Discharge detection method based on partial discharge detector

Abstract

The present application belongs to the technical field of discharge detection, and specifically discloses a discharge detection method based on a partial discharge detector, which comprises the following steps: collecting vibration signals between a counterweight ball and a shell in real time, and collecting discharge detection signals synchronously; intercepting the vibration signals according to a preset time window, and calculating the matching degree of the signals in each window with a preset mechanical impact signal; recording the window with a matching degree exceeding a threshold value as an interference window, and determining an interference mode based on the proportion of the interference window; if the interference is intermittent, filtering and processing the corresponding time period of the discharge detection signal based on the time boundary of the interference window; if the interference is continuous, suspending the detection, restarting after the vibration signals meet the subsidence condition, and re-collecting the discharge detection signal in the first power frequency cycle after the recovery; and outputting a detection result based on the filtered or re-collected discharge detection signal. The present application can effectively distinguish between mechanical coupling interference of the instrument itself and a real discharge signal, and improve the detection accuracy.

Description

Technical Field

[0001] This invention belongs to the field of discharge detection technology, and more specifically, relates to a discharge detection method based on a partial discharge detector. Background Technology

[0002] Partial discharge detectors typically consist of a housing, internal signal acquisition circuitry, a discharge sensor, and a data processing unit. To improve the reliability of field operations, some detectors are also equipped with anti-collision structures such as counterweights and airbags.

[0003] During the discharge detection process, the counterweight ball moves due to changes in the shell's posture or external vibrations. This movement is coupled to the discharge sensor through the mechanical conduction of the shell, forming a false pulse that has a certain similarity to the real partial discharge signal in the time domain.

[0004] Existing partial discharge detection methods only focus on discharge signals from the device under test and external electromagnetic or electrical interference. For example, the noise-resistant discharge detection method disclosed in Chinese invention patent application number 202310049838.9 mainly focuses on distinguishing between switching noise and discharge signals in power electronic systems, using a voltage step differential method for noise reduction. That is, regarding external electrical interference sources, current methods do not consider the coupling interference generated by the mechanical movement of the instrument's own anti-collision structure. When the instrument experiences accidental impacts or the counterweight ball undergoes unexpected displacement during testing, false pulses are mixed into the collected discharge signal, which can easily be misjudged as partial discharge events inside the equipment. This can lead to false alarms, unnecessary power outages for maintenance, or misassessments of insulation status, reducing the reliability of on-site test data and the effectiveness of test results. Summary of the Invention

[0005] In view of this, in order to solve the above problems, a discharge detection method based on a partial discharge detector is proposed.

[0006] The objective of this invention can be achieved through the following technical solution: This invention provides a discharge detection method based on a partial discharge detector, the method comprising: real-time acquisition of vibration signals between the counterweight ball and the shell in the anti-collision structure corresponding to the partial discharge detector, and synchronous acquisition of discharge detection signals output by the discharge sensor in the discharge detector.

[0007] Vibration signals are captured according to a preset time window, and the matching degree between the vibration signal and the preset mechanical impact signal in each window is calculated.

[0008] Windows whose matching degree exceeds the preset matching threshold are recorded as interference windows, and the interference mode is determined based on the proportion of interference windows.

[0009] If the interference mode is intermittent, the corresponding time period in the discharge detection signal is filtered based on the time boundary of the interference window. If the interference mode is continuous, the partial discharge detection is paused and the vibration signal is monitored. When the vibration signal meets the conditions for dissipation, the discharge detection is restarted and the discharge detection signal is reacquired in the first complete power frequency cycle after recovery.

[0010] The partial discharge detection result is output based on the filtered discharge detection signal or the re-acquired discharge detection signal.

[0011] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention solves the problem of not distinguishing the coupling interference generated by the mechanical movement of the detector itself from the real discharge signal by real-time acquisition of the vibration signal between the counterweight ball and the shell in the anti-collision structure of the partial discharge detector, and calculates the matching degree of the vibration signal and the preset mechanical impact signal and the identification of the interference window. It achieves accurate identification of the source of false pulses and avoids the misjudgment of insulation status caused by false pulses being misjudged as partial discharge events inside the equipment.

