Multi-criterion small-current grounding system fault line selection method based on zero-sequence component

Abstract

The invention discloses a zero-sequence component-based multi-criterion low-current grounding system fault line selection method, and mainly relates to the technical field of power grid grounding fault line selection. Comprising the following steps of collecting zero-sequence voltage U0 at a neutral point, judging whether line selection is started or not according to the value of U0, and if the line selection is judged to be started, collecting zero-sequence current of each outgoing line; performing CEEMDAN decomposition on the collected zero sequence current to obtain a modal component of each outgoing line; extracting an energy peak value of the intrinsic mode energy of the zero-sequence current of each outgoing line and a line mode energy total value by using a mode energy method; the method has the advantages that the grey wolf optimization algorithm GWO is introduced, the optimal combination of parameters is searched by optimizing the LSSVM, and therefore the line selection precision can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid grounding fault line selection, and specifically to a multi-criterion small current grounding system fault line selection method based on zero sequence components. Background Art

[0002] The medium voltage distribution network is a key network connecting the transmission network and the electrical load. The 3-66 kV medium voltage distribution network generally operates with a neutral point not grounded or grounded through an arc suppression coil. After a single-phase grounding fault occurs in the small current grounding system, it can operate with the fault for 1-2 hours. Under transient fault conditions, the short-circuit point can extinguish the arc and cut off the fault by itself, greatly reducing the power outage probability. Moreover, after a single-phase grounding fault, the line voltage can still remain symmetrical and does not affect the normal operation of the inter-phase load. Therefore, it has strong power supply reliability.

[0003] Due to this characteristic of the neutral point not being directly grounded, most do not adopt an automatic reclosing device to ensure uninterrupted power supply. However, when a single-phase grounding fault occurs in the small current grounding system, the voltage of the non-fault phase to the ground increases. When an intermittent arc grounding occurs, it can cause arc overvoltage, threatening some insulation of the system and easily expanding into an inter-phase short circuit. The faults in the small current grounding system have problems such as weak current and unstable fault arcs, and there are various interferences under complex power grid conditions and harsh on-site working conditions, resulting in incorrect detection data. Coupled with the limitations and poor robustness of the line selection method, it has caused great difficulties in fault detection. Therefore, the difficult problem of fault line selection has not been well solved.

[0004] In the prior art, usually, multiple line selection methods such as the zero sequence current group amplitude comparison and phase comparison method, harmonic component method, negative sequence current method, zero sequence admittance method, and first half-wave method are used for single-phase fault line selection. However, due to the complex fault conditions and weak fault characteristics in the small current grounding system, the applicable range of a single line selection method has limitations and the line selection accuracy rate is low, unable to meet the needs of actual engineering.

[0005] Therefore, there is an urgent need for a multi-criterion small current grounding system fault line selection method based on zero sequence components to solve the above problems. Summary of the Invention

[0006] The purpose of the present invention is to provide a multi-criterion small current grounding system fault line selection method based on zero sequence components, which can find the optimal combination of parameters by optimizing LSSVM, thereby improving the line selection accuracy.

[0007] To achieve the above purpose, the present invention is realized through the following technical solutions:

[0008] On the one hand, a multi-criterion small current grounding system fault line selection method based on zero sequence components is provided, including the following steps:

[0009] S1: Collect the zero-sequence voltage U0 at the neutral point, and determine whether to start line selection based on the value of U0. If it is determined to start line selection, collect the zero-sequence current of each outgoing line;

[0010] S2: Perform CEEMDAN decomposition on the collected zero-sequence current to obtain the modal components of each outgoing line;

[0011] S3: Use the modal energy method to extract the energy peak of the inherent modal energy of the zero-sequence current of each outgoing line and the total line modal energy;

[0012] S4: Use the spectrogram to obtain the peaks of the zero-sequence current and the characteristic frequency band of all lines, and then complete the identification of the faulty line.

[0013] Preferably, in step S1, determining whether to start line selection according to the value of U0 is specifically:

[0014] Set the maximum nominal voltage threshold of the zero-sequence voltage U0. When the zero-sequence voltage U0 is greater than the maximum threshold, select to start line selection; otherwise, do not start line selection.

