquartile range v IQR It is expressed as: (2) The wind speed data points v i Data points that meet any of the following conditions are considered abnormal data and are removed. v i And the corresponding P i : Condition 1:

[0010] Condition 2: .

[0011] Step 2.3 is as follows: The data obtained in step 2.2 are v bin =0.5m / s is the step size of the wind speed sub-interval, and the wind speed v is changed from the cut-in wind speed v in Cut-out wind speed v out Divide into several sub-intervals; take 50kW as the neighborhood radius and 8 as the minimum number of neighborhood samples, and calculate the power data points in the sub-intervals. P i The DBSCAN clustering algorithm was used to remove outliers.

[0012] Step 3 is as follows: Eliminate pitch angle β Data that is not zero; Selecting Tip Speed ​​Ratio lData in the range not affected by wind speed changes, blade tip speed ratio l It is expressed as: (3) In the formula, R = D / 2; According to the selected tip speed ratio data, the data in the most wind speed range is retained accordingly.

[0013] Step 4 specifically includes the following steps: Step 4.1: Convert the data in the optimal wind speed range to the wind speed in steps of 0.4 m / s. v Divide i sub-intervals; Step 4.2: For each wind speed sub-interval i The data within the yaw error i =2° is the step length, and the yaw error interval Divide into j subintervals, where the yaw error i Calculated by the following formula: (4) Step 4.3: Calculate each wind speed interval i Each yaw error subinterval j The average power of m In the wind speed sub-range, n The power mean of the yaw error subintervals , expressed as: (5) In the formula, k For the m In the wind speed sub-interval, n The number of power points in the yaw error subinterval; P a express k The power corresponding to the ath power point among the power points; At the same time, the sum of the power mean values ​​in each optimal wind speed interval is calculated in the yaw sub-interval. n The sum of the power means in the yaw error subintervals is expressed as: (6) Step 4.4: Identify the maximum value of the sum of the means obtained by equation (6), and the corresponding yaw error is the static yaw error of the wind turbine; If the maximum value of the sum of the power means corresponds to the n yaw error subintervals, then the static yaw error value It is expressed as: (7) After eliminating the static yaw error, the power generation P + The theoretical improvement value is expressed as: (8).

[0014] The beneficial effects of the present invention are: The static yaw error discrimination method of the present invention is based on the analysis of SCADA data of wind turbine sets. By performing algorithm statistics, cleaning and screening on the historical SCADA operation data of the wind turbine sets, the static yaw error value of the wind turbine sets is analyzed and calculated. The power generation loss caused by the static yaw error of the wind turbine sets is eliminated in a high-efficiency and low-cost method, the unbalanced load borne by the wind turbine sets is reduced, and the operating efficiency of the wind farm is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of static yaw error; Figure 2 It is a schematic diagram of the overall flow of the static yaw error determination method based on wind turbine SCADA data analysis of the present invention. DETAILED DESCRIPTION

[0016] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] The present invention is based on a static yaw error identification method for wind turbine SCADA data analysis. By acquiring the SCADA data of the wind turbine, and performing statistical analysis and screening on the historical operation data of the wind turbine, abnormal values ​​under abnormal operation conditions such as acquisition errors, fault shutdowns, and power restrictions are eliminated to obtain cleaned operation data. Then, the operation data of the horizontal distribution interval of the blade tip speed ratio in the maximum wind energy tracking stage is selected when the pitch angle of the unit is zero and the power characteristics of different wind speeds and yaw error intervals are analyzed, and the static yaw error value of the wind turbine is calculated and identified. Figure 2 As shown, the specific steps include: Step 1: Obtain wind turbine parameters and SCADA data.

[0018] Wind turbine parameters, including rotor diameter D , Wind rotor swept area A , wind energy utilization coefficient C P , wind turbine theoretical power curve (wind speed-power, vP Curve), cut-in wind speed v in And the optimal tip speed ratio l opt ; SCADA data is the historical operation data of wind turbines, including powerP , wind speed v , Air density of the operating environment r , wind wheel speed oh , wind direction angle f , Cabin angle c and pitch angle β .

[0019] Step 2: Identify and clean abnormal data in SCADA data. The specific steps include: Step 2.1: Based on the parameters of the wind turbine and the actual operating conditions, remove the abnormal data in the SCADA data; Data that meets any of the following conditions is considered abnormal data in SCADA data: Condition 1: v < v in ; Condition 2: P <0; Condition 3: C P-SCADA > C Pmax ; In the formula, , A=πD 2 / 4 , C Pmax is the wind energy utilization coefficient in the parameters of the wind turbine C P The maximum value of .

