A new energy large base alternating current collection line multi-interval quantity fault detection method

Abstract

The application discloses a new energy large base AC collection line multi-interval quantity fault detection method, and belongs to the technical field of line fault detection. The new energy large base AC collection line multi-interval quantity fault detection method comprises the following steps: analyzing the sudden current characteristics of power electronic sources in the new energy large base after a fault, and determining the differences in the sudden current characteristics of different types of power electronic sources; using the differences in the sudden current characteristics, extracting the sudden current characteristics by using an eigenvalue decomposition algorithm; analyzing the differences in the sudden current characteristics of the fault collection line, the non-fault collection line and the low-voltage side line of the main transformer, establishing a fault detection setting calculation mode, and realizing accurate identification of faults in the region. The new energy large base AC collection line multi-interval quantity fault detection method solves the problems that the existing detection method is not suitable for 35kV AC collection line detection in the new energy large base, and the fault detection efficiency and precision are low.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of line fault detection, and in particular to a multi-interval quantity fault detection method for an AC collection line of a large new energy base. BACKGROUND

[0002] The construction of large new energy bases such as Shaguo Desert has become a new form and important measure for the construction of new power systems. Flexible direct current transmission is an important way to transmit power from large new energy bases. When a 35kV AC collection line of a large new energy base fails, the short-circuit current is provided by power electronic conversion equipment of different types, which presents the characteristics of limited amplitude and waveform distortion, making it difficult for overcurrent fault detection to act correctly.

[0003] To solve the problem of correct action of traditional overcurrent fault detection affected by the limited amplitude and waveform distortion of the fault characteristics of multiple types of power electronic sources, existing solutions can be divided into two categories: time-domain quantity fault detection and frequency-domain quantity fault detection. In the frequency-domain quantity-based fault detection, the non-power frequency quantity of the fault characteristics of new energy sources is considered, and a fault detection based on the frequency difference of multi-interval current is proposed. However, in a system without synchronous power sources, the fault detection performance decreases because different power electronic converters have non-power frequency quantities. A new principle of fault detection based on high-frequency impedance can also be used, but it requires a high sampling rate and is not resistant to noise. Considering that the sampling frequency of the fault detection device of the AC system is usually within 2.4kHz, the fault detection based on fault transient time-domain quantities is more easily applied in engineering. By analyzing the current fault component, a new principle of fault detection is proposed using the characteristics of the current fault component. However, this method is difficult to resist the influence of noise and outliers. In existing methods, the fault detection setting value is corrected by a step setting value or the real-time calculation of the system impedance to adaptively correct the fault detection setting value. However, this method is only suitable for scenarios where new energy is connected to a large system and is difficult to apply to 35kV AC collection lines of large new energy bases without synchronous power sources. Artificial intelligence algorithms can also be used to identify fault lines, but the operation time of artificial intelligence algorithms is relatively long, and the fault detection performance is highly dependent on data quality, so it is difficult to apply to actual engineering. SUMMARY

[0004] The purpose of the present application is to provide a multi-interval quantity fault detection method for an AC collection line of a large new energy base, which solves the problem of low fault detection efficiency and low precision of existing detection methods that are not suitable for 35kV AC collection lines of large new energy bases.

[0005] To achieve the above purpose, the present application provides a multi-interval quantity fault detection method for an AC collection line of a large new energy base, comprising the following steps:

[0006] S1, analyze the characteristics of the sudden current of the power electronic source in the new energy base after the fault, and determine the differences in the characteristics of the sudden current of different types of power electronic sources;

[0007] S2, use the differences in the characteristics of the sudden current to extract the characteristics of the sudden current on the multi-interval line by using the algorithm based on eigenvalue decomposition;

[0008] S3, analyze the measurement error and the differences in the characteristics of the sudden current of the fault collection line, the non-fault collection line, and the low-voltage side line of the main transformer, establish a fault detection setting calculation method, and realize accurate identification of the faults inside and outside the area.