[0012] (2) The present invention determines the interference mode based on the proportion of the interference window. By filtering the discharge detection signal time period corresponding to the intermittent interference, it solves the problem that false pulses caused by intermittent mechanical impacts are mixed into the discharge signal and thus misjudged as partial discharge events. It achieves effective suppression of intermittent false pulses while preserving the true discharge characteristics, and reduces the false alarm rate.

[0013] (3) This invention addresses the problem of unreliable discharge detection results caused by strong mechanical interference by actively pausing partial discharge detection and continuously monitoring vibration signals until the conditions for dissipation are met, thereby avoiding the continued output of erroneous data under invalid detection conditions and preventing unnecessary power outages for maintenance.

[0014] (4) By different adaptive processing of intermittent and continuous interference, the present invention enables the final output discharge detection signal to maintain reliability and effectiveness under different interference conditions, fully considering the coupling interference of the instrument's own structure, thereby improving the effectiveness of on-site detection data and the accuracy of insulation status assessment conclusions. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall implementation process of the present invention;

[0016] Figure 2 This is a schematic diagram of the interference mode determination process of the present invention;

[0017] Figure 3 This is a schematic diagram of the process for determining the time boundary of the interference window in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Currently, the discharge detection signals acquired by partial discharge detectors are typically analyzed directly, focusing on the discharge characteristics from the device under test and external electromagnetic or electrical interference. However, the counterweight ball in the detector's own anti-collision structure moves when the shell's posture changes or is affected by external vibrations. This movement is mechanically coupled to the discharge sensor, forming spurious pulses that are similar to real partial discharge signals in the time domain. This coupling interference caused by the instrument's own mechanical movement has not yet been identified, making it impossible to distinguish between spurious pulses and real discharge signals.

[0020] This invention discloses a discharge detection method based on a partial discharge detector. By collecting vibration signals between the counterweight ball and the shell in the anti-collision structure, and identifying interference modes based on the matching degree between the vibration signals and preset mechanical impact signals, different processing methods such as filtering or pausing detection are applied to different interference modes. This can effectively eliminate the influence of the instrument's own mechanical coupling interference on discharge detection and improve the accuracy and reliability of partial discharge detection results.

[0021] Please refer to details. Figure 1 As shown, the present invention provides a discharge detection method based on a partial discharge detector. The method includes: S1, real-time acquisition of vibration signals between the counterweight ball and the shell in the anti-collision structure corresponding to the partial discharge detector, and synchronous acquisition of discharge detection signals output by the discharge sensor in the discharge detector.

[0022] The vibration signal is collected by a vibration sensor fixedly installed on the inner wall of the housing, and is a sequence of vibration amplitude values.

[0023] Specifically, the partial discharge detector includes a housing, a discharge sensor disposed within the housing, and a shock-absorbing structure containing a counterweight ball. This counterweight ball can move within the housing and generates mechanical vibration when the instrument is subjected to impact. To monitor this vibration, one or more vibration sensors (e.g., piezoelectric accelerometers or MEMS accelerometers) are fixedly mounted on the inner wall of the housing near the area where the counterweight ball moves. The output of the vibration sensor is converted from analog to digital to form a discrete amplitude sequence arranged in time order.

[0024] The vibration signal and the discharge detection signal output by the discharge sensor (such as a high-frequency current transformer, an ultra-high frequency antenna, or a coupling capacitor) are synchronously acquired using the same clock source to ensure strict correspondence on the time axis.

[0025] The above method can obtain the mechanical vibration signal generated by the movement of the counterweight ball and the discharge detection signal that may be mixed with mechanical coupling interference, thus providing a data basis for subsequent interference identification and processing.

[0026] S2. Extract vibration signals according to the preset time window and calculate the matching degree between the vibration signal and the preset mechanical impact signal in each window.