[0015] Preferably, step S2 includes:

[0016] S21: Add white noise to the original signal x(t) and perform EMD decomposition to obtain the first-order IMF component of EMD;

[0017] S22: Perform averaging on the first-order IMF component of EMD obtained in step S21 to obtain the first-order IMF component of CEEMDAN;

[0018] S23: Calculate the first residual signal and the second-order IMF component of CEEMDAN;

[0019] S24: According to steps S21 - S24, obtain the jth residual signal and calculate the (j + 1)th order IMF component of CEEMDAN;

[0020] S25: Repeat step S24. When the residual component finally meets the iteration termination condition, obtain the final residual component.

[0021] Preferably, step S3 is specifically:

[0022] When the original current signal is decomposed by CEEMDAN into and a residual component r, at this time, the energies of the inherent modal components after decomposition of the original signal x(t) are respectively:

[0023]

[0024] where m = 1, 2, …, j; n is the length of this time series signal; k is the sampling point;

[0025] {E 1 ,E 2 ,…,E j ,E r} forms an automatic division of the signal energy in the frequency domain, and the energy E of the original signal x(t) to be decomposed is the sum of the energies of each frequency band:

[0026]

[0027] Preferably, in the step S4, the process of drawing the spectrogram includes:

[0028] S41: The preliminary extraction of the characteristics of the zero-sequence current is completed by using the basic amplitude ratio method;

[0029] S42: The peak-valley values in the time domain and the peak values in the frequency domain of the characteristic frequency bands of all lines are comprehensively compared to obtain the second characteristic;

[0030] S43: The spectrogram is drawn by integrating the characteristics of steps S41 and S42.

[0031] Preferably, the step S41 is specifically:

[0032] By using the basic amplitude ratio method, for different fault conditions, the peak-valley values in the time domain and the peak values in the frequency domain of the zero-sequence current of each line are collected to complete the preliminary extraction of the characteristics of the zero-sequence current.

[0033] Preferably, the step S42 is specifically:

[0034] The fault line is determined by using the amplitude difference of the components of the transient zero-sequence current at each detection point within the frequency band;

[0035] Based on the modal energy method, for the fault line and the non-fault outgoing line, the amplitudes of the characteristic frequency bands with the largest modal energy after CEEMDAN decomposition are compared.

[0036] Preferably, in the step S4, the identification of the fault line includes:

[0037] The zero-sequence current fault characteristics obtained from the single-phase grounding fault of the resonant grounding system trained under different conditions are introduced to optimize the LSSVM parameters by GWO. After obtaining the optimal parameter combination, a GWO-LSSVM fault line selection model is trained;

[0038] The test set data is substituted into the trained classification model for data classification to determine whether the current line is a fault line. If all outgoing lines show non-fault characteristics, it is determined that the busbar has a fault; otherwise, the fault line is output.

[0039] On the other hand, a line selection system based on the multi-criterion small current grounding system fault line selection method based on zero-sequence components as described above is provided, including:

[0040] A judgment module, configured to: collect zero-sequence voltage U at the neutral point 0 , and judge whether to start line selection according to the value of U 0 . If it is judged to start line selection, collect the zero-sequence current of each outgoing line;

[0041] A decomposition module, configured to: perform CEEMDAN decomposition on the collected zero-sequence current to obtain the modal components of each outgoing line;

[0042] A calculation module, configured to: use the modal energy method to extract the energy peak value of the inherent modal energy of the zero-sequence current of each outgoing line and the total line modal energy value;

[0043] An identification module, configured to: obtain the peak values of the zero-sequence current and the characteristic frequency band of all lines by using the spectrogram, and then complete the identification of the fault line.

