Doubly-fed fan Crowbar circuit action identification method based on difference hash algorithm

Abstract

The invention relates to the technical field of power system automation, and discloses a doubly-fed fan Crowbar circuit action identification method based on a difference hash algorithm, and the method comprises the steps: screening out a low-voltage ride-through test result of a three-phase symmetric and asymmetric fault through reading a doubly-fed fan hardware-in-loop simulation or field fault ride-through test result; respectively outputting the reactive power and the reactive current of the test result as pictures, fitting the reactive power and the reactive current of the doubly-fed fan Crowbar circuit without action, and outputting the reactive power and the reactive current as pictures; reducing the size of the picture, converting the picture into a grey-scale map, obtaining the fingerprint of the picture by using a difference hash algorithm, and calculating the Hamming distance of reactive power and reactive current, so as to judge whether the Crowbar circuit of the doubly-fed fan acts or not; according to the method, whether the Crowbar circuit acts in the low-voltage ride-through period can be identified from the doubly-fed fan fault ride-through test result, the result is not influenced by the doubly-fed fan capacity, the low-voltage ride-through depth, the fan active output size and the like, and the method has the advantage of high accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of power system automation, and in particular to a method for identifying the Crowbar circuit action of a doubly-fed wind turbine based on a difference hash algorithm. Background Art

[0002] With the construction of new power systems, large-scale new energy is connected to the grid, and the penetration rate of new energy is high in most regions. It is urgent to build a grid simulation model containing new energy. The hardware and control strategies of direct-drive wind turbines and photovoltaic inverters are relatively small; however, the hardware and control strategies used for low voltage ride-through of doubly fed wind turbines are quite different, and the control strategy has a greater impact on whether the Crowbar circuit is used. The low voltage ride-through strategies of doubly fed wind turbines vary greatly from company to company, and the internal control strategies are the core secrets of each company and are not open to the public; whether the Crowbar circuit is activated during the low voltage ride-through of doubly fed wind turbines is one of the core contents of simulation modeling.

[0003] The existing "Guidelines for Power System Safety and Stability" and "Technical Regulations for Wind Farm Access to Power System (Part I: Onshore Wind Power)" require wind turbines to carry out simulation modeling. The existing Crowbar circuit action identification method for double-fed wind turbines during low voltage ride-through mainly relies on manual identification or Crowbar circuit action signals, etc. There are also a few methods based on reactive power and reactive current results of low voltage ride-through for identification, but this method is easily affected by the capacity and output of the double-fed wind turbine. Summary of the invention

[0004] In order to address the deficiencies in the prior art, the present invention provides a method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm, which accurately identifies whether the Crowbar circuit of a doubly-fed wind turbine is in action during low voltage ride-through, lays a foundation for modeling and simulation of doubly-fed wind turbines, and is easy to promote and apply.

[0005] The technical solution adopted by the present invention is:

[0006] The method for identifying the action of the Crowbar circuit of a doubly-fed wind turbine based on a difference hash algorithm comprises the following steps:

[0007] Step 1: Read the fault voltage ride-through test results:

[0008] Read the hardware-in-the-loop simulation or on-site fault voltage ride-through test results of the doubly-fed wind turbine, and filter out the low voltage ride-through test results of three-phase symmetrical and asymmetrical faults;

[0009] Step 2: Select reactive power and current data:

[0010] Data selection: Based on the low voltage ride-through test results of three-phase symmetrical and asymmetrical faults, the reactive power and reactive current data 200ms before the fault and 500ms after the fault are selected and output as pictures;

[0011] Calculate reactive power and current: read the initial values ​​of reactive power and reactive current before the fault, and the voltage amplitude during the low voltage ride-through period, calculate the reactive power and reactive current values ​​during the low voltage ride-through period, assume that the Crowbar circuit of the doubly fed wind turbine does not operate during the low voltage ride-through period, fit the reactive power and reactive current data 200ms before the fault and 500ms after the fault, and output them as pictures;

[0012] Step 3, reduce the image size:

[0013] Using the bilinear interpolation method, the tested reactive power and reactive current images, as well as the fitted reactive power and reactive current images, are reduced to a size of 32×32;

[0014] Step 4, convert to grayscale image:

[0015] Convert the reduced test reactive power and reactive current images, and the fitted reactive power and reactive current images into grayscale images;

[0016] Step 5, get the image fingerprint:

[0017] The differential hash algorithm is used to obtain the fingerprints of the test reactive power and reactive current grayscale images and the fitted reactive power and reactive current grayscale images;

[0018] Step 6, calculate the Hamming distance:

[0019] Calculate the Hamming distance d1 between the test reactive power and the fitted reactive power, and calculate the Hamming distance d2 between the test reactive current and the fitted reactive current;

[0020] Step 7: Determine whether the Crowbar circuit is working:

[0021] Whether the Crowbar circuit of the doubly fed wind turbine is operating is determined by the Hamming distance value d1 between the tested reactive power and the fitted reactive power and the Hamming distance value d2 between the tested reactive current and the fitted reactive current.

[0022] Further, in step 1, the method for screening the low voltage ride through test results of the doubly fed wind turbine three-phase fault includes the following steps:

[0023] Step 1.1, data screening: based on whether the fundamental positive sequence voltage per unit value in the double-fed wind turbine hardware-in-the-loop simulation or on-site fault voltage ride-through test results is less than 0.9. If it is less than 0.9, it is low voltage ride-through test data;

[0024] Step 1.2, phase voltage calculation: Use the full-wave Fourier algorithm to calculate the fundamental positive sequence voltage per unit value during the fault period as U1, and the three-phase phase voltage per unit values ​​during the fault period as U A , U B , U C ;

[0025] Step 1.3, fault diagnosis: If U A -U1|、|U B -U1|、|U C -U1| are all less than 0.01, which indicates that they are the low voltage ride-through data of a three-phase symmetrical fault; otherwise, they are the low voltage ride-through data of an asymmetrical fault.

[0026] Furthermore, in step 2, the process of calculating the reactive power and reactive current values ​​during low voltage ride-through is as follows:

[0027] Reactive current during low voltage ride through for:

[0028]

[0029] The reactive power Q1 during low voltage ride-through is:

[0030]

[0031] Where U t is the line voltage amplitude during low voltage ride-through; is the reactive current value before the fault; I N is the rated current of the equipment; Q0 is the reactive power before the fault; Q1 is the reactive power during the low voltage ride-through period.

[0032] Furthermore, in step 3, the image size is reduced by using the bilinear interpolation method, mainly using the function imresize() in the matlab software to reduce the test reactive power and reactive current images, as well as the fitted reactive power and reactive current images to a size of 32×32.

[0033] Furthermore, in step 5, the specific steps of obtaining the fingerprint of the image using the differential hash algorithm are as follows:

[0034] Step 5.1, using discrete cosine transform to separate the component rate set of the 32×32 image after the test reactive power is reduced, and the set is a 32×32 matrix;

[0035] Step 5.2, retain the upper left corner 8×8 matrix of the 32×32 matrix after discrete cosine transform;

[0036] Step 5.3, calculate the average value of the 8×8 matrix pixels after discrete cosine transformation;

[0037] Step 5.4, after discrete cosine transformation, the pixels of the 8×8 matrix that are greater than the average value are recorded as 1, otherwise they are recorded as 0;

[0038] Step 5.5, obtain the information fingerprint of the test reactive power graph: combine 64 fingerprint information bits and keep the order consistent.

[0039] Furthermore, in step 6, the Hamming distance is used to compare whether each bit of the fingerprint is identical;

[0040] If they are not the same, the Hamming distance value is increased by 1; after the 64 fingerprint information bits are compared, the Hamming distance value is obtained; the smaller the Hamming distance, the higher the similarity of the graphics; the larger the Hamming distance, the higher the similarity of the graphics.

[0041] Further, in step 7, the Hamming distance value d1 is calculated for the 64 fingerprint information obtained by testing reactive power and fitting reactive power; the Hamming distance value d2 is calculated for the 64 fingerprint information obtained by testing reactive current and fitting reactive current; if the Hamming distance values ​​d1 and d2 are both less than or equal to 5, it means that the Crowbar circuit of the doubly fed wind turbine is not actuated; otherwise, it means that the Crowbar circuit of the doubly fed wind turbine is actuated.

[0042] Compared with the prior art, the method for identifying the Crowbar circuit action of a doubly-fed wind turbine based on the difference hash algorithm has the following beneficial effects:

[0043] (1) The method for identifying the Crowbar circuit action of a doubly fed wind turbine based on the differential hash algorithm uses the perceptual hash algorithm combined with the Hamming distance to automatically identify whether the Crowbar circuit of the doubly fed wind turbine is in action according to the reactive power and reactive current of the fault ride-through test results. It is not affected by the capacity of the doubly fed wind turbine, the low voltage ride-through depth, the active output of the wind turbine, etc., and has the advantage of high accuracy.