[0027] Understandably, this embodiment first requires windowing processing of the continuously acquired vibration signals. The preset time window is preferably set to two complete power frequency cycles to ensure synchronization with the discharge signal. Each window is sequentially extracted with a sliding step size (e.g., 20ms, i.e., 50% overlap) to ensure that no possible impact events are missed. The power frequency cycle refers to the frequency cycle of the AC power supply connected to the device under test, which is usually 50Hz or 60Hz, corresponding to a cycle length of 20ms.

[0028] Considering that the mechanical vibration generated by the counterweight ball impacting the shell manifests in the time domain as a series of peaks with fixed time intervals and progressively decreasing amplitudes, this waveform is determined by the moment of impact and subsequent bounces. In contrast, the vibration waveforms generated by other mechanical vibration sources (such as the shell colliding with external objects, loosening of internal components, or transportation vibrations) differ significantly from the counterweight ball impact signal in terms of the time interval between peaks and the amplitude decay pattern.

[0029] Based on this, the embodiments of the present invention accurately identify whether the vibration signal in the current window originates from the mechanical impact of the counterweight ball by introducing a matching degree between the actual vibration signal and the preset mechanical impact signal, thereby distinguishing between real instrument interference and irrelevant vibration.

[0030] Specifically, the calculation process for the matching degree between the vibration signal and the preset mechanical impact signal is as follows: S21, extract the occurrence time and amplitude value of all peak points (i.e., local maximum amplitude values) from the vibration signal within the current time window, and construct the actual peak sequence in chronological order. At the same time, extract the occurrence time and amplitude value of each peak point from the preset mechanical impact signal to construct the reference peak sequence.

[0031] Understandably, the preset mechanical impact signal is obtained in advance by means of the following method: with the detector housing in a stationary state, a miniature electromagnet fixed to the housing attracts a counterweight ball to a specific position (e.g., a preset distance from the inner wall of the housing). Then, the electromagnet is de-energized, releasing the counterweight ball, causing it to impact the inner wall of the housing at a consistent speed and angle under the action of gravity or spring force. A vibration sensor records the entire process from the start of the impact until the vibration decays to the background noise level, which is then used as the preset mechanical impact signal. To ensure reliability, the above operation can be repeated multiple times, and the averaged signals from the multiple recordings are used as the final preset mechanical impact signal.

[0032] S22. Pair the actual peak sequence with the reference peak sequence in chronological order and count the number of successful pairings. The time deviation of the pairing shall not exceed three sampling periods. Peaks exceeding this deviation are considered to be from other vibration sources and shall not participate in the pairing.

[0033] Understandably, the propagation of the vibration signal within the housing and the sensor's response delay will cause a slight offset between the actual peak occurrence time and the reference, typically not exceeding three sampling cycles. Therefore, this pairing time deviation is set.

[0034] As a preferred example, to further improve the accuracy of pairing and avoid interference from random vibration peaks caused by non-counterweight ball impacts on the matching degree calculation, the actual peak sequence can be further screened. Specifically, when pairing in chronological order, for the current peak in the actual peak sequence, if its vibration signal amplitude is less than the vibration signal amplitude of its previous successfully paired peak (if any), and simultaneously less than the vibration signal amplitude of the peak at the corresponding position in the benchmark peak sequence, then the peak is considered not to conform to the amplitude decay law that should exist after the counterweight ball impact, and therefore is excluded and not included in subsequent pairing and matching degree calculations.

[0035] S23. If the number of successful pairings is 0, then the matching degree of the current window is directly determined to be 0. Otherwise, take the maximum value between the peak number of the actual peak sequence and the peak number of the reference peak sequence, divide the number of successfully paired peaks by this maximum value, and obtain the peak time similarity.

[0036] It is understandable that in actual vibration, some peak values ​​may be lost or extra peak values ​​may be generated due to interference. Using the method of dividing the number of successful pairings by the maximum number of peak values ​​in the two sequences can reflect the degree of matching in the time sequence and avoid misjudgment due to differences in the number of peak values.

[0037] S24. For each successfully paired peak, if the amplitude values ​​of the two peaks are the same, then the amplitude similarity of the pair is assigned a value of 1. Otherwise, the smaller amplitude is divided by the larger amplitude, and the resulting ratio is used as the amplitude similarity of the pair.

