High-Resistance Grounding Fault Detection Method Based on Integral Characteristic of Zero-Sequence Current

Abstract

The present invention discloses a high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current, which relates to the field of high-resistance grounding fault detection in 10 kV distribution networks, and includes: performing low-pass filtering on the acquired original zero-sequence current signal to obtain the low-frequency zero-sequence current signal; segmenting the obtained low-frequency zero-sequence current signal, integrating each segment, fitting the sin function according to the integral curve, and obtaining the distortion area; calculating the historical distortion area sequence, and establishing an adaptive threshold to determine whether a fault or disturbance occurs; establishing positive and negative signs of the distortion area, a sign of the degree of area distortion, and a sign of the continuous number of distortion areas to determine whether the continuous segment is a fault characteristic period; setting a time window to determine whether the number of occurrences of the fault characteristic period exceeds the limit; The present invention determines whether a fault or disturbance occurs according to the relationship between the distortion area of the current segment and the adaptive threshold, then distinguishes between high-resistance grounding faults and non-high-resistance grounding faults or system disturbances through the fault characteristic period identifier, and finally determines whether a high-resistance grounding fault occurs according to whether the number of occurrences of the fault characteristic period within the specified time window exceeds the limit. The present invention can effectively improve the reliability and sensitivity of high-resistance grounding fault detection.

Description

Technical Field

[0001] The present invention relates to the technical field of high-resistance grounding fault detection in 10 kV distribution networks, and particularly to a high-resistance grounding fault detection method based on the integral characteristics of zero-sequence current. Background Art

[0002] High-resistance grounding faults are one of the most common single-phase grounding faults in 10 kV distribution networks, mostly occurring when tree growth contacts overhead lines or when a wire breaks and touches non-metallic conductors such as cement ground or grassland. An electrical connection is easily formed between the rough surface of the wire and the grounding medium, accompanied by non-linear air arc breakdown. The long-term existence of the arc may trigger malignant accidents such as forest fires and personal electric shock. The characteristics of high-resistance grounding faults are weak, and it is difficult for traditional protection devices to detect and remove faults in a timely manner. Among the existing detection methods, based on the arc angle, characteristic quantities such as the concavity and convexity, slope, and distortion time of the zero-sequence current are calculated to quantify the non-linear distortion characteristics caused by the arc, which has high sensitivity. However, the fault detection method is easily affected by system disturbances, and at the same time, the fault state may repeatedly alternate between fault and non-fault, making it difficult to guarantee the reliability of the detection method. Summary of the Invention

[0003] To solve the problems existing in the prior art, the purpose of the present invention is to provide a high-resistance grounding fault detection method based on the integral characteristics of zero-sequence current. The present invention can reliably determine whether a high-resistance grounding fault occurs, while excluding the influence of system disturbances, achieving the purpose of quickly and effectively detecting faults. Moreover, the present invention uses a digital quantity as the output, which is more conducive to on-site maintenance personnel to judge whether a fault has occurred.

[0004] To achieve the above purpose, the technical solution adopted by the present invention is: a high-resistance grounding fault detection method based on the integral characteristics of zero-sequence current, including the following steps:

[0005] Step 1: Use the Fourier transform filtering method to perform low-pass filtering on the original waveform of the collected zero-sequence current data to obtain the filtered current after removing background noise and high-frequency signals;

[0006] Step 2: Segment the filtered current, integrate the zero-sequence current waveform in each section, then fit the sin function according to the integral curve, and calculate the distortion area according to the fitting curve and the integral curve;

[0007] Step 3: Calculate the adaptive threshold according to the historical distortion area data, compare it with the current section distortion area, and judge whether a fault or disturbance has occurred according to the result. If it is judged that a fault or disturbance has occurred, go to Step 4; otherwise, return to Step 1, judge as normal operation and record the area in the historical distortion area data;

[0008] Step 4: Calculate the fault characteristic quantity. According to the positive and negative identification of the distortion area, the degree of area distortion identification, and the continuous quantity identification of the distortion area, determine whether the continuous section simultaneously meets the three identifications. If the continuous section simultaneously meets the three identifications, determine that the current continuous section is the fault characteristic period; otherwise, return to Step 1 and determine it as a non-high-resistance grounding fault.

