A line fault single-end traveling wave identification and wave head calibration method based on waveform Turing sensitivity angle detection

Abstract

The present application relates to a kind of line fault single-ended traveling wave identification and wave head calibration method based on waveform Turing sensitive angle detection, belong to transmission line fault location technical field.First, time-domain waveform is imaged, in the perspective of digital image, the most sensitive angle of wave head mutation is detected using Radon transformation variable-angle projection, and the non-axis symmetry characteristics on the two sides of the most sensitive angle of wave head are used for detection effectiveness check, wave head group containing multiple wave heads can be effectively detected, with self-checking capability, it can be applied to various bus outgoing line types.The wave head calibration method proposed in the present application has strong robustness, and the detection result is not affected by fault conditions, signal conditioning link, algorithm itself parameter.

Description

TECHNICAL FIELD

[0001] The application relates to a line fault single-end traveling wave recognition and wave head calibration method based on a waveform Turing sensitive angle detection and belongs to the technical field of power transmission line fault positioning. BACKGROUND

[0002] With the proposal of the double-carbon target, it is an inevitable choice to vigorously develop new energy. With the transformation and upgrading of the traditional power system to a new power system dominated by new energy, the system inertia no longer increases with the scale and even shows a downward trend, so it is crucial to reduce the impact on the double-high power grid. Reliable, timely and accurate positioning of power transmission line faults plays a very important role in reducing line inspection time, shortening repair and outage time, effectively reducing power transmission channel congestion and power limit duration, etc. Traveling wave fault location has the advantages of high theoretical accuracy, being unaffected by system operation mode, transition resistance and CT saturation, etc., and has been widely used in power systems. Compared with double-end traveling wave location, single-end traveling wave location has the advantages of low cost, wide coverage and no need for multi-point communication, etc. The mainstream current double-end traveling wave location also configures a single-end traveling wave location function.

[0003] Single-end location uses the time difference between the initial traveling wave and the subsequent fault point reflection wave arriving at the observation point to locate the fault, so it is necessary to detect the subsequent wave head position. However, the subsequent wave head is affected by the folding and reflection of the line topological structure and impedance discontinuous points, resulting in difficulty in identifying the fault point reflection wave head, unreliable location results, and difficulties in engineering practicality of single-end traveling wave fault identification and location. If the series of homopolarity waveform mutations caused by the initial traveling wave and the subsequent fault point reflection wave group can be accurately identified, reliable fault location can be performed using the single-end location method. For the initial traveling wave and the subsequent fault point reflection wave group, the traditional time domain method realizes wave head identification, but the wave head fault type, fault distance and other factors inevitably lead to weak wave head mutation characteristics, resulting in location failure, and it is difficult to find multiple wave heads, and lacks self-checking ability. The application is based on the inherent sensitive angle and non-axial symmetry characteristics of the fault traveling wave wave head in the perspective of the waveform image, the characteristics can be characterized by the projection line integral intensity of the waveform image in the limited direction, and the Radon transform is used to detect the mutation most sensitive angle of the waveform image and identify the non-axial symmetry on both sides of the most sensitive angle, to realize a reliable identification method for the initial traveling wave and the multiple fault reflection wave group of single-end fault. It can effectively suppress other directions and interference such as pulses and white noise, effectively improve the detection reliability and identification ability of the subsequent wave head, improve the redundancy of the single-end location formula, can independently check the results, has a wide application scenario, and provides a new idea for the expansion of single-end traveling wave analysis fault location. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a line fault single-end traveling wave recognition and wave head calibration method based on waveform sensitive angle detection, which is based on a waveform diagram, recognizes the most sensitive angle of the wave head according to the mutual corresponding relationship between the wave head and the most sensitive angle of the mutation, and checks the detection effectiveness according to the non-axial symmetry characteristics of the two sides of the wave head, so that each wave head of the fault traveling wave can be effectively recognized, the robustness of detection is high, the accuracy is high, and the algorithm has strong adaptability. The method can effectively solve the problem of difficult recognition of the second wave head in single-end distance measurement, can effectively detect multiple wave heads with the same level as the initial mutation, makes up for the shortcomings of mainstream methods such as wavelet transform that can only find the first two wave heads in most cases, each wave head has mutual checking capability, and the reliability of the wave head calibration result is greatly improved.