[0009] Preferably, in S1, the power electronic source includes wind and light units and a flexible direct current converter station, the wind and light units adopt current inner loop control after the fault occurs, and present a controlled current source characteristic, and the short-circuit current sudden change characteristic is represented as:

[0010] ;

[0011] In the formula, the subscript FPRES represents a full power type power source, is the short-circuit current of the full power type power source, and respectively represent the d-axis current instruction value and the q-axis current instruction value of the full power type power source after the fault, d q is the d-axis current actual value and the q-axis current actual value of the full power type power source before the fault, d is the three-phase phase sequence, is the damping ratio of the second-order system, ω d is the damping oscillation frequency, β is the damping angle, is the initial phase angle after the fault, ω is the power frequency angle frequency, t is the time, A is the calculation coefficient related to the damping oscillation frequency, is the calculation coefficient related to the damping ratio of the second-order system;

[0012] The flexible direct current converter station adopts constant voltage-frequency control, and the sudden current is represented as:

[0013] ;

[0014] In the formula, ω + represents the positive sequence angle frequency, ω - represents the negative sequence angle frequency, Z eq is the equivalent impedance related to the short-circuit type, ​​​a proportional coefficient related to the short-circuit type, a current in the φ phase of the HVDC converter station, an electromotive force of the HVDC converter station, a positive-sequence initial phase angle, a negative-sequence initial phase angle, a short-circuit current provided by the wind-solar generator, the wind-solar generator.

[0015] Preferably, the maximum current-carrying capacity of the wind-solar generator is 1.5 times the rated current thereof.

[0016] Preferably, the maximum short-circuit current provided by the HVDC converter station for overcurrent lockout is 1.5 times the rated current thereof.

[0017] Preferably, in the S2, the sudden change current feature on the multi-interval line is extracted by using an algorithm based on eigenvalue decomposition, and specifically includes the following steps:

[0018] S21, constructing a current sample array in a time window into a form of an inverse Toeplitz matrix:

[0019] ;

[0020] wherein N is the number of current samples of the current sample array I , and is the i-th current sample value. N

[0021] S22, performing Huber weighting processing on each sample value in the matrix to reduce the weight of abnormal values and reduce the influence of abnormal values.

[0022] S23, performing standardization processing on the matrix after the weighting processing to obtain a standardized matrix; and extracting a maximum eigenvalue by eigenvalue decomposition to represent the short-circuit current sudden change feature.

[0023] Preferably, in the S22,

[0024] the Huber weighting factor is:

[0025] ;

[0026] wherein is the Huber weighting factor of the matrix with N rows and M columns, is the current sample value of the matrix with N rows and M columns, is the median. Me

[0027] ​​​Preferably, in the S23,

[0028] The normalized matrix is:

[0029] ;

[0030] The normalized matrix is subjected to eigenvalue decomposition:

[0031] ;

[0032] In the formula, B is the eigenvalue matrix after eigenvalue decomposition, and the superscript T represents the transpose of the matrix, is the eigenvalue of the matrix B , λ 1> λ 2>…> λ N .

[0033] Preferably, in the S3, the fixed value setting calculation mode is based on the size relationship of the short-circuit current mutation characteristics of the fault collection line, the non-fault collection line and the low-voltage side line of the main transformer after the fault occurs, combined with the influence of the new energy zero output scene, to form a fault detection criterion, and the fault detection criterion is:

[0034] ;

[0035] In the formula, λ 1( I L ) is the maximum eigenvalue of the current sampling value on the collection line obtained by eigenvalue decomposition, is the fault detection setting value of the first fault detection criterion, is the fault detection setting value of the second fault detection criterion, I MTL is the current sampling value on the low-voltage side line of the main transformer, is the maximum eigenvalue obtained by eigenvalue decomposition after setting processing of the current sampling value on the low-voltage side line of the main transformer;

[0036] When the fault detection criterion is met, it is determined as an intra-zone fault; otherwise, it is an extra-zone fault.

[0037] Preferably, the S3 is λ set is 10, is 0.7.

[0038] The new energy large base alternating current collection line multi-interval fault detection method has the following advantages and positive effects:

[0039] 1. The application utilizes the difference of the sudden current characteristics of multiple types of power electronic sources, extracts fault characteristics by combining eigenvalue decomposition algorithm, accurately distinguishes internal fault and external fault through detection criterion, avoids misoperation and refusal, and improves detection accuracy.