[0009] Step 5: If the number of fault characteristic periods ≥ 5 appears within five power frequency periods or nine continuous sections, determine that a high-resistance grounding fault has occurred; otherwise, return to Step 1 to monitor the system operation.

[0010] As a further improvement of the present invention, Step 1 is specifically as follows:

[0011] Use the Fourier transform filtering method for low-pass filtering to retain as much as possible the harmonic components causing nonlinear distortion, and obtain the zero-sequence current filtering signal within 0 Hz - 500 Hz.

[0012] As a further improvement of the present invention, Step 2 is specifically as follows:

[0013] Split the filtered current into waveforms of each section with only the positive half-cycle, that is, take the absolute value of the filtered current. The specific calculation formula is as follows:

[0014] I abs (t) = |I(t)

[0015] In the formula, I(t) is the filtered current, and I abs (t) is the absolute value of the filtered current;

[0016] Integrate I abs (t) within each section to obtain the integral curve varying with time. The specific calculation formula is as follows:

[0017]

[0018] In the formula, I n (t) is the integral function, t is the t-th sampling point of the current section, i is the sampling point, and Δt is the time step;

[0019] Use the starting point coordinates (t0, I0) of the integral curve as the starting point coordinates (-π / 2, -1) of the sin function, and the end point (t1, I1) corresponds to the coordinates (π / 2, 1) of the sin function for sin function fitting. The specific calculation formula of the fitted sin function is as follows:

[0020]

[0021] In the formula, f(t) is the fitted sin function, ω is the angular frequency of the sin function, and f s is the sampling rate;

[0022] The calculation formula for the distortion area S is as follows:

[0023]

[0024] Where N is the number of sampling data points in the section [t0, t1];

[0025] As a further improvement of the present invention, step 3 is specifically as follows:

[0026] Calculate the sequence A of the distortion areas of the ten sections before the x-th section of data, specifically as follows:

[0027] A = {|S(t0 - T n )|, |S(t0 - 2T n )|... |S(t0 - 10T n )|}

[0028] Where T n is the calculation step size and T n = T / 2, S(t0 - kT n ) is the distortion area of the (x - k)-th section of data, k ∈ [1, 10], and k is a positive integer;

[0029] Assume that the sequence A follows a normal distribution X ~ N(μ, σ 2 ), then the calculation formula for its probability density function P(S(r)) is specifically as follows:

[0030]

[0031] Where S(r) is the distortion area of the r-th section, r ∈ [x - 1, x - 10], r is a positive integer, μ is the mean value, and σ is the standard deviation;

[0032] The larger S(r) is, the smaller the probability of the event represented by this area is. Considering the high-resistance grounding fault in the distribution network as a small-probability event, set the adaptive threshold Det for the distortion area of the x-th section based on the probability density function of the historical area data sequence A. The calculation formula for the adaptive threshold Det is specifically as follows:

[0033] Det = μ + 4σ

[0034] As a further improvement of the present invention, the fault or disturbance detection criterion is specifically as follows:

[0035] If the distortion area S of the x-th section > Det, it is judged as a fault or disturbance; if S ≤ Det, it is judged as normal operation;

[0036] As a further improvement of the present invention, step 4 is specifically as follows:

[0037] Calculate the fault characteristic quantity:

[0038] Positive and negative sign identification P of distortion area m :

[0039] If the distortion area S > 0, then mark the positive and negative sign identification P of the distortion area m = 1; if the distortion area S ≤ 0, then mark the positive and negative sign identification P of the distortion area m = 0;

[0040] Degree of area distortion identification P n , and the calculation formula of the area distortion degree Ds is as follows:

[0041]

[0042] If the area distortion degree Ds > α, then mark the degree of area distortion identification P n = 1; if the area distortion degree Ds ≤ α, then mark the degree of area distortion identification P n = 0; where α is a threshold, and α takes 1% - 3%;