[0005] The technical solution of the present application is: a line fault single-end traveling wave recognition and wave head calibration method based on waveform sensitive angle detection, first, reading the phase current fault traveling wave collected by the wave recording device, under a proper time window, intercepting the fault waveform before and after the initial mutation, and making a waveform diagram; secondly, making Radon transformation on the waveform diagram within a limited angle range; thirdly, sequentially making local extremum search on the above transformation result, sequentially finding the most sensitive direction of each wave head mutation, and checking the detection effectiveness according to the non-axial symmetry characteristics of the two sides of the wave head in the most sensitive mutation direction, and repeating the above steps until the detection upper limit is reached, then stopping, then through pixel-time coordinate transformation, the corresponding time of the wave head is obtained, and the accurate time of the wave head is determined in the refined interval, and finally the wave head arrival time difference required for single-end distance measurement is obtained.

[0006] The specific steps are as follows:

[0007] Step 1: First, fault line selection and phase selection are performed, the full length of the fault line and the type information of the outgoing line of the opposite bus are read, and the relative polarity of the subsequent wave head search is determined. The initial traveling wave before ams and the traveling wave after bms of the fault phase current are intercepted, and the traveling wave data is displayed in the form of a waveform diagram, and the parameters are initialized. Among them, a can be 0.2 ms, and b can be 1.8 ms.

[0008] Step 2: Radon transformation is performed on the waveform image in the angle range [θ min ,θ max ] with a step size θ Δ , and the projection line integral result matrix R of the waveform diagram under different angles is obtained. Among them, θ min can be 75°, θ max can be 95°, and the step size θ Δ can be 1°.

[0009] Step 3: The search range is from the minimum angle to the starting search angle and from the starting vertical distance to the maximum vertical distance, that is, θ∈(θ min ,θ ref ), ρ∈(ρ ref ,ρmax ), intercept the maximum value of the corresponding local segment search in matrix R, and obtain the corresponding most sensitive angle θ seni .

[0010] Step4: Determine whether the trend of the line integral value in the symmetric angle range on both sides of the most sensitive angle satisfies formula (1). If it does, go to Step 5; if not, set the corresponding line integral value to zero according to formula (2) before reaching the search boundary, return to Step 3, and continue to search backward; if the search boundary is reached, stop.

[0011]

[0012] where k θ is the change rate of the line integral value of the wave front projected at a certain angle on one side of the most sensitive angle obtained by least squares fitting, which can be calculated according to formula (3):

[0013] R(ρ seni -Δρ:ρ seni +Δρ,θ seni )=0 (2)

[0014]

[0015] where x i is the step size of the data to be fitted, starting from 1 and increasing backward; y i is the line integral value of the wave front in the ±ε θ angle range on both sides of the most sensitive angle, and k θ and b θ are the slope and intercept obtained by fitting, respectively.

[0016] Step5: Determine whether the number of identified wave fronts meets the ranging and result checking requirements. If so, go to Step6; if not, update the search range reference according to formula (4) and return to Step3.

[0017]

[0018] where ρ seni and θ seni are the starting points of the row and column search range of the current detected wave front in matrix R, and Δρ is the near-end fault dead zone

[0019] Step 6: find the intersection of the most sensitive angle characteristic straight line and the waveform image and determine the horizontal coordinate of the start point of each wave head. The wave head space pixel distance-time difference conversion is carried out according to formula (5), and the wave head arrival time is calibrated in the fine interval corresponding to the original recording data. The multiple wave head arrival time differences, wave speeds and line lengths are substituted into the corresponding distance measurement formula, and according to the type of the opposite bus, it is judged whether it is necessary to identify according to the slope of the first wave head. The distance measurement result of the first time difference and the variance of multiple distance measurement results are output.

[0020]

[0021] Wherein, N is the sampling point number of the intercepted waveform data, M is the effective pixel point number of the horizontal coordinate in the waveform graph, m0 is the first data point in the horizontal pixel of the graph, and v is the wave speed.

