Power distribution network fault line selection method and system

Abstract

The invention relates to a power distribution network fault line selection method and system. The method comprises the steps of judging whether a zero sequence voltage value of a neutral point exceeds a preset proportion of a phase voltage value or not according to a preset power distribution network operation regulation; if so, determining that a grounding fault occurs, and triggering a line selection program to record the transient zero-sequence current of each line; decomposing the transient zero sequence current by using variational mode decomposition (VMD) to obtain an intrinsic mode function (IMF) component corresponding to each line; a plurality of IMF components are provided; performing fast Fourier transform (FFT) analysis on the IMF component to obtain a power frequency component in each transient zero-sequence current, and removing the power frequency component to obtain a non-power frequency component of each line; after the non-power-frequency components are smoothed, a correlation coefficient matrix of each line is obtained through calculation; and inputting the correlation coefficient matrix into a K-means clustering algorithm, and obtaining a fault line according to an output clustering result. And the accuracy and efficiency of ground fault line selection are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grids, and in particular relates to a method and system for fault line selection in a distribution network. Background Art

[0002] Medium voltage distribution networks in mountainous areas face many challenges due to complex terrain, crisscrossing lines and changeable meteorological conditions. Mountain environments may increase the risk of grounding faults, making fast and accurate line selection critical for grid faults. In mountainous areas, due to the irregularity of the terrain, there may be various geological conditions and soil types, which will affect the characteristics of the grounding system. The current grounding system has gradually become a mainstream grounding method recognized by power workers in most countries due to its high safety and power supply reliability.

[0003] According to the data of distribution network operation, single-phase grounding fault is the most common fault in the low-current grounding system. After the fault occurs, because the fault current is usually small, the relay protection device does not need to act immediately, which is conducive to improving the power supply reliability of the system. Especially when a transient grounding fault occurs in the system, the grounding point can extinguish the arc by itself, and the staff does not need to take specific measures. This is of great help in improving the power supply reliability and reducing the power outage time of users. However, if the power grid has intermittent arc grounding, it will cause arc overvoltage, endanger the system insulation, and easily expand to phase-to-phase short circuit. Therefore, it is necessary to select the fault line in time to ensure the safe and stable operation of the system.

[0004] In order to identify the faulty line, the traditional method is for the distribution network staff to manually switch off each line to find the faulty line. After switching on the line, observe the ground fault indication. If the signal changes at this time, then the line at this time is the faulty line. The traditional fault line selection method can eventually find the faulty line, but the switch-off method requires cutting off the power supply of all outgoing lines one by one, which may cause operating overvoltage and reduce the reliability of power supply. Summary of the invention

[0005] In order to remedy the defects of the prior art, the present invention provides a method and system for fault line selection in a distribution network.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0007] In a first aspect, a method for fault line selection in a distribution network is provided, comprising:

[0008] According to the preset distribution network operation regulations, determine whether the zero-sequence voltage value of the neutral point exceeds the preset proportion of the phase voltage value;

[0009] If it exceeds, it is determined that a ground fault has occurred, and the line selection program is triggered to record the transient zero-sequence current of each line;

[0010] The transient zero-sequence current is decomposed using variational mode decomposition (VMD) to obtain the intrinsic mode function (IMF) components corresponding to each line; the IMF components include multiple components;

[0011] Perform fast Fourier transform (FFT) analysis on the IMF components to obtain the power frequency components in each transient zero-sequence current, and remove the power frequency components to obtain the non-power frequency components of each line;

[0012] After smoothing the non-power frequency components, the correlation coefficient matrix of each line is calculated;

[0013] The correlation coefficient matrix is ​​input into the K-means clustering algorithm, and the fault line is obtained according to the output clustering result.

[0014] Furthermore, the IMF component is subjected to fast Fourier transform FFT analysis to obtain the power frequency component in each transient zero-sequence current, and the power frequency component is removed to obtain the non-power frequency component of each line, including:

[0015] The second power frequency cycle of the first IMF component among the IMF components is selected for FFT analysis to obtain the power frequency component in each transient zero-sequence current;

[0016] After removing the power frequency components, the non-power frequency components of each line are obtained.

[0017] Furthermore, the correlation coefficient matrix of each line is calculated after smoothing the non-power frequency component, including:

[0018] After smoothing the non-power frequency components of each line, the non-power frequency components of each line are obtained as C j (k), k=1, 2, ..., N; N is the number of sampling points; j represents the jth line;

[0019] Calculate the non-power frequency component C of each line j The correlation coefficient between the first quarter periods of (k) is obtained to obtain the correlation coefficient matrix M.