[0043] Continuous quantity identification N of distortion area p :

[0044] Calculate the positive and negative sign identification of the distortion area and the degree of area distortion identification of the (x + 1)-th segment of data. If the data of the x-th and (x + 1)-th segments simultaneously satisfy P m = P n = 1, then mark the continuous quantity identification N of the distortion area p = 1; otherwise, mark the continuous quantity identification N of the distortion area p = 0;

[0045] As a further improvement of the present invention, the fault feature period detection criterion is as follows:

[0046] If the data of the continuous section (the x-th and (x + 1)-th segments) satisfies the following formula:

[0047]

[0048] Then determine that the continuous section is the fault feature period, and mark the fault feature period identification F T = 1;

[0049] As a further improvement of the present invention, step 5 is specifically as follows:

[0050] If the number of fault feature periods in five power frequency periods or nine continuous sections ≥ 5, that is, the number of times the output fault feature period identification F T = 1 exceeds 5 times, then determine that a high-resistance grounding fault has occurred.

[0051] The beneficial effects of the present invention are:

[0052] Considering that the high-resistance grounding fault is one of the most common single-phase grounding faults in the 10kV distribution network, the high-resistance grounding fault will be accompanied by the breakdown of non-linear air arc. The long-term existence of the arc may trigger malignant events such as forest fires and biological electric shock, resulting in significant economic losses and casualties. The characteristics of high-resistance grounding faults are weak, and the traditional detection methods have low sensitivity and are easily affected by system disturbances. If the occurrence of the fault can be detected in time, the influence of system disturbances can be effectively eliminated, and misjudgment can be avoided, which will significantly improve the stability of the safe operation of the system. The high-resistance grounding fault detection method proposed by the present invention is to isolate the zero-sequence current low-frequency signal in the original signal, segment the signal, obtain the distortion area according to the integral and fitting curve, and then establish an adaptive threshold for the current section through the historical distortion area data to reliably judge whether a fault or disturbance occurs. Subsequently, the fault characteristic quantity is calculated, and the non-high-resistance grounding fault and disturbances such as system capacitance, load switching, and transformer inrush current are distinguished according to the positive and negative identification of the distortion area, the degree of area distortion identification, and the continuous number identification of the distortion area, and the fault characteristic period is judged. Finally, the occurrence of high-resistance grounding fault is judged according to the over-limit of the number of fault characteristic periods within the specified time window, which improves the sensitivity and reliability of high-resistance grounding fault detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 is the flowchart of the embodiment of the present invention;

[0054] Figure 2 is the schematic diagram of the distortion area S in the embodiment of the present invention;

[0055] Figure 3 is the probability density distribution diagram of the historical distortion area sequence A in the embodiment of the present invention;

[0056] Figure 4 is the change diagram of the adaptive threshold and the distortion area in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0058] Embodiment

[0059] Figure 1 As shown, a high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current includes the following steps:

[0060] Step 1: Use the Fourier transform filtering method to perform low-pass filtering on the original waveform of the collected zero-sequence current data to obtain the filtered current after filtering out background noise and high-frequency signals;

[0061] Step 2: Segment the filtered current, integrate the zero-sequence current waveform within each segment, then fit the sin function according to the integral curve, and calculate the distortion area based on the fitting curve and the integral curve, as Figure 2 shown below:

[0062] Split the filtered current into waveform segments with only positive half-cycles, that is, take the absolute value of the filtered current. The specific calculation formula is as follows:

[0063] I abs (t) = |I(t)

[0064] where I(t) is the filtered current, and I abs (t) is the absolute value of the filtered current;

[0065] Integrate I abs (t) within each segment to obtain an integral curve that changes with time. The specific calculation formula is as follows:

[0066]

[0067] where I n (t) is the integral function, t is the t-th sampling point in the current segment, i is the sampling point, and Δt is the time step;

[0068] Use the starting point coordinates (t0, I0) of the integral curve as the starting point (-π / 2, -1) coordinates of the sin function, and the ending point (t1, I1) corresponding to the sin function (π / 2, 1) coordinates to fit the sin function. The specific calculation formula for the fitted sin function is as follows:

[0069]

[0070] where f(t) is the fitted sin function, ω is the angular frequency of the sin function, and f s is the sampling rate;

[0071] The specific calculation formula for the distortion area S is as follows:

[0072]

[0073] where N is the number of sampling data points within the segment [t0, t1];

[0074] Step 3: Calculate the adaptive threshold based on the historical distortion area data, compare it with the current segment distortion area, and determine whether a fault or disturbance has occurred according to the result. If it is determined that a fault or disturbance has occurred, proceed to Step 4; otherwise, return to Step 1, determine normal operation and include the area in the historical distortion area data, as follows:

[0075] Calculate the sequence A of the distortion areas of the previous ten segments before the x-th segment of data, specifically as follows:

[0076] A = {|S(t0 - T n )|, |S(t0 - 2T n )|... |S(t0 - 10T n )|}

[0077] In the formula, T n is the calculation step size and T n = T / 2, S(t0 - kT n ) is the distortion area of the (x - k)-th segment of data, k ∈ [1, 10], and k is a positive integer;

[0078] Assume that the sequence A follows a normal distribution X ~ N(μ, σ 2 ), then the calculation formula of its probability density function P(S(r)) is specifically as follows:

[0079]

[0080] In the formula, S(r) is the distortion area of the r-th segment, r ∈ [x - 1, x - 10], r is a positive integer, μ is the mean value, and σ is the standard deviation;

[0081] The larger S(r) is, the smaller the probability of the event represented by this area is. Regarding the occurrence of a high-resistance grounding fault in the distribution network as a small-probability event, then set the adaptive threshold Det of the distortion area of the x-th segment based on the probability density function of the historical area data sequence A. The probability density distribution diagram of the sequence A is as Figure 3 shown. The calculation formula of the adaptive threshold Det is specifically as follows:

[0082] Det = μ + 4σ

[0083] The adaptive threshold and the change process of the distortion area are as Figure 4 shown. If the distortion area S(t) of the t-th segment > Det, it is judged as a fault or disturbance; if S(t) ≤ Det, it is judged as normal operation;

[0084] Step 4: Calculate the fault characteristic quantities: the positive and negative identification of the distortion area, the identification of the area distortion degree, and the identification of the continuous number of distortion areas. Judge whether the continuous section simultaneously meets the three identifications. If the continuous section simultaneously meets the three identifications, then judge that the current continuous section is the fault characteristic period; otherwise, return to Step 1 and judge as a non-high-resistance grounding fault; specifically as follows:

[0085] The positive and negative identification P of the distortion area m :

[0086] If the distortion area S > 0, then mark the positive and negative identification P of the distortion area m= 1; If the distortion area S ≤ 0, then mark the positive / negative sign P of the distortion area m = 0;

[0087] Positive / negative sign P of the area distortion degree n , The calculation formula of the area distortion degree Ds is as follows:

[0088]

[0089] If the area distortion degree Ds > α, then mark the positive / negative sign P of the area distortion degree n = 1; If the area distortion degree Ds ≤ α, then mark the positive / negative sign P of the area distortion degree n = 0; Where α is the threshold value, and α takes 1% - 3%;

[0090] Continuous quantity identifier N of the distortion area p :

[0091] Calculate the positive / negative sign of the distortion area and the positive / negative sign of the area distortion degree of the (x + 1)-th segment of data. If the data of the x-th and (x + 1)-th segments simultaneously satisfy P m = P n = 1, then mark the continuous quantity identifier N of the distortion area p = 1; Otherwise, mark the continuous quantity identifier N of the distortion area p = 0;

[0092] If the data of the continuous section (the x-th and (x + 1)-th segments) satisfies the following formula:

[0093]

[0094] Then determine that the continuous section is the fault feature period, and mark the fault feature period identifier F T = 1;

[0095] Step 5. If the number of fault feature periods in five power frequency periods or nine continuous sections ≥ 5, then it is determined that a high-resistance grounding fault has occurred; otherwise, return to Step 1 to monitor the system operation, as follows:

[0096] If the number of fault feature periods in five power frequency periods or nine continuous sections ≥ 5, that is, the number of times the output fault feature period identification F T = 1 exceeds 5 times, then it is determined that a high-resistance grounding fault has occurred.