[0044] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0045] A new transient line selection method based on complete ensemble empirical mode decomposition with added adaptive white noise (CEEMDAN) and least squares support vector machine (LSSVM) is proposed. CEEMDAN is introduced into the zero-sequence current feature extraction process, and the constructed feature vector is input into the least squares support vector machine classifier to realize the determination of the fault line. Subsequently, on the basis of the above method, in order to improve the line selection accuracy, the grey wolf optimization algorithm (GWO) is further introduced, and a new multi-criterion fusion line selection method based on complete ensemble empirical mode decomposition with added adaptive white noise (CEEMDAN) and grey wolf optimization algorithm (GWO) optimized least squares support vector machine (LSSVM) is proposed. This algorithm can improve the line selection accuracy by optimizing the optimal combination of parameters of LSSVM. Brief Description of the Drawings

[0046] Figure 1 is the method flow chart of the present invention;

[0047] Figure 2 is the system structure schematic diagram of the present invention. Detailed Embodiments

[0048] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by this application.

[0049] In the present invention, terms such as "upper", "lower", "left", "right", "front", "rear", "vertical", "horizontal", "side", "bottom", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. They are only relational terms determined for the convenience of describing the structural relationships of various components or elements of the present invention, and do not specifically refer to any component or element in the present invention, and should not be construed as a limitation to the present invention.

[0050] In the present invention, terms such as "fixed connection", "connected", "joined" should be understood in a broad sense, which may mean a fixed connection, an integral connection or a detachable connection; it may be directly connected or indirectly connected through an intermediate medium. For those skilled in relevant scientific research or technology in this field, the specific meanings of the above terms in the present invention can be determined according to specific circumstances, and should not be construed as a limitation to the present invention.

[0051] Embodiment:

[0052] As Figure 1 shown, this embodiment provides a fault line selection method for a small current grounding system based on zero-sequence components, including the following steps:

[0053] S1: Collect the zero-sequence voltage U0 at the neutral point, and judge whether to start line selection according to the value of U0. If it is judged to start line selection, collect the zero-sequence current of each outgoing line;

[0054] S2: Perform CEEMDAN decomposition on the collected zero-sequence current to obtain the modal components of each outgoing line;

[0055] S3: Use the modal energy method to extract the energy peak value of the intrinsic modal energy of the zero-sequence current of each outgoing line and the total line modal energy value;

[0056] S4: Use the spectrogram to obtain the peak values of the zero-sequence current and the characteristic frequency band of all lines, and then complete the identification of the faulty line.

[0057] Among them, in step S1, judging whether to start line selection according to the value of U0 is specifically:

[0058] Set the maximum nominal voltage threshold of the zero-sequence voltage U0. When the zero-sequence voltage U0 is greater than the maximum threshold, select to start line selection; otherwise, do not start line selection. In this embodiment, the maximum threshold of U0 is set to 0.3UN, that is, when U0 is greater than 0.3UN, start line selection, and at the same time collect the zero-sequence current of each outgoing line at this time.

[0059] Step S2 is specifically:

[0060] (1) Add white noise ε 0 E i ω i (t) to the original signal x(t), where ε 0is the amplitude coefficient of the white noise added for the first time, ω i (t) is the white noise), and perform EMD decomposition to obtain the first-order IMF component After averaging, it becomes the first-order IMF component of CEEMDAN, that is:

[0061]

[0062] where N is the signal length;

[0063] (2) The first residual signal is:

[0064]

[0065] (3) Add white noise ε to r 1 (t) 1 E 1 ω i (t) and continue to decompose to obtain the second-order IMF component:

[0066]

[0067] (4) And so on, the j-th residual signal is:

[0068]

[0069] (5) The (j + 1)-th order IMF component is:

[0070]

[0071] (6) Loop steps (4) and (5). When the residual component finally meets the iteration termination condition, the algorithm terminates. Suppose there are J IMF components at this time, and the final residual component is:

[0072]

[0073] Step S3 is specifically as follows: When the original current signal is decomposed by CEEMDAN into and a residual component r, at this time, the energies of the intrinsic mode components after the decomposition of the original signal x(t) are respectively:

[0074]

[0075] where m = 1, 2,..., j; n is the length of this time series signal; k is the sampling point;

[0076] {E 1 , E 2 , …, E j , E r}An automatic division of the signal energy in the frequency domain is formed, and the energy E of the original signal x(t) to be decomposed is the sum of the energies of each frequency band:

[0077]

[0078] When a single-phase fault occurs, regardless of how the fault conditions change, the inherent modal energy value of the fault line is always higher than that of the non-fault line.