[0020] Step 2.2: The data obtained in step 2.1 are further subjected to the quartering method to eliminate the abnormal values ​​that deviate from the theoretical power curve of the wind turbine; The data obtained in step 2.1 are P bin is the step size of the power subinterval, and the power P Divide into several sub-intervals, P bin Set to: (1) In the formula, r is the power step coefficient, which is set after data statistics and analysis r =0.002, P l is the rated power of the wind turbine; For each power data point in the sub-interval P i ,Will P i Corresponding wind speed data points vi Use quartiles to filter out outliers. Specifically, sort all wind speed data in the sub-interval by numerical value and obtain the upper quartile of the wind speed. v UQ and lower quartile v LQ , then the interquartile range v IQR It is expressed as: (2) The wind speed data points v i Data points that meet any of the following conditions are considered abnormal data and are removed. v i And the corresponding P i : Condition 1:

[0021] Condition 2: .

[0022] Step 2.3: The data obtained in step 2.2 is further clustered using the DBSCAN (Density-based spatial clustering of applications with noise) algorithm to remove outliers.

[0023] The data obtained in step 2.2 are v bin =0.5m / s is the step size of the wind speed sub-interval, and the wind speed v is changed from the cut-in wind speed v in Cut-out wind speed v out Divide into several sub-intervals; take 50kW as the neighborhood radius and 8 as the minimum number of neighborhood samples, and calculate the power data points in the sub-intervals. P i The DBSCAN clustering algorithm was used to remove outliers.

[0024] Step 3: Eliminate and filter the operation data of wind turbines under the influence of control strategies in the SCADA data after identification and cleaning. Specifically: Eliminate pitch angle β The power characteristics of wind turbines are not only affected by the yaw system, but also by the control system. After reaching the rated power, the wind turbine will perform pitch control. At this time, the power of the wind turbine is not only affected by the yaw error, but also by the reduced wind energy capture capability due to blade pitch control.

[0025] Selecting Tip Speed ​​Ratio lThe data in the range not affected by wind speed changes. Tip speed ratio l It is the ratio of the wind turbine blade tip linear velocity to the wind speed, expressed as: (3) In the formula, R = D / 2; As an important parameter in the wind turbine control strategy, l When the change is large, the output power of the wind turbine P Therefore, the data in the interval where the tip speed ratio is not affected by wind speed changes is selected to minimize the influence of control factors on the static yaw error value judgment result.

[0026] According to the selected blade tip speed ratio data, the data in the most wind speed range is retained. Within this wind speed range, the blade tip speed ratio of the wind turbine l The change is small, the unit is in the maximum power point tracking MPPT control mode, is not affected by the pitch control, other factors have little impact on the power, the power loss caused by the static yaw error is the most obvious, and the static yaw error is easy to identify.

[0027] Step 4: Static yaw error value analysis and judgment. Specifically includes the following steps: Step 4.1: Convert the data in the optimal wind speed range to the wind speed in steps of 0.4 m / s. v Divide i subintervals; the wind speed values ​​in each subinterval are similar. In a smaller wind speed variation range, the tip speed ratio changes less, the control strategy has less impact on the power of the wind turbine, and the static yaw error causes a more obvious loss in power.

[0028] Step 4.2: For each wind speed sub-interval i The data within the yaw error i =2° is the step length, and the yaw error interval Divide into j subintervals, where the yaw error i Calculated by the following formula: (4) Step 4.3: Calculate each wind speed interval i Each yaw error subinterval j The average power of m In the wind speed sub-range, n The power mean of the yaw error subintervals , expressed as: (5) In the formula, k For them In the wind speed sub-interval, n The number of power points in the yaw error subinterval; P a express k The power corresponding to the ath power point among the power points; At the same time, the sum of the power mean values ​​in each optimal wind speed interval is calculated in the yaw sub-interval. n The sum of the power means in the yaw error subintervals is expressed as: (6) Step 4.4: Identify the maximum value of the sum of the means obtained by equation (6), and the corresponding yaw error is the static yaw error of the wind turbine.

[0029] The principle is: at wind speed v If the wind energy utilization coefficient C P unchanged, when the yaw error When the wind speed is zero, the wind turbine output power is the optimal value; on the contrary, v If the wind energy utilization coefficient C P When the output power of the wind turbine is the optimal value, the corresponding yaw error is the static yaw error caused by the measurement error of the wind measuring equipment. Therefore, the maximum value of the sum of the means obtained by identifying formula (6) corresponds to the static yaw error of the wind turbine.

[0030] If the maximum value of the sum of the power means corresponds to the n yaw error subintervals, then the static yaw error value It is expressed as: (7) After eliminating the static yaw error, the power generation P + The theoretical improvement value is expressed as: (8).