[0022] The application is characterized in that the phase current traveling wave data collected by the fault recording device of the transmission line is intercepted under a proper time window, and image processing is carried out. Due to the influence of the frequency variable parameter of the transmission line, the current traveling wave propagating along the two sides of the transmission line is distorted and attenuated, and then the intensity and the mutation slope of each wave head decrease with the increase of the traveling wave propagation distance, so that each wave head has a corresponding mutation most sensitive angle. The waveform graph is projected at different angles by Radon transformation, the mutation most sensitive direction of each wave head is found out, and the non-axial symmetry characteristic of the wave head on the two sides of the corresponding most sensitive angle is verified, so that multiple wave heads of the same polarity can be effectively detected, the effective distinction between the fault wave head and the noise can be realized, and the ability to verify the correctness of itself is also possessed.

[0023] The application has the following beneficial effects:

[0024] 1. The method for detecting the mutation most sensitive angle of the wave head and then calibrating the wave head can better grasp the relative independence of the wave head, can effectively detect the wave head, and compared with the time domain detection means such as wavelet transformation, the application can reliably detect multiple wave heads and has the self-checking ability.

[0025] 2. The non-axial symmetry characteristic on the two sides of the most sensitive angle of the wave head is creatively used for the effectiveness verification of wave head detection, so that the noise can be effectively excluded, the anti-noise ability is strong, and the wave head detection accuracy is high.

[0026] 3. The wave head recognition of the traditional time domain data is converted into the detection of the most sensitive change direction and the axial symmetry of the geometric graph in the waveform image space domain, the other directions and the interference such as pulse and white noise can be effectively inhibited, the detection reliability and the recognition ability of the subsequent wave head are greatly improved, the redundancy of the single-end distance measurement formula is improved, the result can be self-verified, a new idea is provided for the single-end traveling wave analysis fault distance measurement and the extended application. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 This is a flowchart of the algorithm of the present invention;

[0028] Figure 2 This is a fault waveform diagram captured by the present invention;

[0029] Figure 3 This is a schematic diagram of the result of the Radon transform waveform used in this invention;

[0030] Figure 4 This is a schematic diagram (a) illustrating the wavehead detection effect achieved by the present invention;

[0031] Figure 5 This is a schematic diagram (b) illustrating the wavehead detection effect achieved by the present invention;

[0032] Figure 6 This is a schematic diagram (c) illustrating the wavehead detection effect achieved by the present invention;

[0033] Figure 7 This is a schematic diagram d showing the wavefront detection effect achieved by the present invention. Detailed Implementation

[0034] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0035] Example 1: As Figure 1 As shown, a method for single-end traveling wave identification and wavefront calibration of line faults based on waveform graph sensitivity angle detection is proposed. First, the phase current fault traveling wave collected by the waveform recording device is read. Under an appropriate time window, the fault waveform before and after the initial sudden change is captured and a waveform graph is plotted. Second, Radon transform is performed on the waveform graph within a limited angle range. Then, local extremum search is performed on the above transformation results to find the most sensitive direction of each wavefront sudden change. The detection effectiveness is checked by the non-axisymmetric characteristic of the wavefront on both sides of its most sensitive sudden change direction. This process is repeated until a pre-set detection upper limit is reached. Then, the corresponding time of the wavefront is obtained by pixel-time coordinate transformation, and the accurate time of the wavefront is determined in a refined range. Finally, the wavefront arrival time difference required for single-end ranging is obtained.

[0036] The specific steps are as follows:

[0037] Step 1: First, select the fault line and phase, read the total length of the fault line and the outgoing line type information of the opposite bus, and determine the relative polarity of the subsequent wavefront search. Extract the traveling wave data of the initial traveling wave of the fault phase current for a total duration of (a+b) ms (ams before and bms after), and display it in waveform format. Initialize the parameters. Here, a can be 0.2ms and b can be 1.8ms.

[0038] Step 2: Within the angle range [θ] min ,θ max [Inner step size θ] ΔRadon transform is done to the waveform image to obtain the projection line integral result matrix R of the waveform image at different angles. Where θ min Take 75°, θ max Take 95°, step θ Δ Take 1°.

[0039] Step 3: the minimum angle to the starting search angle and the starting vertical distance to the maximum vertical distance are taken as the search range, i.e. θ∈(θ min , θ ref ), ρ∈(ρ ref , ρ max ), the maximum value in the corresponding local segment of the matrix R is obtained by intercepting, and the corresponding most sensitive angle θ seni is obtained.