[0079] In step S4, a spectrogram is plotted and the peak values of the zero-sequence current and the characteristic frequency band of all lines are obtained. The specific method is as follows:

[0080] (1) First, the basic amplitude comparison method is used. The amplitude of the transient current generated by the fault is greater than the power frequency component, and the compensation effect of the arc suppression coil on the high-frequency signal decreases significantly. For different fault conditions, the peak and valley values of the zero-sequence current of each line in the time domain and the peak value in the frequency domain are collected to complete the preliminary extraction of the characteristics of the zero-sequence current.

[0081] (2) The over-compensation effect of the arc suppression coil makes the amplitude of the zero-sequence current measured on the fault line not necessarily significantly greater than that of the non-fault outgoing line. The fault line is determined by using the amplitude difference of the components of the transient zero-sequence current at each detection point within a specific frequency band; based on the modal energy method, in the comparison of the amplitudes of the characteristic frequency bands with the largest modal energy after CEEMDAN decomposition, the fault line is significantly larger than the non-fault outgoing line, and its transient information after the fault is more obvious. Therefore, comparing the peak and valley values in the time domain and the peak values in the frequency domain of the characteristic frequency bands of all lines is another fault characteristic adopted by the invention for line selection.

[0082] (3) Combining the two characteristics, a spectrogram is plotted for analysis and calculation.

[0083] In step S4, the identification of the fault line is completed, including:

[0084] Optimization of classification parameters: The zero-sequence current fault characteristics obtained from the single-phase grounding faults of the resonant grounding system under different conditions are trained. The parameters of the LSSVM are optimized by introducing the GWO. After obtaining the optimal parameter combination, a GWO-LSSVM fault line selection model is trained.

[0085] Identification of the fault line: The data in the test set is substituted into the trained classification model for data classification to determine whether the current line is a fault line. If all outgoing lines show non-fault characteristics, it is determined that there is a bus fault; otherwise, the fault line is output. (This data set refers to the historical fault data of the small current grounding system applied, which is classified according to different fault types.)

[0086] As Figure 2 shown, this embodiment also provides a multi-criterion small current grounding system fault line selection system based on zero-sequence components, including:

[0087] A judgment module, configured to: collect zero-sequence voltage U at the neutral point 0 , and determine whether to start line selection according to the value of U 0 . If it is determined to start line selection, collect the zero-sequence current of each outgoing line;

[0088] A decomposition module, configured to: perform CEEMDAN decomposition on the collected zero-sequence current to obtain the modal components of each outgoing line;

[0089] A calculation module, configured to: use the modal energy method to extract the energy peak of the inherent modal energy of the zero-sequence current of each outgoing line and the total line modal energy;

[0090] An identification module, configured to: obtain the peaks of the zero-sequence current and the characteristic frequency band of all lines by using the spectrogram, and then complete the identification of the faulty line.

[0091] The above is a specific description of the preferred embodiment of the present invention. However, the present invention is not limited to the described embodiment. Those skilled in the art can make various equivalent deformations or substitutions without departing from the spirit of the present invention, and these equivalent deformations or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for fault line selection in a small current grounding system based on multi-criteria zero-sequence components, characterized in that: The following steps are involved: S1: Collect the zero-sequence voltage U0 at the neutral point, and determine whether to start line selection based on the value of U0. If it is determined to start line selection, collect the zero-sequence current of each outgoing line; S2: Decompose the collected zero-sequence current by CEEMDAN to obtain the modal components of each outgoing line; S3: using the modal energy method to extract the energy peak value of the inherent modal energy of each outgoing line zero-sequence current and the total value of the line modal energy; S4: Use the spectrum diagram to obtain the peak values ​​of zero-sequence current and characteristic frequency bands of all lines, and then complete the fault line identification.