[0097] In this embodiment, first, the collected original zero-sequence current signal is subjected to low-pass filtering to obtain a zero-sequence current low-frequency signal within 0 Hz - 500 Hz. Then, this zero-sequence current low-frequency signal is segmented, the waveform of each segment is integrated, a sin function is fitted according to the integral curve, and then the distortion area is calculated based on the integral curve and the fitted curve. Subsequently, an adaptive threshold is calculated by establishing a historical distortion area data sequence to determine whether a fault or disturbance occurs in the current segment. Then, positive and negative identification of the distortion area, identification of the degree of area distortion, and identification of the continuous number of distortion areas are established according to the characteristics of non-high-resistance grounding faults and system disturbances such as capacitor switching, load switching, and transformer inrush current, respectively, to determine whether the continuous segment is a fault characteristic period. Finally, by setting a time window and judging the number of times the fault characteristic period appears within the specified time window, it is determined whether a high-resistance grounding fault occurs, improving the reliability and sensitivity of high-resistance grounding fault detection.

[0098] This embodiment has the following advantages: Low-pass filtering the original zero-sequence current signal can effectively filter out background noise and better retain harmonic signals that cause non-linear distortion. Segmenting the zero-sequence current signal can improve the subsequent processing efficiency. Integrating the waveform of each segment, fitting a sin function according to the integral curve, and calculating the distortion area based on the integral curve and the fitted curve can quantify the non-linear characteristics caused by the arc. The adaptive threshold calculated based on the historical distortion area data takes into account both the reliability and sensitivity of fault detection. Calculating the positive and negative identification of the distortion area, identification of the degree of area distortion, and identification of the continuous number of distortion areas can reliably distinguish non-high-resistance grounding faults and system disturbances. Setting a time window to judge whether the number of times the fault characteristic period appears exceeds the limit can prevent the fault state from fluctuating between fault and non-fault or the arc from repeatedly extinguishing unstably and reigniting, preventing missed judgments, and improving the reliability and sensitivity of high-resistance grounding fault detection.

[0099] The above-described embodiments merely represent specific implementation manners of the present invention. Their descriptions are relatively specific and detailed, but should not be construed as limiting the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current, characterized in that It includes the following steps: Step 1: Use the Fourier transform filtering method to perform low-pass filtering on the collected zero-sequence current data to obtain the filtered current after removing background noise and high-frequency signals; Step 2: Segment the filtered current, integrate the zero-sequence current waveform in each segment, then fit the sin function according to the integral curve, and calculate the distortion area based on the fitting curve and the integral curve; The specific content of Step 2 is as follows: Split the filtered current into waveform segments with only positive half-cycles, that is, take the absolute value of the filtered current. The specific calculation formula is as follows: I abs (t) = |I(t)| Wherein, I(t) is the filtered current, and I abs (t) is the absolute value of the filtered current; Integrate I abs (t) within each section to obtain an integral curve that varies with time. The specific calculation formula is as follows: Wherein, I n (t) is an integral function, t is the t-th sampling point in the current section, i is the sampling point, and Δt is the time step; Use the starting coordinate (t0, I0) of the integral curve as the starting coordinate (-π / 2, -1) of the sin function, and the ending coordinate (t1, I1) corresponds to the sin function (π / 2, 1) coordinate for sin function fitting. The specific calculation formula of the fitting sin function is as follows: where \(f(t)\) is the fitted sine function, \(\omega\) is the angular frequency of the sine function, and \(f\) s is the sampling rate; The specific calculation formula of the distortion area S is as follows: In the formula, N is the number of sampling data points in the segment [t0, t1]; Step 3: Calculate the adaptive threshold based on the historical distortion area data, compare it with the current segment distortion area, and judge whether a fault or disturbance has occurred according to the result. If it is judged that a fault or disturbance has occurred, go to Step 4, otherwise return to Step 1, and judge as normal operation and the area is included in the historical distortion area data; Step 4: Calculate the fault characteristic quantity, and judge whether the continuous segment simultaneously satisfies the three identifiers according to the positive and negative identifier of the distortion area, the degree of area distortion identifier, and the continuous quantity identifier of the distortion area. If the continuous segment simultaneously satisfies the three identifiers, judge that the current continuous segment is the fault characteristic period, otherwise return to Step 1 and judge as a non-high-resistance grounding fault; Step 5: If the number of fault characteristic periods ≥ 5 appears within five power frequency periods or nine continuous segments, judge that a high-resistance grounding fault has occurred, otherwise return to Step 1 to monitor the system operation.