[0040] 2. The application amplifies the sudden characteristics by constructing inverse Toeplitz matrix, reduces the influence of abnormal values by combining Huber weighting, reduces misoperation caused by noise and abnormal values, standardizes processing and focuses on main sudden characteristics through eigenvalue decomposition, further filters out interference components, and improves anti-interference ability.

[0041] 3. The application is aimed at new energy zero output scene, and the setting value λ set ensures that external fault detection does not misoperate; considering measurement error and phase angle difference, the setting value k set ensures that internal fault detection does not refuse, adapts to complex environment of no synchronous power supply and multiple power electronic source access, and meets the detection needs of new energy large base 35kV AC collection line.

[0042] 4. The application realizes rapid fault identification in the transient stage through 10ms window length current sampling and matrix eigenvalue fast calculation, shortens fault response time, and improves fault detection efficiency.

[0043] The technical solutions of the application will be further described in detail below with the help of the drawings and examples. DESCRIPTION OF DRAWINGS

[0044] Figure 1 is the flow chart of the application;

[0045] Figure 2 is the new energy flexible direct current grid connection topology graph of the simulation experiment of the application;

[0046] Figure 3 is the internal fault detection calculation value of the simulation experiment;

[0047] Figure 4 is the external fault detection calculation value of the simulation experiment. DETAILED DESCRIPTION

[0048] In the present application, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. If there is any inconsistency, the meaning explained in the specification or the meaning derived from the content described in the specification shall prevail. In addition, the terms used herein are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

[0049] The embodiments of the application will be described in detail below with reference to the drawings.

[0050] As Figure 1As shown. A new energy large base AC collection line multi-interval fault detection method, comprising the following steps:

[0051] S1, analyze the sudden current characteristics of power electronic sources in new energy large base after fault, and determine the differences in sudden current characteristics of different types of power electronic sources.

[0052] Power electronic sources include wind and light units and flexible DC converter stations. The wind and light units adopt current inner loop control after the fault occurs, showing controlled current source characteristics, and the short-circuit current sudden change characteristics are represented as:

[0053] ;

[0054] In the formula, the subscript FPRES represents the full power type power source, is the short-circuit current of the full power type power source, and respectively, are the full power type power source before fault d axis, q axis current command value, is the actual value of the full power type power source before fault d axis current, is the three-phase phase sequence, is the damping ratio of the second-order system, ω d is the damping oscillation frequency, β is the damping angle, is the initial phase angle after fault, ω is the power frequency angle frequency, t is the time, A is the calculation coefficient related to the damping oscillation frequency, is the calculation coefficient related to the damping ratio of the second-order system.

[0055] The short-circuit current sudden change characteristics provided by the wind and light units are affected by the low-pass control strategy, and the sudden change current amplitude is related to the maximum current capacity of the wind and light units. Considering the characteristics of the power electronic elements of the wind and light units, the maximum current capacity is 1.5 times the rated current.

[0056] The flexible DC converter station adopts constant voltage-frequency control, which needs to reduce the voltage command value to prevent the short-circuit current from being too large, and the sudden change current is represented as:

[0057] ;

[0058] In the formula, ω + represents the positive sequence angle frequency, ω - represents the negative sequence angle frequency, Z eq is the equivalent impedance related to the short-circuit type, is a proportional coefficient related to the short-circuit type, is a current of the HVDC converter station in the φ phase, is an electromotive force of the HVDC converter station, is a positive-sequence initial phase angle, is a negative-sequence initial phase angle, is a short-circuit current provided by the wind-solar generator, represents the wind-solar generator.

[0059] The amplitude of the sudden change current provided by the HVDC converter station is affected by the maximum current of the HVDC converter station after the fault, in order to prevent overcurrent from causing the HVDC converter station to be blocked, the maximum short-circuit current provided by the HVDC converter station is 1.5 times the rated current of the HVDC converter station, and compared with the sudden change current of the new energy, due to the difference in control characteristics and the difference in the maximum current capacity characteristics, it presents different sudden change current characteristics.