[0044] (2) It does not rely on manual identification to determine whether the Crowbar circuit is activated during the low voltage ride-through of the doubly-fed wind turbine. At the same time, it can distinguish the operation of the Crowbar circuit during the low voltage ride-through of three-phase symmetrical and asymmetrical faults, which is convenient for the simulation modeling of the doubly-fed wind turbine.

[0045] (3) It is not affected by the size of reactive power and reactive current, but only related to the trend of reactive power and reactive current changes caused by the action of the Crowbar circuit. It has the advantages of high accuracy and ease of use.

[0046] (4) The invention adopts a differential hash algorithm to identify the operation of the Crowbar circuit of a doubly-fed wind turbine, which has high accuracy and fast calculation speed, is easy to use, and is easy to promote and use. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0048] in:

[0049] Figure 1 is a flow chart of the action recognition method of the present invention;

[0050] Figure 2 This is the waveform of the fault ride-through test result of the doubly-fed wind turbine of Application Example 1 of the present invention;

[0051] Figure 3 This is a waveform diagram of reactive power actually measured in Application Example 1 of the present invention;

[0052] Figure 4 This is a waveform diagram of reactive current measured in Application Example 1 of the present invention;

[0053] Figure 5 This is a fitted reactive power waveform diagram of Application Example 1 of the present invention;

[0054] Figure 6 This is a fitted reactive current waveform diagram of Application Example 1 of the present invention;

[0055] Figure 7 This is the waveform of the fault ride-through test result of the doubly-fed wind turbine of Application Example 2 of the present invention;

[0056] Figure 8 This is a waveform diagram of reactive power actually measured in Application Example 1 of the present invention;

[0057] Fig. 9 This is a waveform diagram of reactive current measured in Application Example 1 of the present invention;

[0058] Fig.10 This is a fitted reactive power waveform diagram of Application Example 1 of the present invention;

[0059] Fig.11 This is a fitted reactive current waveform diagram of Application Example 1 of the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] At present, whether the Crowbar circuit of a doubly-fed wind turbine is operating is mainly judged manually based on the reactive power, reactive circuit waveform during low voltage ride-through, or the Crowbar circuit operation signal of the wind turbine; this embodiment provides a Crowbar circuit operation identification method for a doubly-fed wind turbine based on a differential hash algorithm.

[0062] This doubly-fed wind turbine Crowbar circuit action identification method based on the differential hash algorithm does not rely on manual identification, and can automatically and quickly identify whether the Crowbar is in action during three-phase symmetrical and asymmetric low voltage ride-through, which helps to identify the transient model parameters of the doubly-fed wind turbine and lays the foundation for accurate transient simulation modeling of the doubly-fed wind turbine.

[0063] Specifically, Figure 1 As shown, the method for identifying the action of the Crowbar circuit of a doubly fed wind turbine based on the difference hash algorithm is performed in the following steps:

[0064] Step 1: Read the fault voltage ride-through test results:

[0065] Read the hardware-in-the-loop simulation or on-site fault voltage ride-through test results of the doubly-fed wind turbine, and filter out the low voltage ride-through test results of three-phase symmetrical and asymmetrical faults.

[0066] Furthermore, the screening method of the low voltage ride through test results of the three-phase fault of the doubly fed wind turbine specifically includes the following steps:

[0067] Step 1.1, data screening: based on whether the fundamental positive sequence voltage per unit value in the hardware-in-the-loop simulation or on-site fault voltage ride-through test results of the doubly fed wind turbine is less than 0.9. If it is less than 0.9, it is low voltage ride-through test data.

[0068] Step 1.2, phase voltage calculation: Use the full-wave Fourier algorithm to calculate the fundamental positive sequence voltage standard value during the fault period as U1, and the three-phase phase voltage standard values ​​during the fault period as U A , U B , U C .

[0069] Step 1.3, fault diagnosis: If U A -U1|、|U B -U1|、|U C -U1| are all less than 0.01, which indicates that they are the low voltage ride-through data of a three-phase symmetrical fault; otherwise, they are the low voltage ride-through data of an asymmetrical fault.