[0038] It should be noted that the amplitude ratio between the actual peak value and the reference peak value generated by the same impact is not linearly affected by the absolute value of the impact velocity, but only reflects the relative degree of deviation. Here, the ratio of the smaller value to the larger value is used to ensure that the amplitude similarity always lies between 0 and 1, which facilitates subsequent unified comparison.

[0039] S25. Calculate the arithmetic mean of the amplitude similarity of all successfully paired peaks as the final signal amplitude similarity.

[0040] S26. The product of peak time similarity and signal amplitude similarity is used as the matching degree between the vibration signal and the preset mechanical impact signal. The closer the matching degree is to 1, the more consistent the actual vibration and the impact characteristics of the counterweight ball are.

[0041] It should be noted that peak time similarity reflects the consistency of the actual vibration event and the mechanical impact signal in terms of their occurrence patterns, while signal amplitude similarity reflects the closeness of their impact intensities. Both are used to quantify the similarity of the mechanical impact signal from both time and intensity dimensions. If only the time pattern matches but the amplitude differs significantly, it may be an accidental overlap of random vibration noise. If only the amplitudes are similar but the time patterns do not correspond, it may be interference from other vibration sources. Therefore, this embodiment of the invention uses multiplication to calculate the matching degree, ensuring that the matching degree is high when both are high and low when either is low, thus simultaneously reflecting the similarity in both dimensions.

[0042] A higher matching degree indicates that the actual vibration is more consistent with the impact characteristics of the counterweight ball. A lower matching degree indicates that the vibration within the window may originate from other mechanical vibration sources. The matching degree calculation in this step provides a basis for subsequently distinguishing between interference windows and non-interference windows.

[0043] S3. Window with a matching degree exceeding the preset matching threshold is recorded as interference window, and the interference mode is determined based on the proportion of interference window.

[0044] Each preset time window corresponds to a matching degree value. The matching degree value is between 0 and 1. The higher the value, the more consistent the vibration signal within that window is with the mechanical impact signal generated by the counterweight ball.

[0045] For example, the preferred range for the matching degree threshold is 0.4 to 0.6. The default value in this invention is 0.5, which means that the vibration signal and the preset mechanical impact signal must have a combined similarity of more than half in terms of time and amplitude before it can be identified as real interference. The window is recorded as the interference window, indicating that there is mechanical vibration caused by the impact of the counterweight ball in the window.

[0046] It should be noted that for vibration sensors with high sensitivity, the threshold can be increased to 0.6 to reduce the probability of false triggering, while for vibration sensors with low sensitivity, the threshold can be decreased to 0.4 to improve the detection capability of weak impacts.

[0047] Further, please refer to Figure 2 As shown, in order to determine the persistence of interference, it is necessary to determine the interference mode based on the proportion of the interference window. The specific process is as follows: S31, the time axis is divided into continuous power frequency period intervals with the power frequency period as the time length.

[0048] S32. Within each power frequency cycle interval, the ratio of the number of interference windows falling within that interval to the total number of windows within that interval is calculated to obtain the interference percentage for that power frequency cycle.

[0049] S33. If the proportion of interference exceeds the preset proportion threshold for at least two consecutive power frequency cycles, it is determined to be a continuous interference mode.

[0050] S34. If the proportion of interference in at least one power frequency cycle exceeds the proportion threshold, but does not meet the condition of two consecutive cycles, it is determined to be an intermittent interference mode.

[0051] S35. If the interference percentage of all power frequency cycles is lower than the percentage threshold, it is judged as an interference-free mode.