[0040] Step 4: whether the line integral value in the symmetrical angle range on both sides of the most sensitive angle meets the formula (1) is determined. If it meets, Step 5 is entered; if it does not meet, the corresponding line integral value is set to zero according to formula (2) when the search boundary is not reached, and Step 3 is returned to continue searching backward; if the search boundary is reached, it is stopped.

[0041]

[0042] Where k θ is the change rate of the line integral value of the wave front at a certain angle projection on one side of the most sensitive angle obtained by least square fitting, which can be obtained according to formula (3):

[0043] R(ρ seni -Δρ:ρ seni +Δρ,θ seni )=0 (2)

[0044]

[0045] In the formula, x i is the step length of the data to be fitted, which starts from 1 and increases backward; y i is the line integral value of the wave front in the angle range of ±ε θ on both sides of the most sensitive angle, and k θ and b θ are the slope and intercept obtained by fitting, respectively.

[0046] Step 5: whether the number of recognized wave fronts meets the ranging and result checking requirements is determined. If it does, Step 6 is entered; if it does not, the search range reference is updated according to formula (4), and Step 3 is returned.

[0047]

[0048] Where ρ seni and θ seniThe start of the current detected wave head in the row, column search range in the matrix R, and the near-end fault dead zone

[0049] Step 6: Find the intersection of the most sensitive angle feature straight line and the waveform image and determine the horizontal coordinates of the wave head start point. Perform wave head spatial pixel distance-time difference conversion according to formula (5), and complete the wave head arrival time calibration in the fine interval corresponding to the original recording data. Substitute the multiple wave head arrival time differences, wave speeds, and line lengths into the corresponding distance measurement formula, and according to the type of the opposite bus, determine whether it is necessary to identify according to the slope of the first wave head. Output the distance measurement result of the first time difference and the variance of multiple distance measurement results.

[0050]

[0051] Where N is the number of sampling points for waveform data, M is the number of effective horizontal pixel points in the waveform graph, m0 is the first data point in the horizontal pixel of the waveform graph, and v is the wave speed.

[0052] On the basis of the above technical solutions, further specific implementation is described.

[0053] Select a fault traveling wave recording data, and intercept the data before and after the fault phase current wave mutation to make a waveform graph. Take the first wave head position as the benchmark, intercept 2000 data points before 0.2 ms and after 1.8 ms of the initial mutation, and make a 1200*900 pixel waveform graph, as shown in Figure 2 .

[0054] Perform Radon transformation on the waveform graph with angle constraint:

[0055] Make a waveform graph with a resolution of 1200*900, make a distance scatter with a 1-pixel unit according to the diagonal pixel 1500 of the image, and perform Radon angle projection on the waveform graph from 95° with a step of -1° in the most sensitive angle range of (71°, 95°), and get a 1500*25 matrix R, as shown in Figure 3 .

[0056] Find the position of the maximum element in the matrix R (ρ max1 , θ max1 ) = (399, 20), which indicates that the line integral value is maximum at the 399th point on the projection axis when the projection angle is 0 ° °, and the corresponding mutation most sensitive angle θ sen1 = 90°.

[0057] Wave head non-axisymmetric feature confirmation: calculate the local area of the current detected wave head in θ sen1The sequence of maximum line integral values ​​within the ±3° projection range was used to obtain the rate of change of the line integral values ​​corresponding to the two sides of the most sensitive angle of the current detection wavefront, as shown in Table 1. The rate of change on both sides satisfies Equation 1: Confirm that the line corresponding to the most sensitive direction currently detected is the wavefront.

[0058]

[0059] Table 1: Confirmation of Local Asymmetric Features of Wave Head

[0060] Based on the current detected wavefront, the search domain is updated according to equation (2) to find the first subsequent wavefront, where Δρ is the equivalent pixel length of the 6km near-end fault dead zone. It is discretized into approximately 40 sampling points in the time domain coordinates at a sampling rate of 1MHz, and then converted to the image according to equation (5) to obtain the pixel length ΔL = 23px corresponding to the dead zone Δρ. Within the updated range, the maximum value of the matrix R is searched to obtain the most sensitive direction of the wavefront change. The wavefront is then confirmed by non-axisymmetric features. The results are shown in Table 1. The iteration continues to search backwards until 3 wavefronts are detected. The corresponding sensitivity angles of each detected wavefront are shown in the following lines. Figure 4 As shown.