[0020] Furthermore, the non-power frequency components of each line are smoothed, including:

[0021] A smoothing processing method based on a moving average filter is adopted to perform moving smoothing filtering on the non-power frequency components of each line.

[0022] Furthermore, the correlation coefficient matrix of each line is calculated, including:

[0023] The correlation coefficient is used to describe the correlation between the non-power frequency components of each line, and the value range of the correlation coefficient is between -1 and 1;

[0024] The formula for calculating the correlation coefficient is:

[0025] ;

[0026] in, represents the correlation coefficient between line i and line j, Indicates N The average value of Indicates N The average value of

[0027] According to the correlation coefficient Get the correlation coefficient matrix M:

[0028] ;

[0029] Wherein, m represents the total number of lines.

[0030] Furthermore, the correlation coefficient matrix is ​​input into the K-means clustering algorithm, and the fault line is obtained according to the output clustering result, including:

[0031] Use the correlation coefficient matrix M as the input of the K-means clustering algorithm;

[0032] The K-means clustering algorithm is used to perform cluster analysis on the data set consisting of faulty lines and non-faulty lines. According to the differences between the faulty lines and non-faulty lines, the data set is divided into a faulty set and a non-faulty set. The cluster where a single line is located is the faulty line cluster, and the cluster where the remaining lines are located is the non-faulty line cluster. The clustering results are output;

[0033] The fault line is obtained according to the clustering results.

[0034] In a second aspect, a distribution network fault line selection system is provided, comprising:

[0035] A judgment module, used to judge whether the zero-sequence voltage value of the neutral point exceeds a preset proportion of the phase voltage value according to a preset distribution network operation procedure;

[0036] The transient zero-sequence current recording module is used to determine that a ground fault has occurred if the zero-sequence voltage value of the neutral point exceeds a preset proportion of the phase voltage value, and trigger the line selection program to record the transient zero-sequence current of each line;

[0037] The current decomposition module is used to decompose the transient zero-sequence current using variational mode decomposition VMD to obtain the intrinsic mode function IMF components corresponding to each line; the IMF components include multiple;

[0038] The power frequency component elimination module is used to perform fast Fourier transform (FFT) analysis on the IMF component to obtain the power frequency component in each transient zero-sequence current, and eliminate the power frequency component to obtain the non-power frequency component of each line;

[0039] A correlation coefficient matrix generation module is used to calculate the correlation coefficient matrix of each line after smoothing the non-power frequency components;

[0040] The fault line clustering analysis module is used to input the correlation coefficient matrix into the K-means clustering algorithm and obtain the fault line according to the output clustering results.

[0041] Furthermore, the power frequency component elimination module includes:

[0042] A power frequency component acquisition unit and a power frequency component elimination unit;

[0043] A power frequency component acquisition unit is used to select the second power frequency period of the first IMF component in the IMF components for FFT analysis to obtain the power frequency component in each transient zero-sequence current;

[0044] The power frequency component elimination unit is used to obtain the non-power frequency components of each line after eliminating the power frequency components.

[0045] Furthermore, the correlation coefficient matrix generation module includes:

[0046] Non-power frequency component processing unit and correlation coefficient matrix generating unit;

[0047] The non-power frequency component processing unit is used to smooth the non-power frequency components of each line to obtain the non-power frequency components of each line as C j (k), k=1, 2, ..., N; N is the number of sampling points; j represents the jth line;

[0048] Correlation coefficient matrix generation unit, used to calculate the non-power frequency component C of each line j The correlation coefficient between the first quarter periods of (k) is obtained to obtain the correlation coefficient matrix M.

[0049] Furthermore, the correlation coefficient matrix generating unit is specifically used to describe the correlation degree between the non-power frequency components of each line by using the correlation coefficient, and the value range of the correlation coefficient is between -1 and 1;

[0050] The formula for calculating the correlation coefficient is:

[0051] ;

[0052] in, represents the correlation coefficient between line i and line j, Indicates N The average value of Indicates N The average value of

[0053] According to the correlation coefficient Get the correlation coefficient matrix M:

[0054] ;

[0055] Wherein, m represents the total number of lines.