[0052] Understandably, the vibration caused by a single impact of the counterweight ball typically lasts for 1 to 2 power frequency cycles. If a high percentage of interference occurs in two or more consecutive power frequency cycles, it indicates that the counterweight ball is subjected to continuous external excitation, which constitutes persistent interference. Accordingly, this invention preferably sets the percentage threshold to 0.5, meaning that interference is considered to exist in a cycle when at least half of the window is marked as an interference window within a power frequency cycle. If this condition is met for two consecutive cycles, it is determined to be persistent interference. For example, in precision detection scenarios with high data integrity requirements, the threshold can be lowered to 0.4 to trigger pause protection more sensitively. In rapid inspection scenarios where a small amount of residual interference can be tolerated, the threshold can be raised to 0.6 to reduce unnecessary detection interruptions.

[0053] The above determination can accurately distinguish between brief, accidental impacts and continuous mechanical shocks, thus providing a basis for different subsequent discharge signal processing strategies.

[0054] S4. If the interference mode is intermittent interference, filter the corresponding time period in the discharge detection signal based on the time boundary of the interference window. If the interference mode is continuous interference, pause partial discharge detection and continue to monitor the vibration signal. When the vibration signal meets the conditions for dissipation, restart the discharge detection and reacquire the discharge detection signal in the first complete power frequency cycle after recovery.

[0055] For example, if the interference mode is determined to be intermittent interference, it indicates that there is an impact within a certain power frequency cycle. In this case, filtering the corresponding time period in the discharge detection signal can effectively recover the real discharge signal.

[0056] Before filtering, this step first requires determining the precise start and end range of each interference window on the time axis, i.e., the time boundaries, in order to locate the interference-affected periods in the discharge detection signal. Please refer to [link to relevant documentation]. Figure 3 As shown, the specific determination method is as follows: extract the first peak point and the last peak point of the vibration signal within each interference window, and take the occurrence time of the first peak point as the starting time of the time boundary.

[0057] Starting from the moment the last peak occurs, the vibration signal amplitude at each moment is compared with the preset background noise threshold value along the time axis. The background noise threshold value is set to twice the root mean square value of the vibration signal amplitude output by the vibration sensor when there is no external vibration. This root mean square value is calculated by collecting the vibration signal amplitude for a period of time (e.g., within 1 second) during the initialization of the detector.

[0058] If the vibration signal amplitudes at at least two adjacent time points are both below the background noise threshold, then the first of these time points is taken as the end time of the time boundary. Otherwise, the end time of the interference window is taken as the end time of the time boundary.

[0059] It should be noted that after the counterweight ball impacts, the vibration signal undergoes a process of rapid rise, decay after reaching a peak, and then returning to noise levels. Using the first peak point as the starting moment allows for accurate location of the instant of impact. Using two consecutive adjacent moments where the amplitude is below the noise threshold as the termination condition eliminates premature termination caused by single-point noise fluctuations, ensuring that the vibration has essentially subsided.

[0060] Furthermore, the specific processing procedure for filtering the corresponding time periods in the discharge detection signal is as follows: A1. Obtain the start and end times of each interference window. In the vibration signal, extract the occurrence times corresponding to all peak points located between the start and end times to construct a list of interference peak times. These times correspond to the impact peaks generated by the counterweight ball impact and subsequent bounces.

[0061] A2. In the discharge detection signal, for each moment in the list of interference peak moments, perform the following sub-steps: A21. If the moment is the first sampling moment of the discharge detection signal, replace its amplitude with the amplitude of the next adjacent moment.

[0062] A22. If the current moment is the last sampling moment of the discharge detection signal, then its amplitude is replaced with the amplitude of the previous adjacent moment.

[0063] A23. Otherwise, replace the discharge detection signal amplitude at that moment with the arithmetic mean of the discharge detection signal amplitudes at the previous and next adjacent moments.

[0064] The amplitude of the discharge detection signal at adjacent time points is taken from the value of the discharge detection signal before replacement.

[0065] A24. Use the replaced discharge detection signal as the filtered discharge detection signal.

[0066] Understandably, the mechanical coupling interference generated by the impact of the counterweight ball manifests as a spurious pulse synchronized with the vibration peak moment in the discharge detection signal. By extracting the vibration peak moment and replacing the amplitude of the corresponding moment in the discharge detection signal with the average of the adjacent moments, this spurious pulse can be effectively suppressed while preserving the true discharge characteristics. Averaging using the original values ​​before replacement avoids the influence of the replacement order on the calculation results.