[0070] Step 2: Select reactive power and current data:

[0071] Data selection: Based on the low voltage ride-through test results of three-phase symmetrical and asymmetrical faults, the reactive power and reactive current data 200ms before the fault and 500ms after the fault were selected and output as pictures.

[0072] Calculate reactive power and current: read the initial values ​​of reactive power and reactive current before the fault and the voltage amplitude during the low voltage ride-through, and calculate the reactive power and reactive current values ​​during the low voltage ride-through; assume that the Crowbar circuit of the doubly fed wind turbine does not operate during the low voltage ride-through, fit the reactive power and reactive current data 200ms before the fault and 500ms after the fault, and output them as a picture.

[0073] Furthermore, the reactive power and reactive current values ​​during low voltage ride-through are calculated. The specific process is as follows:

[0074] Reactive current during low voltage ride through for:

[0075]

[0076] The reactive power Q1 during low voltage ride-through is:

[0077]

[0078] Where U t is the line voltage amplitude during low voltage ride-through; is the reactive current value before the fault; I N is the rated current of the equipment; Q0 is the reactive power before the fault; Q1 is the reactive power during the low voltage ride-through period.

[0079] Step 3, reduce the image size:

[0080] The tested reactive power and reactive current images, as well as the fitted reactive power and reactive current images, are reduced to a size of 32×32 using a bilinear interpolation method.

[0081] Specifically, the double-line interpolation method is used to reduce the image size, and the function imresize() in the matlab software is mainly used to reduce the test reactive power and reactive current images, as well as the fitted reactive power and reactive current images to a size of 32×32.

[0082] Step 4, convert to grayscale image:

[0083] The reduced test reactive power and reactive current images, as well as the fitted reactive power and reactive current images are converted into grayscale images for subsequent processing.

[0084] Step 5, get the image fingerprint:

[0085] The differential hash algorithm is used to obtain the fingerprints of the test reactive power and reactive current grayscale images and the fitted reactive power and reactive current grayscale images.

[0086] The specific steps to obtain the fingerprint of an image using the differential hash algorithm are as follows:

[0087] Take the reactive power test picture as an example to introduce:

[0088] Step 5.1, using discrete cosine transform to separate the component rate set of the 32×32 image after the test reactive power is reduced, and the set is a 32×32 matrix;

[0089] Step 5.2, retain the upper left corner 8×8 matrix of the 32×32 matrix after discrete cosine transform;

[0090] Step 5.3, calculate the average value of the 8×8 matrix pixels after discrete cosine transformation;

[0091] Step 5.4, after discrete cosine transformation, the pixels of the 8×8 matrix that are greater than the average value are recorded as 1, otherwise they are recorded as 0;

[0092] Step 5.5, obtain the information fingerprint of the test reactive power graph: combine 64 fingerprint information bits and keep the order consistent.

[0093] The steps of testing reactive current images, fitting reactive power and reactive current fingerprint extraction can be obtained in the same way.

[0094] Step 6, calculate the Hamming distance:

[0095] The Hamming distance d1 between the test reactive power and the fitted reactive power is calculated, and the Hamming distance d2 between the test reactive current and the fitted reactive current is calculated.

[0096] Specifically, the Hamming distance is to compare whether each bit of the fingerprint is the same; if not, the Hamming distance value is increased by 1; after the 64 fingerprint information bits are compared, the Hamming distance value can be obtained. The smaller the Hamming distance, the higher the similarity of the image quantity; the larger the Hamming distance, the higher the similarity of the image quantity.

[0097] Step 7: Determine whether the Crowbar circuit is working:

[0098] Whether the Crowbar circuit of the doubly fed wind turbine is operating is determined by the Hamming distance value d1 between the tested reactive power and the fitted reactive power and the Hamming distance value d2 between the tested reactive current and the fitted reactive current.

[0099] The Hamming distance value d1 is calculated for the 64 fingerprint information obtained by testing reactive power and fitting reactive power; the Hamming distance value d2 is calculated for the 64 fingerprint information obtained by testing reactive current and fitting reactive current; if the Hamming distance values ​​d1 and d2 are both less than or equal to 5, it means that the Crowbar circuit of the doubly fed wind turbine is not activated; otherwise, it means that the Crowbar circuit of the doubly fed wind turbine is activated.