[0031] Example 1 This embodiment provides a static yaw error determination method based on wind turbine SCADA data analysis, such as Figure 2 As shown, the specific steps include: Step 1: Obtain wind turbine parameters and SCADA data; Step 2: Identify and clean abnormal data in SCADA data; Step 3: Eliminate and filter the operation data of wind turbines under the influence of control strategies in the SCADA data after identification and cleaning; Step 4: Analysis and determination of static yaw error value.

[0032] Example 2 Based on Example 1, in step 1, the parameters of the wind turbine generator set include the rotor diameter D , Wind rotor swept area A , wind energy utilization coefficient C P , wind turbine theoretical power curve, cut-in wind speed v in And the optimal tip speed ratio l opt ;The theoretical power curve of wind turbine is the wind speed-power curve; SCADA data is the historical operation data of wind turbines, including power P , wind speed v , Air density of the operating environment r , wind wheel speed oh , wind direction angle f , Cabin angle c and pitch angle β .

[0033] Example 3 Based on Example 2, step 2 specifically includes the following steps: Step 2.1: Based on the parameters of the wind turbine and the actual operating conditions, remove the abnormal data in the SCADA data; Step 2.2: The data obtained in step 2.1 are further subjected to the quartering method to eliminate the abnormal values ​​that deviate from the theoretical power curve of the wind turbine; Step 2.3: The data obtained in step 2.2 is further clustered using the DBSCAN algorithm to remove outliers.

[0034] Example 4 Based on Example 3, In step 2.1, data that meets any of the following conditions is considered abnormal data in SCADA data: Condition 1: v < v in ; Condition 2: P <0; Condition 3: C P-SCADA > C Pmax ; In the formula, , A=πD 2 / 4 , C Pmax is the wind energy utilization coefficient in the parameters of the wind turbineC P The maximum value of .

[0035] Step 2.2 is as follows: The data obtained in step 2.1 are P bin is the step size of the power subinterval, and the power P Divide into several sub-intervals, P bin Set to: (1) In the formula, r is the power step coefficient, r =0.002, P l is the rated power of the wind turbine; For each power data point in the sub-interval P i ,Will P i Corresponding wind speed data points v i Use quartiles to filter out outliers. Specifically, sort all wind speed data in the sub-interval by numerical value and obtain the upper quartile of the wind speed. v UQ and lower quartile v LQ , then the interquartile range v IQR It is expressed as: (2) The wind speed data points v i Data points that meet any of the following conditions are considered abnormal data and are removed. v i And the corresponding P i : Condition 1:

[0036] Condition 2: .

[0037] Step 2.3 is as follows: The data obtained in step 2.2 are v bin =0.5m / s is the step size of the wind speed sub-interval, and the wind speed v is changed from the cut-in wind speed v in Cut-out wind speed v outDivide into several sub-intervals; take 50kW as the neighborhood radius and 8 as the minimum number of neighborhood samples, and calculate the power data points in the sub-intervals. P i The DBSCAN clustering algorithm was used to remove outliers.

[0038] Example 5 Based on Example 4, step 3 is specifically as follows: Eliminate pitch angle β Data that is not zero; Selecting Tip Speed ​​Ratio l Data in the range not affected by wind speed changes, blade tip speed ratio l It is expressed as: (3) In the formula, R = D / 2; According to the selected tip speed ratio data, the data in the most wind speed range is retained accordingly.

[0039] Example 6 Based on Example 5, step 4 specifically includes the following steps: Step 4.1: Convert the data in the optimal wind speed range to the wind speed in steps of 0.4 m / s. v Divide i sub-intervals; Step 4.2: For each wind speed sub-interval i The data within the yaw error i =2° is the step length, and the yaw error interval Divide into j subintervals, where the yaw error i Calculated by the following formula: (4) Step 4.3: Calculate each wind speed interval i Each yaw error subinterval j The average power of m In the wind speed sub-range, n The power mean of the yaw error subintervals , expressed as: (5) In the formula, k For the m In the wind speed sub-interval, n The number of power points in the yaw error subinterval; P a express k The power corresponding to the ath power point among the power points; At the same time, the sum of the power mean values ​​in each optimal wind speed interval is calculated in the yaw sub-interval. n The sum of the power means in the yaw error subintervals is expressed as: (6) Step 4.4: Identify the maximum value of the sum of the means obtained by equation (6), and the corresponding yaw error is the static yaw error of the wind turbine; If the maximum value of the sum of the power means corresponds to the n yaw error subintervals, then the static yaw error value It is expressed as: (7) After eliminating the static yaw error, the power generation P + The theoretical improvement value is expressed as: (8).

Claims

1. A static yaw error determination method based on wind turbine SCADA data analysis is characterized in that: The specific steps include: Step 1: Obtain wind turbine parameters and SCADA data; Step 2: Identify and clean abnormal data in SCADA data; Step 3: Eliminate and filter the operation data of wind turbines under the influence of control strategies in the SCADA data after identification and cleaning; Step 4: Analysis and determination of static yaw error value.