[0056] The beneficial effects achieved by the present invention are:

[0057] According to the preset distribution network operation procedures, it is determined whether the zero-sequence voltage value of the neutral point exceeds the preset proportion of the phase voltage value; if it exceeds, it is determined that a ground fault has occurred, and the line selection program is triggered to record the transient zero-sequence current of each line; the transient zero-sequence current is decomposed using variational mode decomposition VMD to obtain the intrinsic mode function IMF component corresponding to each line; the IMF component includes multiple components; the IMF component is subjected to fast Fourier transform FFT analysis to obtain the power frequency component in each transient zero-sequence current, and the power frequency component is eliminated to obtain the non-power frequency component of each line; after smoothing the non-power frequency component, the correlation coefficient matrix of each line is calculated; the correlation coefficient matrix is ​​input into the K-means clustering algorithm, and the fault line is obtained according to the output clustering result. Combining VMD with FFT avoids the interference of transient signals. Taking the ideal transient zero-sequence current as an example, it is proved that the power frequency component and the noise signal can be accurately separated to obtain the non-power frequency component of the transient zero-sequence current. It is not affected by various fault conditions and has strong anti-interference ability. The method is simple and reliable and does not require a large number of complex calculations, which greatly improves the efficiency of fault line selection and also improves the accuracy of fault line selection. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flow chart of a method for selecting a fault line in a power distribution network according to the present invention;

[0059] Figure 2 This is the transient equivalent circuit diagram of the arc suppression coil grounding system fault;

[0060] Figure 3 It is a fault simulation model of overhead lines of medium-voltage mountain distribution network of the present invention;

[0061] Figure 4 It is the zero sequence current diagram of four lines of the present invention;

[0062] Figure 5 It is a non-power frequency component diagram of four circuits of the present invention;

[0063] Figure 6 is the non-power frequency component C of each line after smoothing by the present invention j (i) Schematic diagram;

[0064] Figure 7 Schematic diagram of arc current and arc voltage of the present invention;

[0065] Figure 8 It is a structural diagram of the distribution network fault line selection system of the present invention. DETAILED DESCRIPTION

[0066] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0067] like Figure 1 As shown, an embodiment of the present invention provides a method for fault line selection in a distribution network, comprising:

[0068] 101, judging whether the zero-sequence voltage value of the neutral point exceeds a preset proportion of the phase voltage value according to a preset distribution network operation procedure;

[0069] When a single-phase grounding fault occurs in a medium-voltage mountain distribution network, the steady-state values ​​measured are generally small, especially when the grounding resistance is large. It is very easy to make misjudgments when using steady-state information for line selection. Because the transient process signals are relatively rich, it is necessary to reasonably use the information contained in the transient process, which is very helpful for improving the accuracy of line selection. If a grounding fault occurs in the arc suppression coil grounding system, the transient zero-sequence current at the grounding point contains capacitive current and inductive current. In order to clearly explain the transient process of the arc suppression coil grounding system, a system fault transient equivalent circuit is established, such as Figure 2 As shown. Figure 2 In the figure, C0 is the total capacitance of the neutral point grounding system through the arc suppression coil, L0 is the equivalent inductance in the zero-sequence loop, R0 is the equivalent resistance in the zero-sequence loop, and R P , L P are the active loss resistance and inductance of the arc suppression coil, U 0f is the equivalent zero-sequence power supply voltage, i0 is the transient zero-sequence current at the fault point, i L is the transient inductor current, i C is the transient capacitor current. The transient zero-sequence current of the fault is mainly composed of i C and i L It consists of two parts, and its characteristics depend on the specific conditions of the grounding phase angle and system parameters. C and the transient inductor current i L Both contain power frequency components, and the two will compensate each other. In the initial stage of single-phase grounding, the characteristics of the fault zero-sequence current are mainly determined by i C The mathematical expression of transient ground current i0 is:

[0070] ;

[0071] In the above formula, I Cm is the amplitude of the capacitor current, I Lm is the amplitude of the inductor current, is the angular frequency of the transient free oscillation component, is the attenuation coefficient of the free oscillation component, , among which is the decay time constant of the capacitor current, is the decay time constant of the inductor current;

[0072] It is the power frequency component of the zero-sequence current after compensation. Its magnitude and polarity depend on I Cm and I Lm The difference in magnitude of ; It is the transient high-frequency oscillation component of the zero-sequence current of the fault, and its magnitude is related to the initial phase angle of the fault; It is the decaying DC component of the zero-sequence current of the fault, produced by the arc suppression coil and only exists in the fault line. Its magnitude is related to the initial phase angle of the fault. About, when When , its value is maximum; when When , its value is minimum. For grounding through arc suppression coils, when a single-phase grounding fault occurs in the line, a transient zero-sequence current with a rapid decay will flow through the grounding point. The amplitude of the grounding current at the initial stage of the fault is several times the steady-state amplitude, but this current does not last long, and it will basically transition to the steady-state stage in about 1 cycle, and the amplitude will be greatly reduced. In a system grounded through an arc suppression coil, the polarity of the power frequency component of the fault line is uncertain and is determined according to the size of the arc suppression coil compensation current; therefore, the power frequency component not only cannot be used as the basis for selecting the fault line, but the existence of the power frequency component will also interfere with the accuracy of the fault line, so the present invention proposes to use non-power frequency components for fault line selection.