[0100] Through the above steps, a complete identification method is constructed, which can effectively analyze the state of the Crowbar circuit of the doubly fed wind turbine under low voltage ride-through conditions.

[0101] Furthermore, in order to verify the accuracy of the Crowbar circuit action identification method of the doubly fed wind turbine based on the difference hash algorithm, the following actual verification was carried out:

[0102] Actual verification 1

[0103] Read the on-site measured fault ride-through test data of A double-fed wind turbine, such as Figure 2 As shown, the screening is for three-phase symmetrical low voltage ride-through test:

[0104] Select the reactive power and reactive current data 200ms before and 500ms after the fault, and output them as pictures, such as Figure 3 , 4 As shown;

[0105] Read the initial values ​​of reactive power and reactive current before the fault, the voltage amplitude during the low voltage ride-through period, calculate the reactive power and reactive current values ​​during the low voltage ride-through period, fit the reactive power and reactive current data 200ms before the fault and 500ms after the fault, and output them as pictures, such as Figure 5 , 6 As shown;

[0106] Use bilinear interpolation to reduce the image size and convert it into grayscale;

[0107] The differential hash algorithm is used to obtain the fingerprints of the test reactive power and reactive current grayscale images and the fitted reactive power and reactive current grayscale images;

[0108] Calculate the Hamming distance d1=6 between the test reactive power and the fitted reactive power, and calculate the Hamming distance d2=6 between the test reactive current and the fitted reactive current;

[0109] Since the Hamming distance values ​​d1 and d2 are both greater than 5, the Crowbar circuit of the doubly fed wind turbine operates during the three-phase symmetrical voltage ride-through period.

[0110] Actual verification 2

[0111] Read the on-site measured fault ride-through test data of B double-fed wind turbine, such as Figure 7 As shown, the screening is a three-phase symmetrical low voltage ride-through test;

[0112] Select the reactive power and reactive current data 200ms before and 500ms after the fault, and output them as pictures, such as Figure 8 , 9 As shown;

[0113] Read the initial values ​​of reactive power and reactive current before the fault, the voltage amplitude during the low voltage ride-through period, calculate the reactive power and reactive current values ​​during the low voltage ride-through period, fit the reactive power and reactive current data 200ms before the fault and 500ms after the fault, and output them as pictures, such as Fig.10 , 11 As shown;

[0114] Use bilinear interpolation to reduce the image size and convert it into grayscale;

[0115] The differential hash algorithm is used to obtain the fingerprints of the test reactive power and reactive current grayscale images and the fitted reactive power and reactive current grayscale images;

[0116] Calculate the Hamming distance d1=1 between the test reactive power and the fitted reactive power, and calculate the Hamming distance d2=1 between the test reactive current and the fitted reactive current;

[0117] Since the Hamming distance values ​​d1 and d2 are both less than 5, the Crowbar circuit of the doubly fed wind turbine does not operate during the three-phase symmetrical voltage ride-through period.

[0118] In summary, the doubly fed wind turbine Crowbar circuit action identification method based on the differential hash algorithm does not rely on manual identification, and can automatically and quickly identify whether the Crowbar is acting during three-phase symmetrical and asymmetric low voltage ride-through. The differential hash algorithm is combined with the Hamming distance to identify the doubly fed wind turbine Crowbar circuit action, which is not affected by the doubly fed wind turbine capacity, low voltage ride-through depth, wind turbine active output, etc., and has the advantage of high accuracy.

[0119] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the same elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm, characterized in that: The following steps are involved: Step 1: Read the fault voltage ride-through test results: Read the hardware-in-the-loop simulation or on-site fault voltage ride-through test results of the doubly-fed wind turbine, and filter out the low voltage ride-through test results of three-phase symmetrical and asymmetrical faults; Step 2: Select reactive power and current data: Data selection: Based on the low voltage ride-through test results of three-phase symmetrical and asymmetrical faults, the reactive power and reactive current data 200ms before the fault and 500ms after the fault are selected and output as pictures; Calculate reactive power and current: read the initial values ​​of reactive power and reactive current before the fault, and the voltage amplitude during the low voltage ride-through period, calculate the reactive power and reactive current values ​​during the low voltage ride-through period, assume that the Crowbar circuit of the doubly fed wind turbine does not operate during the low voltage ride-through period, fit the reactive power and reactive current data 200ms before the fault and 500ms after the fault, and output them as pictures; Step 3, reduce the image size: Using the bilinear interpolation method, the tested reactive power and reactive current images, as well as the fitted reactive power and reactive current images, are reduced to a size of 32×32; Step 4, convert to grayscale image: Convert the reduced test reactive power and reactive current images, and the fitted reactive power and reactive current images into grayscale images; Step 5, get the image fingerprint: The differential hash algorithm is used to obtain the fingerprints of the test reactive power and reactive current grayscale images and the fitted reactive power and reactive current grayscale images; Step 6, calculate the Hamming distance: Calculate the Hamming distance d1 between the test reactive power and the fitted reactive power, and calculate the Hamming distance d2 between the test reactive current and the fitted reactive current; Step 7: Determine whether the Crowbar circuit is working: Whether the Crowbar circuit of the doubly fed wind turbine is operating is determined by the Hamming distance value d1 between the tested reactive power and the fitted reactive power and the Hamming distance value d2 between the tested reactive current and the fitted reactive current.