2. According to the method of claim 1, a multi-criteria small current grounding system fault line selection method based on zero-sequence component is characterized in that: In step S1, whether to start line selection is determined according to the value of U0, specifically: Set the maximum nominal voltage threshold of the zero-sequence voltage U0. When the zero-sequence voltage U0 is greater than the maximum threshold, the line selection is started, otherwise, the line selection is not started.

3. According to the method of claim 1, the method is characterized in that: The step S2 comprises: S21: Add white noise to the original signal x(t) and perform EMD decomposition to obtain the first-order IMF component of EMD; S22: average the first-order IMF components of the EMD obtained in step S21 to obtain the first-order IMF components of CEEMDAN; S23: Calculate the first residual signal and the second-order IMF component of CEEMDAN; S24: According to steps S21-S24, the jth residual signal is obtained, and the j+1th order IMF component of CEEMDAN is calculated; S25: Repeat step S24, and when the residual component finally meets the iteration termination condition, the final residual component is obtained.

4. According to the method of claim 1, the method is characterized in that: The step S3 is specifically: When the original current signal is decomposed into …, and a residual component r. At this time, the energy of the inherent modal components after decomposition of the original signal x(t) is: Wherein, m=1, 2, ..., j; n is the length of this time series signal; k is the sampling point; {E1, E2, ..., E j , E r } forms an automatic division of signal energy in the frequency domain, then the energy E of the decomposed original signal x(t) is the sum of the energy of each frequency band:

5. According to the method of claim 1, the method is characterized in that: In step S4, the process of drawing the spectrum diagram includes: S41: Use the basic amplitude ratio method to complete the preliminary extraction of zero-sequence current characteristics; S42: comprehensively comparing the peak-to-valley values ​​in the time domain and the peak values ​​in the frequency domain of all line characteristic frequency bands to obtain a second characteristic; S43: Combining the features of steps S41 and S42, a spectrum diagram is drawn.

6. The method for fault line selection in a small current grounding system based on multi-criteria zero-sequence components according to claim 5 is characterized in that: The step S41 is specifically as follows: The basic amplitude ratio method is used to collect the peak-to-valley values ​​of the zero-sequence current in the time domain and the peak value in the frequency domain of each line for different fault conditions to complete the preliminary extraction of the characteristics of the zero-sequence current.

7. The method for fault line selection in a small current grounding system based on multiple criteria of zero sequence components according to claim 5 is characterized in that: The step S42 is specifically as follows: The fault line is determined by using the amplitude difference of the transient zero-sequence current components in the frequency band at each detection point; Based on the modal energy method, the amplitude of the characteristic frequency band with the largest modal energy after CEEMDAN decomposition is selected for comparison between the faulty line and the non-faulty outgoing line.

8. The method for fault line selection in a small current grounding system based on multi-criteria zero-sequence components according to claim 1, characterized in that: In step S4, the fault line identification is completed, including: The zero-sequence current fault characteristics obtained from the single-phase grounding fault of the resonant grounding system trained under different conditions are introduced into GWO to optimize the LSSVM parameters. After obtaining the optimal parameter combination, the GWO-LSSVM fault line selection model is trained. Substitute the test set data into the trained classification model for data classification to determine whether the current line is a faulty line. If all outgoing lines show non-fault characteristics, the busbar is judged to be faulty; otherwise, the faulty line is output.

9. A line selection system based on the multi-criteria small current grounding system fault line selection method based on zero-sequence component as claimed in claim 1, characterized in that: include: The judgment module is used to collect the zero-sequence voltage U0 at the neutral point and judge whether to start the line selection according to the value of U0. If it is judged that the line selection is started, the zero-sequence current of each outgoing line is collected; The decomposition module is used to: decompose the collected zero-sequence current by CEEMDAN to obtain the modal components of each outgoing line; A calculation module is used to: extract the energy peak value of the inherent modal energy of each outgoing line zero-sequence current and the total value of the line modal energy by using the modal energy method; The identification module is used to obtain the peak values ​​of zero-sequence current and characteristic frequency bands of all lines by using the spectrum diagram, so as to complete the fault line identification.

Images