[0060] S2, by using the difference in the sudden change current characteristics, an algorithm based on eigenvalue decomposition is used to extract the sudden change current characteristics on the multi-interval line.

[0061] The algorithm based on eigenvalue decomposition is used to extract the sudden change current characteristics on the multi-interval line, which specifically includes the following steps:

[0062] S21, in order to quickly identify the sudden change characteristics of the short-circuit current in the transient stage, the current sampling array in a time window is constructed in the form of a two-dimensional array, in order to amplify the difference in sudden change characteristics, the current sampling array is constructed in the form of an inverse Toeplitz matrix as follows:

[0063] ;

[0064] In the formula, N is the number of current samples of the current sampling array I, is the first current sampling value. N

[0065] After the fault occurs, with the update of the sampling points in the time window, the fault sampling points are updated along the left diagonal line sampling points of the constructed inverse Toeplitz matrix, with the left diagonal line sampling points as the center, and the change speed of the fault sampling points of the matrix is higher than that of the one-dimensional array.

[0066] S22, Huber weighting is performed on each sampling value in the matrix to reduce the weight of the abnormal value and reduce the influence of the abnormal value.

[0067] By Huber weighting of each sampling value, the weight of the abnormal value is reduced to reduce the influence of the abnormal value, and the Huber weighting factor is:

[0068] ; ​

[0069] wherein, is a matrix is a matrix is a Huber weighting factor, is a matrix is a matrix is a matrix Me is a median.

[0070] S23, standardizing the matrix after the weighting processing to obtain a standardized matrix; extracting a maximum eigenvalue through eigenvalue decomposition to represent the short-circuit current mutation feature.

[0071] The inverse Toeplitz matrix is standardized, and a mean operator and a Huber weighting factor are introduced to obtain an improved standardized matrix:

[0072] ;

[0073] Eigenvalues obtained by eigenvalue decomposition of the matrix reflect the mutation characteristics of the current signal under different frequency modes. The standardized matrix is subjected to eigenvalue decomposition:

[0074] ;

[0075] wherein, B is an eigenvalue matrix after eigenvalue decomposition, and the superscript T represents the transpose of the matrix, is a matrix B is an eigenvalue of the matrix λ 1> λ 2>…> λ N .

[0076] The larger eigenvalue usually corresponds to the main mutation characteristics of the current signal, and the smaller eigenvalue may be related to the noise component, so the maximum eigenvalue can be used to represent the short-circuit current mutation feature of the corresponding line. The eigenvalue calculated by eigenvalue decomposition of the improved standardized matrix can effectively reduce the influence of abnormal values and avoid fault detection misoperation caused by abnormal values.

[0077] S3, analyzing the measurement error and the difference in the mutation current characteristics of the fault collection line, the non-fault collection line and the low-voltage side line of the main transformer, establishing a fault detection setting calculation method, and realizing accurate identification of intra- and inter-zone faults.

[0078] According to the size relationship of the short-circuit current mutation characteristics of the three lines after the fault occurs, and considering the influence of the proposed fault detection under the scene of new energy zero output, a multi-interval information quantity fault detection criterion based on eigenvalue decomposition is constructed as follows:

[0079] ;

[0080] In the formula, λ 1( I L ) is the maximum eigenvalue calculated by eigenvalue decomposition of the current sampling value on the aggregation line, λ set is the fault detection setting value of the first fault detection criterion, is the fault detection setting value of the second fault detection criterion, I MTL is the current sampling value on the low-voltage side of the main transformer, is the maximum eigenvalue obtained by eigenvalue decomposition of the current sampling value on the low-voltage side of the main transformer after setting processing.