2. The static yaw error determination method based on wind turbine SCADA data analysis according to claim 1 is characterized in that: In step 1, the parameters of the wind turbine include the rotor diameter D , Wind rotor swept area A , wind energy utilization coefficient C P , wind turbine theoretical power curve, cut-in wind speed v in And the optimal tip speed ratio λ opt ; The theoretical power curve of the wind turbine is a wind speed-power curve; The SCADA data is the historical operation data of the wind turbine, including power P , wind speed v , Air density of the operating environment ρ , wind wheel speed ω , wind direction angle φ , Cabin angle γ and pitch angle β .

3. The static yaw error determination method based on wind turbine SCADA data analysis according to claim 2 is characterized in that: Step 2 specifically includes the following steps: Step 2.1: Based on the parameters of the wind turbine and the actual operating conditions, remove the abnormal data in the SCADA data; Step 2.2: The data obtained in step 2.1 are further subjected to the quartering method to eliminate the abnormal values ​​that deviate from the theoretical power curve of the wind turbine; Step 2.3: The data obtained in step 2.2 is further clustered using the DBSCAN algorithm to remove outliers.

4. The static yaw error determination method based on wind turbine SCADA data analysis according to claim 3 is characterized in that: In step 2.1, data that meets any of the following conditions is considered abnormal data in SCADA data: Condition 1: v < v in ; Condition 2: P < 0; Condition 3: C P-SCADA > C Pmax ; In the formula, , A=πD 2 / 4 , C Pmax is the wind energy utilization coefficient in the parameters of the wind turbine C P The maximum value of .

5. The static yaw error determination method based on wind turbine SCADA data analysis according to claim 4 is characterized in that: Step 2.2 is as follows: The data obtained in step 2.1 are P bin is the step size of the power subinterval, and the power P Divide into several sub-intervals, P bin Set to: （1） In the formula, r is the power step coefficient, r =0.002, P l is the rated power of the wind turbine; For each power data point in the sub-interval P i ,Will P i Corresponding wind speed data points v i Use quartiles to filter out outliers. Specifically, sort all wind speed data in the sub-interval by numerical value and obtain the upper quartile of the wind speed. v UQ and lower quartile v LQ , then the interquartile range v IQR It is expressed as: （2） The wind speed data points v i Data points that meet any of the following conditions are considered abnormal data and are removed. v i And the corresponding P i : Condition 1: Condition 2: .

6. The static yaw error determination method based on wind turbine SCADA data analysis according to claim 5 is characterized in that: Step 2.3 is as follows: The data obtained in step 2.2 are v bin =0.5m / s is the step size of the wind speed sub-interval, and the wind speed v is changed from the cut-in wind speed v in Cut-out wind speed v out Divide into several sub-intervals; take 50kW as the neighborhood radius and 8 as the minimum number of neighborhood samples, and calculate the power data points in the sub-intervals. P i The DBSCAN clustering algorithm was used to remove outliers.

7. The static yaw error determination method based on wind turbine SCADA data analysis according to claim 6 is characterized in that: Step 3 is as follows: Eliminate pitch angle β Data that is not zero; Selecting Tip Speed ​​Ratio λ The data in the range not affected by wind speed changes, the tip speed ratio λ It is expressed as: （3） In the formula, R = D / 2; According to the selected tip speed ratio data, the data in the maximum wind speed range is retained accordingly.

8. The static yaw error determination method based on wind turbine SCADA data analysis according to claim 7 is characterized in that: Step 4 specifically includes the following steps: Step 4.1: Convert the data in the optimal wind speed range to the wind speed in steps of 0.4 m / s. v Divide i sub-intervals; Step 4.2: For each wind speed sub-interval i The data within the yaw error θ =2° is the step length, and the yaw error interval Divide into j subintervals, where the yaw error θ Calculated by the following formula: （4） Step 4.3: Calculate each wind speed interval i Each yaw error subinterval j The average power of m In the wind speed sub-range, n The power mean of the yaw error subintervals , expressed as: （5） In the formula, k For the m In the wind speed sub-interval, n The number of power points in the yaw error subinterval; P a express k The power corresponding to the ath power point among the power points; At the same time, the sum of the power mean values ​​in each optimal wind speed interval is calculated in the yaw sub-interval. n The sum of the power means in the yaw error subintervals is expressed as: （6） Step 4.4: Identify the maximum value of the sum of the means obtained by equation (6), and the corresponding yaw error is the static yaw error of the wind turbine; If the maximum value of the sum of the power means corresponds to the n yaw error subintervals, then the static yaw error value It is expressed as: （7） After eliminating the static yaw error, the power generation P + The theoretical improvement value is expressed as: （8）。

Images