[0073] According to the distribution network operation regulations, the neutral point zero sequence voltage U0 shall not exceed 15% of the phase voltage U0. N When U0≤0.15U N , it is judged that no ground fault occurs.

[0074] 102, determining that a ground fault occurs, triggering a line selection program to record transient zero-sequence currents of each line;

[0075] 103, using variational mode decomposition VMD to decompose the transient zero-sequence current to obtain intrinsic mode function IMF components corresponding to each line; the IMF components include multiple;

[0076] 104, performing fast Fourier transform (FFT) analysis on the IMF components to obtain the power frequency components in each transient zero-sequence current, and eliminating the power frequency components to obtain the non-power frequency components of each line;

[0077] The power frequency component of the transient zero-sequence current cannot be directly decomposed through VMD decomposition, and further processing is required; the idea of ​​VMD is that the signal to be decomposed is composed of sub-signals of different IMFs. The analysis of steady-state harmonic signals usually uses FFT. In order to accurately obtain the power frequency component and avoid the interference of transient signals, the second power frequency cycle of the first IMF component in the IMF component is selected for FFT analysis to obtain the power frequency component in each transient zero-sequence current. The obtained power frequency component is eliminated, and finally the non-power frequency component of the transient zero-sequence current can be obtained. In order to verify that the signal decomposition method of VMD-FFT proposed in the present invention can accurately obtain the non-power frequency component of the zero-sequence current, taking the ideal zero-sequence current signal as an example, the first IMF component (IMF1) obtained by VMD decomposition is subjected to FFT analysis, and the non-power frequency component is obtained according to the above steps. The power frequency component and the white noise signal are directly removed from the ideal zero-sequence current to obtain the non-power frequency component under the ideal state. Accurate signal decomposition and extraction are the prerequisites for correctly selecting the fault line. The two curves of the non-power frequency component under ideal conditions and the non-power frequency component obtained by the signal decomposition method combined with VMD-FFT are almost identical, indicating that this method can accurately eliminate the power frequency component and noise signal, and can be used to extract the non-power frequency component of transient zero-sequence current, laying a good foundation for the subsequent selection of fault lines.

[0078] 105, after smoothing the non-power frequency components, the correlation coefficient matrix of each line is calculated;

[0079] Since transient zero-sequence current is greatly disturbed by noise signals, it is necessary to smooth each transient zero-sequence current in order to avoid the influence of high-frequency oscillation signals and noise signals on the correlation analysis. To achieve this goal, a smoothing method based on a moving average filter is adopted. The original signal is transformed by a function, that is, after a moving smoothing filter process, a smooth signal curve can be obtained. Taking an element point at a certain moment as an example, the smoothing method first takes g elements as an interval with the point as the center; then a second-order polynomial is fitted to the elements in the interval to obtain a weighted regression curve, and this step is repeated for the remaining element points to obtain their respective weighted regression curves; finally, the centers of all regression curves are connected together to obtain the smoothed signal curve; specifically,

[0080] After smoothing the non-power frequency components of each line, the non-power frequency components of each line are obtained as C j (k), k=1, 2, ..., N; N is the number of sampling points; j represents the jth line;

[0081] Calculate the non-power frequency component C of each line j (k) The correlation coefficient between the first quarter periods, and the correlation coefficient matrix M is obtained;

[0082] The correlation coefficient is used to describe the correlation between the non-power frequency components of each line, and the value range of the correlation coefficient is between -1 and 1;

[0083] The formula for calculating the correlation coefficient is:

[0084] ;

[0085] in, represents the correlation coefficient between line i and line j, Indicates N The average value of Indicates N The average value of

[0086] According to the correlation coefficient Get the correlation coefficient matrix M:

[0087] ;

[0088] Wherein, m represents the total number of lines.