[0061] The starting horizontal pixel is obtained by finding the intersection point. The horizontal pixel is converted into time difference according to formula (5) to obtain the time scale of each wavefront relative to the initial data point, as shown in Table 2.

[0062]

[0063]

[0064] Table 2: Coordinates of the starting point and arrival time of the wave head

[0065] Given that the busbar at the opposite end of the faulty line has multiple outgoing lines, and taking the wave velocity v = 0.298 m / μs, substituting the arrival time difference of adjacent wavefronts into the ranging formula, the fault distances are found to be 24.88 km and 25.03 km, with a variance of 0.0056. The ranging result of the arrival time difference of the first two adjacent wavefronts is 24.88 km, and the reference wavelet transform result is also 24.88 km. This indicates that the wavefront calibration is effective, the single-end ranging is reliable, and it meets the requirements of the ranging engineering.

[0066] Similarly, the other implementation effects can be obtained, such as... Figures 5-7 As shown.

[0067] The specific embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A line fault single-end traveling wave identification and wave head calibration method based on waveform Turing sensitivity angle detection, characterized in that: Firstly, the phase current fault traveling wave collected by the recording device is read, and the fault waveforms before and after the initial mutation are intercepted under the appropriate window, and the waveform diagram is drawn; secondly, the Radon transform is performed on the waveform diagram within the limited angle range; then the local extremum search is performed on the above transform results, and the most sensitive direction of each wave head mutation is found in turn, and the detection effectiveness is checked according to the non-axisymmetric characteristics of the wave head on both sides of the most sensitive mutation direction, and the above-mentioned reciprocal cycle is repeated until the detection upper limit set is reached, that is, stopped, and then the wave head corresponding time is obtained through the pixel-time coordinate transformation, and the accurate time of the wave head is determined in the refined interval, and finally the wave head arrival time difference required for single-ended distance measurement is obtained; The specific steps are as follows: Step 1: First, fault line selection and phase selection are performed, the full length of the fault line and the type information of the opposite bus outgoing line are read, the relative polarity of the subsequent wave head search is determined, the initial traveling wave before ams and the traveling wave after bms of the fault phase current are intercepted for a total of (a+b) ms time length and displayed in the form of a waveform diagram, and the parameters are initialized; Step 2: within a certain angle range with a step size Radon transform is performed on the waveform graph to obtain a waveform graph projection line integral result matrix R under different angles. Step3: take the minimum angle to the starting search angle and the starting vertical distance to the maximum vertical distance as the search range, that is , , intercept the maximum value of the corresponding local segment in the matrix R, and obtain the corresponding most sensitive angle ; Step 4: Determine whether the trend of the line integral value in the symmetric angle range on both sides of the most sensitive angle meets formula (1), if it meets, go to Step 5; if it does not meet, set the corresponding line integral value to zero according to formula (2) when the search boundary is not reached, return to Step 3 and continue to search backward; if the search boundary is reached, stop; （1） （2） （3） where x i is the step size of the data to be fitted, starting at 1 and increasing backwards; y i are the line integral values of the wave front on either side of the most sensitive angle in the angular range, , are the slope and intercept of the fit, respectively; Step 5: Determine whether the number of identified wave heads meets the requirements of distance measurement and result checking, if yes, go to Step 6, if no, update the search range reference according to formula (4) and return to Step 3; （4） wherein, , are the start of the row, column search range in matrix R where the wave front is currently detected, is the near end fault dead zone; Step 6: Find the intersection of the most sensitive angle characteristic straight line and the waveform image and determine the starting point horizontal coordinate of each wave head, convert the wave head space pixel distance-time difference according to formula (5), and complete the wave head arrival time calibration in the refined interval corresponding to the original recording data, and input the multiple wave head arrival time difference, wave speed and line length into the corresponding distance measurement formula, according to the type of the opposite bus, determine whether the first wave head slope needs to be identified according to the dual solution, output the distance measurement result of the first time difference and the variance of multiple distance measurement results; （5） Wherein, N is the number of sampling points for intercepting waveform data, M is the number of effective pixel points in the horizontal coordinate of the waveform diagram, and m0 is the first data point in the horizontal pixel of the waveform diagram.

Images