2. The high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current according to claim 1, characterized in that The specific content of Step 1 is as follows: Use the Fourier transform filtering method to perform low-pass filtering, and retain as many harmonic components causing nonlinear distortion as possible to obtain the zero-sequence current filtering signal within 0Hz - 500Hz.

3. The high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current according to claim 1, characterized in that, The specific content of Step 3 is as follows: Calculate the sequence A of the distortion areas of the ten segments before the x-th segment data, specifically as follows: A = {|S(t0 - T n )|, |S(t0 - 2T n )|... |S(t0 - 10T n )|} where T n is the calculation step size and T n = T / 2, S(t0 - kT n ) is the distortion area of the x - kth segment of data, k ∈ [1, 10], and k is a positive integer; Assume that the sequence A follows a normal distribution X ∼ N(μ, σ 2 ), then the calculation formula for its probability density function P(S(r)) is as follows: In the formula, S(r) is the distortion area of the r-th segment, r ∈ [x - 1, x - 10], r is a positive integer, μ is the mean value, and σ is the standard deviation; The larger S(r) is, the smaller the probability of the event represented by this area. Considering the occurrence of a high-resistance grounding fault in the distribution network as a small-probability event, set the adaptive threshold Det of the x-th segment distortion area based on the probability density function of the historical area data sequence A. The specific calculation formula of the adaptive threshold Det is as follows: Det = μ + 4σ.

4. The high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current according to claim 3, wherein The fault or disturbance detection criterion is specifically as follows: If the distortion area S of the x-th segment > Det, it is judged as a fault or disturbance; if S ≤ Det, it is judged as normal operation.

5. The high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current according to claim 1, characterized in that The specific content of Step 4 is as follows: Calculate the fault characteristic quantity: Positive and negative sign of distortion area P m : If the distortion area S > 0, then mark the positive / negative sign P of the distortion area m = 1; if the distortion area S ≤ 0, then mark the positive / negative sign P of the distortion area m = 0; Area distortion degree identifier P n , the calculation formula for the area distortion degree Ds is as follows: If the area distortion degree Ds > α, then mark the area distortion degree identifier P n = 1; if the area distortion degree Ds ≤ α, then mark the area distortion degree identifier P n = 0; where α is a threshold value, and α takes 1%; Aberration area continuous quantity identifier N p : Calculate the positive / negative identification of the distortion area and the identification of the degree of area distortion for the (x + 1)-th segment of data. If both the x-th and (x + 1)-th segments of data satisfy P m = P n = 1, then mark the continuous quantity identification of the distortion area N p = 1; otherwise, mark the continuous quantity identification of the distortion area N p = 0.

6. The high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current according to claim 5, wherein The fault characteristic period detection criterion is specifically as follows: If the data of the continuous segment (the x-th and x + 1-th segments) satisfies the following formula: Then it is determined that the continuous section is the fault feature period, and the fault feature period identifier F is marked T = 1 7. The high-resistance grounding fault detection method based on the integral characteristic of zero-sequence current according to claim 1, wherein The specific content of Step 5 is as follows: If the number of fault characteristic periods ≥ 5 within five power frequency cycles or nine consecutive sections, that is, the number of times the output fault characteristic period identification F T = 1 exceeds 5 times, it is determined that a high-resistance grounding fault has occurred.

Images