[0089] When the correlation coefficient is positive, it means that the two components are positively correlated, that is, when one component increases, the other component also increases. The closer the correlation coefficient is to 1, the stronger the positive correlation between the two components. When the correlation coefficient is 0, the two are uncorrelated. When the correlation coefficient is negative, it means that the two components are negatively correlated, that is, when one component increases, the other component decreases. The correlation coefficient is a very important characteristic quantity that describes random signals. It can measure the degree of correlation between two signals. The correlation coefficient matrix M is constructed based on the correlation coefficient between them.

[0090] 106, input the correlation coefficient matrix into the K-means clustering algorithm, and obtain the fault line according to the output clustering result.

[0091] The correlation coefficient matrix is ​​used as the input of the K-means clustering algorithm. The K-means clustering algorithm is used to perform cluster analysis on the data set composed of faulty lines and non-faulty lines. According to the differences between faulty lines and non-faulty lines, the data set can be divided into two categories: faulty and non-faulty. The cluster where a single line is located is the faulty line cluster, and the cluster where the remaining lines are located is the non-faulty line cluster. First, two row vectors are randomly selected from the data set composed of faulty lines and non-faulty lines as the cluster centers P of the faulty line and non-faulty line clusters. Then, the distances from the row vectors in the data set to the centers of the two clusters are calculated respectively, and they are assigned to the corresponding clusters according to the distance. Then, the clusters assigned above are calculated to obtain the new cluster centers of each cluster. All rows in the data set are re-clustered according to the new centers. The above steps are repeated until the cluster centers converge and the clustering results are output. The faulty line is obtained according to the clustering results.

[0092] The implementation principle of the embodiment of the present invention is:

[0093] According to the preset distribution network operation procedures, it is determined whether the zero-sequence voltage value of the neutral point exceeds the preset proportion of the phase voltage value; if it exceeds, it is determined that a ground fault has occurred, and the line selection program is triggered to record the transient zero-sequence current of each line; the transient zero-sequence current is decomposed using variational mode decomposition VMD to obtain the intrinsic mode function IMF component corresponding to each line; the IMF component includes multiple components; the IMF component is subjected to fast Fourier transform FFT analysis to obtain the power frequency component in each transient zero-sequence current, and the power frequency component is eliminated to obtain the non-power frequency component of each line; after smoothing the non-power frequency component, the correlation coefficient matrix of each line is calculated; the correlation coefficient matrix is ​​input into the K-means clustering algorithm, and the fault line is obtained according to the output clustering result. Combining VMD with FFT avoids the interference of transient signals. Taking the ideal transient zero-sequence current as an example, it is proved that the power frequency component and the noise signal can be accurately separated to obtain the non-power frequency component of the transient zero-sequence current. It is not affected by various fault conditions and has strong anti-interference ability. The method is simple and reliable and does not require a large number of complex calculations, which greatly improves the efficiency of fault line selection and also improves the accuracy of fault line selection.

[0094] like Figure 3 The figure shows the most common overhead line fault of the medium voltage mountain distribution network in the present invention, and the simulation model built in the PSCAD / EMTDC simulation environment. In the simulation model, it is set that the line 1 has a phase A grounding fault at a phase angle of 90°, the fault resistance is 100Ω, and the zero-sequence current of the four lines is recorded. Gaussian white noise with an SNR of 30dB is injected into the zero-sequence current to simulate the actual fault situation, and the zero-sequence current of the four lines is obtained as shown in FIG. Figure 4 As shown;

[0095] from Figure 4 It can be seen that the zero-sequence currents of the four lines all contain obvious power frequency components and transient high-frequency components. The maximum amplitudes of the zero-sequence currents of lines 2 and 3 are around 15A, and the maximum amplitude of line 4 is around 28A. Because line 4 is a pure cable line, and the cable line has a large capacitance to the ground, the zero-sequence current of line 4 is relatively larger than that of the other two lines. The maximum amplitude of fault line 1 reaches about 50A, which is much larger than that of other non-fault lines, and the polarity is also opposite to them. 0j Using VMD to decompose, the intrinsic mode function IMF1 decomposed by VMD is mainly low-frequency components, and the power frequency component is its main component. This shows that VMD can effectively decompose the frequency band where the power frequency component in the zero-sequence current is located, achieving the expected effect. IMF1 without high-frequency component interference can accurately obtain and eliminate its power frequency component using FFT analysis, and obtain non-power frequency components such as Figure 5 As shown;

[0096] The present invention can accurately remove the power frequency components and noise signals in the transient zero-sequence current of each line through the decomposition algorithm combined with VMD-FFT, laying a good foundation for subsequent accurate line selection. However, due to the large capacitance of the cable line to the ground, the non-power frequency components contain more high-frequency attenuation oscillation components, which affects the subsequent correlation analysis between lines and requires smoothing;