2. The method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm according to claim 1 is characterized in that: In step 1, the method for screening the low voltage ride through test results of the doubly fed wind turbine three-phase fault includes the following steps: Step 1.1, data screening: based on whether the fundamental positive sequence voltage per unit value in the double-fed wind turbine hardware-in-the-loop simulation or on-site fault voltage ride-through test results is less than 0.

9. If it is less than 0.9, it is low voltage ride-through test data; Step 1.2, phase voltage calculation: Use the full-wave Fourier algorithm to calculate the fundamental positive sequence voltage standard value during the fault period as U1, and the three-phase phase voltage standard values ​​during the fault period as U A , U B , U C ; Step 1.3, fault diagnosis: If U A -U1|、|U B -U1|、|U C -U1| are all less than 0.01, which indicates that they are the low voltage ride-through data of a three-phase symmetrical fault; otherwise, they are the low voltage ride-through data of an asymmetrical fault.

3. The method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm according to claim 1, characterized in that: In step 2, the process of calculating the reactive power and reactive current values ​​during low voltage ride-through is as follows: Reactive current during low voltage ride through for: The reactive power Q1 during low voltage ride-through is: Where U t is the line voltage amplitude during low voltage ride-through; is the reactive current value before the fault; I N is the rated current of the equipment; Q0 is the reactive power before the fault; Q1 is the reactive power during the low voltage ride-through period.

4. The method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm according to claim 1 is characterized in that: In step 3, the image size is reduced by using the bilinear interpolation method, mainly using the function imresize() in the MATLAB software to reduce the test reactive power and reactive current images, as well as the fitted reactive power and reactive current images to a size of 32×32.

5. The method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm according to claim 1, characterized in that: In step 5, the specific steps of using the differential hash algorithm to obtain the fingerprint of the image are as follows: Step 5.1, using discrete cosine transform to separate the component rate set of the 32×32 image after the test reactive power is reduced, and the set is a 32×32 matrix; Step 5.2, retain the upper left corner 8×8 matrix of the 32×32 matrix after discrete cosine transform; Step 5.3, calculate the average value of the 8×8 matrix pixels after discrete cosine transformation; Step 5.4, after discrete cosine transformation, the pixels of the 8×8 matrix that are greater than the average value are recorded as 1, otherwise they are recorded as 0; Step 5.5, obtain the information fingerprint of the test reactive power graph: combine 64 fingerprint information bits and keep the order consistent.

6. The method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm according to claim 1, characterized in that: In step 6, the Hamming distance is used to compare whether each bit of the fingerprint is identical; If they are not the same, the Hamming distance value is increased by 1; after the 64 fingerprint information bits are compared, the Hamming distance value is obtained; the smaller the Hamming distance, the higher the similarity of the graphics; the larger the Hamming distance, the higher the similarity of the graphics.

7. The method for identifying the action of a doubly-fed wind turbine Crowbar circuit based on a differential hash algorithm according to claim 1, characterized in that: In step 7, the Hamming distance value d1 is calculated for the 64 fingerprint information obtained by testing reactive power and fitting reactive power; the Hamming distance value d2 is calculated for the 64 fingerprint information obtained by testing reactive current and fitting reactive current; if the Hamming distance values ​​d1 and d2 are both less than or equal to 5, it means that the Crowbar circuit of the doubly fed wind turbine is not actuated; otherwise, it means that the Crowbar circuit of the doubly fed wind turbine is actuated.

Images