[0081] The setting value calculation of the fault detection criterion needs to ensure that the proposed fault detection does not malfunction under zone-out fault and does not refuse to move under zone-in fault. Among them, the first fault detection criterion mainly ensures that the fault detection does not malfunction under zone-out fault in the non-fault line new energy zero output scene, considering the maximum 10% amplitude error caused by the current transformer, for the setting value , the short-circuit current amplitude I F provided by the non-fault line needs to be considered, which is 1.2 times the rated current provided by the new energy on the line, and the influence of the current transformer amplitude error is considered, and the maximum eigenvalue calculated is:

[0082] ;

[0083] In the formula, ω represents the angular frequency, t represents time, φ represents the initial phase angle. Therefore, a certain margin is reserved, and the setting value of the first fault detection criterion λ set is set to 10.

[0084] The setting of the setting value of the second fault detection criterion needs to consider that the fault detection does not refuse to move under zone-in fault, considering that the short-circuit current of wind and light units and flexible DC stations reaches 130° maximum phase angle difference, since the current sudden change variable calculation does not change the phase angle characteristics, therefore the maximum phase angle difference of the current sudden change variable can still be calculated as 130°, so the sudden change current characteristics of different lines can be obtained as:

[0085] ;

[0086] In the formula, E represents the current transient sudden change characteristic, and the subscript NFL represents the non-fault line, FL represents the fault line, MTL represents the low-voltage side of the main transformer, and RES represents the wind and light power supply.

[0087] Therefore, when the medium-voltage collection line fails, the sudden current amplitude on the fault line side is close to 0.766 times the sudden current amplitude on the low-voltage side of the main transformer, and therefore the setting value cannot be greater than 0.766. To prevent the influence of measurement errors, synchronization delays and other factors, a certain margin is reserved, and the final setting value is set to 0.7.

[0088] Therefore, the fault detection process is as follows: after the fault detection is started, the three-phase current sampling signals of different lines with a window length of 10 ms are extracted, and after the inverse Toeplitz matrix preprocessing, the standardization processing and the maximum eigenvalue are calculated; when the maximum eigenvalue meets the fault detection criterion, the fault detection is performed at this time, and it is determined as an internal fault; when the maximum eigenvalue does not meet the fault detection criterion, the fault detection is reset at this time, and it is determined as an external fault.

[0089] The method described in the application will be described below with specific simulation examples.

[0090] Figure 2 The new energy flexible direct current grid topology for the simulation experiment of the application is shown in FIG. 1. Figure 2 As shown in FIG. 2, the new energy large base flexible direct current system model is built in the RTDS, the fault positions are selected as five fault points F1-F5, F1, F2 and F3 are respectively at the outlet of the collection line 1, the midpoint of the collection line and the end of the collection line, F4 is at the collection line 2, and F5 is at the low-voltage side of the main transformer. Among them, F4 and F5 are external faults for the collection line 1, the current signal sampling frequency is 1 kHz, and the time window is selected as 20 ms.

[0091] The simulation experiment results are shown in FIG. 3. Figure 3 Figure 4 As shown in FIG. 3, when the external three-phase fault occurs, the sudden current of the collection line is small, and therefore the eigenvalue difference is far less than 0, and the fault detection is reliable and does not act; when the internal three-phase fault occurs, the sudden current of the collection line increases, and therefore the eigenvalue difference is far greater than 0, and the proposed fault detection can correctly act.

[0092] Therefore, by using the new energy large base alternating current collection line multi-interval fault detection method described in the application, the problem that the existing detection method is not suitable for the 35kV alternating current collection line detection of the new energy large base, and the fault detection efficiency and precision are low, is solved.

[0093] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the application rather than limit them, and although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the application can still be modified or replaced by equivalents, and these modifications or replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the application.​