[0097] Figure 6 is the non-power frequency component C of each line after smoothing j (i) As can be seen from the figure, the selected smoothing method has a better effect. It can eliminate the interference of some high-frequency components without changing the trend of the line zero-sequence current and improve the accuracy of the correlation calculation. j (i) The correlation coefficient matrix is ​​obtained:

[0098] ;

[0099] Input the correlation coefficient matrix into the K-means algorithm and output the cluster center P:

[0100] ;

[0101] The cluster label is [1, 2, 2, 2], which means that line 1 is the faulty line. The line selection is completed and the line selection result is accurate.

[0102] In the existing line selection technology, few methods consider arc faults. When a single-phase grounding occurs in a low-current grounding system, it is usually accompanied by the generation of an arc. The arc also exhibits alternating extinguishing and reigniting characteristics due to the oscillation of the current at the grounding point. When an arc fault occurs in the system, the relationship between voltage and current is different from that of an ordinary resistance grounding fault. Ordinary resistance is a linear relationship, while arc fault is a time-varying process. There are two main classic models of arc characteristics: Cassie and Mayr. Cassie only considers convection heat dissipation and is mainly suitable for simulating low-resistance arcs in large current systems, while Mayr only considers radial conduction heat dissipation and is mainly suitable for simulating high-resistance arcs in small current grounding systems. Therefore, this article uses the Mayr model.

[0103] Take the case of an arc grounding fault of phase A in line 3 when the phase voltage exceeds the peak value as an example to illustrate the applicability of this method to arc grounding. The arc length is 20 cm, the fault location is 10 km away from the busbar, and the arc current and voltage are as follows: Figure 7 shown; from Figure 7 It can be seen that the arc current exhibits alternating extinguishing and reigniting characteristics when it crosses the zero point. The arc current near the zero point is called the current "zero rest" time. Figure 7It can also be seen that the arc voltage is obviously nonlinear, the arc voltage is very high at the beginning, and then decays quickly. The waveform is seriously distorted and is similar to a rectangular wave.

[0104] If an arc fault occurs at the grounding point, the arc will be extinguished and reignited, accompanied by an arc overvoltage. If the overvoltage lasts too long, it is likely to damage the insulation performance of the equipment and even destroy the stability of the system. Therefore, it is necessary to be able to accurately identify the arc fault and find the fault line. The zero-sequence current signal of the four lines after the arc fault is used with the fault line selection algorithm based on non-power frequency component clustering proposed in this chapter, and the correlation coefficient matrix M is obtained as follows:

[0105]

[0106] Input the correlation coefficient matrix into the K-means algorithm, and the output cluster center P is:

[0107]

[0108] The cluster label is [1, 1, 2, 1], which means that line 3 is the faulty line. The line selection is complete and the line selection result is accurate.

[0109] The above embodiments illustrate the distribution network fault line selection method. The following describes a distribution network fault line selection system that applies the distribution network fault line selection method.

[0110] like Figure 8 As shown, an embodiment of the present invention provides a distribution network fault line selection system, comprising

[0111] A judgment module 801 is used to judge whether the zero-sequence voltage value of the neutral point exceeds a preset proportion of the phase voltage value according to a preset distribution network operation procedure;

[0112] The transient zero-sequence current recording module 802 is used to determine that a ground fault has occurred if the zero-sequence voltage value of the neutral point exceeds a preset proportion of the phase voltage value, and trigger the line selection program to record the transient zero-sequence current of each line;

[0113] The current decomposition module 803 is used to decompose the transient zero-sequence current using variational mode decomposition VMD to obtain the intrinsic mode function IMF component corresponding to each line; the IMF component includes multiple components;

[0114] The power frequency component elimination module 804 is used to perform fast Fourier transform FFT analysis on the IMF component to obtain the power frequency component in each transient zero-sequence current, and eliminate the power frequency component to obtain the non-power frequency component of each line;

[0115] A correlation coefficient matrix generation module 805 is used to calculate the correlation coefficient matrix of each line after smoothing the non-power frequency component;

[0116] The fault line cluster analysis module 806 is used to input the correlation coefficient matrix into the K-means clustering algorithm and obtain the fault line according to the output clustering result.

[0117] Preferably, in some embodiments of the present invention, the power frequency component removal module 804 includes:

[0118] A power frequency component acquisition unit and a power frequency component elimination unit;

[0119] A power frequency component acquisition unit is used to select the second power frequency period of the first IMF component in the IMF components for FFT analysis to obtain the power frequency component in each transient zero-sequence current;

[0120] The power frequency component elimination unit is used to obtain the non-power frequency components of each line after eliminating the power frequency components.

[0121] Preferably, in some embodiments of the present invention, the correlation coefficient matrix generation module includes:

[0122] Non-power frequency component processing unit and correlation coefficient matrix generating unit;

[0123] The non-power frequency component processing unit is used to smooth the non-power frequency components of each line to obtain the non-power frequency components of each line as C j (k), k=1, 2, ..., N; N is the number of sampling points; j represents the jth line;

[0124] Correlation coefficient matrix generation unit, used to calculate the non-power frequency component C of each line j The correlation coefficient between the first quarter periods of (k) is obtained to obtain the correlation coefficient matrix M.

[0125] Preferably, in some embodiments of the present invention, the correlation coefficient matrix generating unit is specifically used to describe the correlation degree between the non-power frequency components of each line using the correlation coefficient, and the value range of the correlation coefficient is between -1 and 1;

[0126] The formula for calculating the correlation coefficient is:

[0127]

[0128] in, represents the correlation coefficient between line i and line j, Indicates N The average value of Indicates N The average value of

[0129] According to the correlation coefficient Get the correlation coefficient matrix M:

[0130]

[0131] Wherein, m represents the total number of lines.

[0132] The implementation principle of the embodiment of the present invention is:

[0133] The judgment module 801 judges whether the zero-sequence voltage value of the neutral point exceeds the preset proportion of the phase voltage value according to the preset distribution network operation procedures; if the zero-sequence voltage value of the neutral point exceeds the preset proportion of the phase voltage value, the transient zero-sequence current recording module 802 determines that a grounding fault has occurred, and triggers the line selection program to record the transient zero-sequence current of each line; the current decomposition module 803 uses variational mode decomposition VMD to decompose the transient zero-sequence current to obtain the intrinsic mode function IMF component corresponding to each line; the IMF component includes multiple components; the power frequency component elimination module 804 performs fast Fourier transform FFT analysis on the IMF component to obtain the power frequency component in each transient zero-sequence current, and eliminates the power frequency component to obtain the non-power frequency component of each line; the correlation coefficient matrix generation module 805 smoothes the non-power frequency component and calculates the correlation coefficient matrix of each line; the fault line clustering analysis module 806 inputs the correlation coefficient matrix into the K-means clustering algorithm, and obtains the fault line according to the output clustering result. Combining VMD with FFT avoids the interference of transient signals. Taking the ideal transient zero-sequence current as an example, it is proved that the power frequency component and the noise signal can be accurately separated to obtain the non-power frequency component of the transient zero-sequence current. It is not affected by various fault conditions and has strong anti-interference ability. The method is simple and reliable and does not require a large number of complex calculations, which greatly improves the efficiency of fault line selection and also improves the accuracy of fault line selection.

[0134] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0135] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0136] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0137] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0138] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims

1. A method for fault line selection in a distribution network, characterized in that: include: According to the preset distribution network operation regulations, determine whether the zero-sequence voltage value of the neutral point exceeds the preset proportion of the phase voltage value; If it exceeds, it is determined that a ground fault has occurred, and the line selection program is triggered to record the transient zero-sequence current of each line; Decomposing the transient zero-sequence current using variational mode decomposition (VMD) to obtain intrinsic mode function (IMF) components corresponding to each line; the IMF components include multiple components; Performing fast Fourier transform (FFT) analysis on the IMF components to obtain power frequency components in each transient zero-sequence current, and removing the power frequency components to obtain non-power frequency components of each line; After smoothing the non-power frequency components, a correlation coefficient matrix of each line is calculated; The correlation coefficient matrix is ​​input into the K-means clustering algorithm, and the fault line is obtained according to the output clustering result.

2. The method for selecting a fault line in a distribution network according to claim 1, characterized in that: The performing of fast Fourier transform FFT analysis on the IMF component to obtain the power frequency component in each transient zero-sequence current, and removing the power frequency component to obtain the non-power frequency component of each line, includes: Selecting the second power frequency cycle of the first IMF component among the IMF components to perform FFT analysis to obtain the power frequency component in each transient zero-sequence current; After the power frequency components are removed, the non-power frequency components of each line are obtained.