Claims

1. A method for detecting multi-interval faults in AC collection lines of a large-scale new energy base, characterized in that, Includes the following steps: S1. Analyze the transient current characteristics of power electronic sources in the new energy base after a fault, and determine the differences in transient current characteristics of different types of power electronic sources. S2. Utilizing the differences in transient current characteristics, an algorithm based on eigenvalue decomposition is used to extract transient current features on multi-segment lines. S3. Analyze the measurement errors and the differences in sudden current characteristics of fault collection lines, non-fault collection lines, and low-voltage side lines of the main transformer, establish a fault detection setting calculation method, and realize accurate identification of faults inside and outside the area. In S3, the setting calculation method is based on the relationship between the magnitude of the short-circuit current changes in the fault-collecting lines, non-fault-collecting lines, and the low-voltage side lines of the main transformer after a fault occurs, combined with the influence of the zero-output scenario of new energy sources, to form a fault detection criterion. The fault detection criterion is as follows: ； In the formula, The maximum eigenvalue is obtained by eigenvalue decomposition of the sampled current values ​​on the line. This is the fault detection setting value for the first fault detection criterion. This is the fault detection setting value for the second fault detection criterion. The current sampling value on the low-voltage side line of the main transformer. The maximum eigenvalue obtained by eigenvalue decomposition after setting the current sampling value on the low-voltage side of the main transformer; If the fault detection criteria are met, the fault is determined to be within the zone; otherwise, it is determined to be outside the zone.

2. The method for detecting multi-interval faults in AC aggregation lines of a large-scale new energy base according to claim 1, characterized in that, In S1, the power electronic source includes wind and solar turbines and flexible DC converter stations. After a fault occurs, the wind and solar turbines adopt current inner loop control, exhibiting controlled current source characteristics. The short-circuit current mutation characteristics are expressed as follows: ； In the formula, the subscript FPRES represents a full-power power supply. This is the short-circuit current of a full-power power supply. and After a full-power power supply failure axis, Shaft current command value, This represents the actual d-axis current before a full-power power supply failure. For the three-phase sequence, For the damping ratio of the second-order system, The frequency of the damped oscillation. The damping angle is... The initial phase angle after the fault. It is the power frequency angular frequency. For time, A The calculated coefficients are related to the damped oscillation frequency. These are the calculation coefficients related to the damping ratio of the second-order system; The flexible DC converter station adopts constant voltage-frequency control, and the sudden current is expressed as: ； In the formula, Represents the positive sequence angular frequency. Represents the negative sequence angular frequency. The equivalent impedance is related to the short-circuit type. A proportionality coefficient related to the type of short circuit. For flexible straight converter station Phase current, The electromotive force of the flexible DC converter station. The first phase angle is in positive sequence. The initial phase angle is negative. Short-circuit current provided for wind and solar turbine units. Represents wind and solar power units.

3. The method for detecting multi-interval faults in AC collection lines of a large-scale new energy base according to claim 2, characterized in that: The maximum current-bearing capacity of the wind and solar turbine is 1.5 times its rated current.

4. The method for detecting multi-interval faults in AC collection lines of a large-scale new energy base according to claim 3, characterized in that: To prevent overcurrent blocking, the maximum short-circuit current provided by the flexible DC converter station is 1.5 times its rated current.

5. The method for detecting multi-interval faults in AC collection lines of a large-scale new energy base according to claim 4, characterized in that, In step S2, the abrupt current characteristics on multi-interval lines are extracted using an eigenvalue decomposition-based algorithm, specifically including the following steps: S21. Array the current sampling data within the time window. Constructed in inverse Toeplitz matrix form: ； In the formula, N is the current sampling array. Number of current samples, For the first N One current sample value; S22. Perform Huber weighting on each sampled value in the matrix to reduce the weight of outliers and reduce their impact. S23. Standardize the weighted matrix to obtain a standardized matrix; extract the largest eigenvalue through eigenvalue decomposition to characterize the short-circuit current mutation characteristics.

6. The method for detecting multi-interval faults in AC collection lines of a large-scale new energy base according to claim 5, characterized in that: In S22, Huber weighting factor is: ； In the formula, For matrix OK Huber weighting factor for the column, For matrix OK The current sampling value of the column, This is the median.

7. The method for detecting multi-interval faults in AC collection lines of a large-scale new energy base according to claim 6, characterized in that, In S23 The standardized matrix is: ； Perform eigenvalue decomposition on the standardized matrix: ； In the formula, This is the eigenvalue matrix after eigenvalue decomposition, where the superscript T represents the transpose of the matrix. For matrix eigenvalues, .

8. The method for detecting multi-interval faults in AC collection lines of a large-scale new energy base according to claim 7, characterized in that: The It is 10. It is 0.7.

Images