3. The method for selecting a fault line in a distribution network according to claim 2, characterized in that: The calculation of the correlation coefficient matrix of each line after smoothing the non-power frequency component includes: After smoothing the non-power frequency components of each line, the non-power frequency components of each line are obtained as C j (k), k=1, 2, ..., N; N is the number of sampling points; j represents the jth line; Calculate the non-power frequency component C of each line j The correlation coefficient between the first quarter periods of (k) is obtained to obtain the correlation coefficient matrix M.

4. The method for selecting a fault line in a distribution network according to claim 3, characterized in that: The smoothing of the non-power frequency components of each line includes: A moving average filter-based smoothing processing method is adopted to perform moving smoothing filtering processing on the non-industrial frequency components of each line.

5. The method for selecting a fault line in a distribution network according to claim 4, characterized in that: The calculation obtains the correlation coefficient matrix of each line, including: The correlation coefficient is used to describe the correlation between the non-power frequency components of each line, and the value range of the correlation coefficient is between -1 and 1; The calculation formula of the correlation coefficient is: ； Among them, the represents the correlation coefficient between line i and line j, Indicates N of the The average value, Indicates N of the The average value of According to the correlation coefficient Get the correlation coefficient matrix M: ； Wherein, m represents the total number of lines.

6. The method for selecting a fault line in a distribution network according to claim 5, characterized in that: The step of inputting the correlation coefficient matrix into a K-means clustering algorithm and obtaining the fault line according to the output clustering result includes: Using the correlation coefficient matrix M as input to a K-means clustering algorithm; Using the K-means clustering algorithm to perform cluster analysis on a data set consisting of faulty lines and non-faulty lines, dividing the data set into a faulty set and a non-faulty set according to the difference between the faulty lines and the non-faulty lines, the cluster where a single line is located is a faulty line cluster, and the clusters where the remaining lines are located are non-faulty line clusters, and outputting clustering results; The fault line is obtained according to the clustering result.

7. A distribution network fault line selection system, characterized in that: include A judgment module, used to judge whether the zero-sequence voltage value of the neutral point exceeds a preset proportion of the phase voltage value according to a preset distribution network operation procedure; A transient zero-sequence current recording module is used to determine that a ground fault has occurred if the zero-sequence voltage value of the neutral point exceeds a preset proportion of the phase voltage value, and trigger a line selection program to record the transient zero-sequence current of each line; A current decomposition module, used to decompose the transient zero-sequence current using variational mode decomposition (VMD) to obtain intrinsic mode function (IMF) components corresponding to each line; the IMF components include multiple components; A power frequency component elimination module is used to perform fast Fourier transform (FFT) analysis on the IMF component to obtain the power frequency component in each transient zero-sequence current, and eliminate the power frequency component to obtain the non-power frequency component of each line; A correlation coefficient matrix generation module, used to calculate the correlation coefficient matrix of each line after smoothing the non-power frequency component; The fault line clustering analysis module is used to input the correlation coefficient matrix into the K-means clustering algorithm and obtain the fault line according to the output clustering result.

8. The distribution network fault line selection system according to claim 7, characterized in that: The power frequency component elimination module comprises: A power frequency component acquisition unit and a power frequency component elimination unit; The power frequency component acquisition unit is used to select the second power frequency period of the first IMF component in the IMF components for FFT analysis to obtain the power frequency component in each transient zero-sequence current; The power frequency component elimination unit is used to eliminate the power frequency component to obtain the non-power frequency component of each line.

9. The distribution network fault line selection system according to claim 8, characterized in that: The correlation coefficient matrix generation module comprises: Non-power frequency component processing unit and correlation coefficient matrix generating unit; The non-power frequency component processing unit is used to smooth the non-power frequency components of each line to obtain the non-power frequency components of each line as C j (k), k=1, 2, ..., N; N is the number of sampling points; j represents the jth line; The correlation coefficient matrix generation unit is used to calculate the non-power frequency component C of each line. j The correlation coefficient between the first quarter periods of (k) is obtained to obtain the correlation coefficient matrix M.

10. The distribution network fault line selection system according to claim 9, characterized in that: The correlation coefficient matrix generating unit is specifically used to describe the correlation degree between the non-power frequency components of each line by using the correlation coefficient, and the value range of the correlation coefficient is between -1 and 1; The calculation formula of the correlation coefficient is: ； Among them, the represents the correlation coefficient between line i and line j, Indicates N of the The average value, Indicates N of the The average value of According to the correlation coefficient Get the correlation coefficient matrix M: ； Wherein, m represents the total number of lines.

Images