[0067] This invention addresses the problem of false pulses mixed with discharge signals due to intermittent mechanical shocks being misjudged as partial discharge events by filtering the discharge detection signal time period corresponding to intermittent interference. It effectively suppresses intermittent false pulses while preserving the true discharge characteristics, thereby reducing the false alarm rate.

[0068] As another example, if the interference mode is determined to be continuous interference, it indicates that the counterweight ball is subjected to continuous or multiple impacts. In this case, the filtering process cannot effectively recover the true discharge signal. Therefore, the partial discharge detection is suspended, and the vibration signal is monitored to wait for the interference to subside.

[0069] Specifically, the fading conditions are as follows: during the period of suspended discharge detection, the vibration signal is divided into preset time windows, the root mean square value of the vibration signal amplitude at all times in each window is calculated, and the matching degree between the vibration signal in that window and the preset mechanical impact signal is calculated.

[0070] The fading condition is determined to be met when several consecutive preset time windows simultaneously meet the following two conditions: the root mean square value of the vibration signal amplitude at all times within the window is lower than the preset background noise threshold value.

[0071] The matching degree between the vibration signal within the window and the preset mechanical impact signal is lower than the preset matching threshold.

[0072] The number of consecutive preset time windows is no less than two, and the present invention preferably sets it to three, so as to avoid misjudgment caused by a single noise fluctuation.

[0073] It should be noted that, in order to prevent the detection from being permanently paused due to the vibration signal not meeting the decay conditions for a long time, this embodiment of the invention sets up a timeout exit mechanism: if the pause duration exceeds the preset timeout threshold (e.g., 60 seconds), the pause state is forcibly exited, the currently collected discharge detection signal is output and a warning command is initiated to prompt the operator to check the instrument status.

[0074] Once the extinction conditions are met, the discharge detection is restarted, and the discharge detection signal is reacquired during the first complete power frequency cycle after recovery.

[0075] This invention addresses the problem of unreliable discharge detection results caused by strong mechanical interference by actively pausing partial discharge detection and continuously monitoring vibration signals until the conditions for dissipation are met, thus preventing unnecessary power outages for maintenance.

[0076] Furthermore, since the discharge detection signal may be affected by residual baseline drift during continuous interference, the signal re-acquired in the first power frequency cycle after restarting may still have some deviation. To further improve detection accuracy, this embodiment adds a differential correction step after re-acquiring the discharge detection signal: the complete power frequency cycle in which the matching degree between the last vibration signal and the preset mechanical impact signal is lower than the matching threshold in each preset time window before the occurrence of continuous interference is marked as the interference-free reference power frequency cycle. The discharge detection signal in this cycle is not affected by the mechanical interference of the counterweight ball and can be used as a reference benchmark.

[0077] The discharge detection signal collected within the interference-free reference power frequency cycle is used as the reference discharge detection signal.

[0078] The discharge detection signal re-acquired during the first complete power frequency cycle after recovery will be compared with the reference discharge detection signal at corresponding moments on the time axis, and the time difference will be calculated step by step.

[0079] The differential signal is used as the discharge detection result for that power frequency cycle, replacing the original re-acquired signal for subsequent output.

[0080] It should be noted that if there is no interference-free reference power frequency cycle that meets the above conditions, the differential calculation is skipped, and the newly acquired discharge detection signal is directly used as the discharge detection result of that power frequency cycle, and a prompt message is generated to indicate that the reference reference is missing.

[0081] Understandably, before and after continuous interference, changes in the DC bias of the discharge sensor or the ambient temperature may cause baseline drift. Differential processing can eliminate common-mode baseline error, making the recovered signal consistent with the signal under interference-free conditions.

[0082] This step, through differentiated adaptive handling of intermittent and continuous interference, ensures that the final output discharge detection signal maintains reliability and effectiveness under different interference conditions. It fully considers the coupling interference of the instrument's own structure, thereby improving the effectiveness of on-site detection data and the accuracy of insulation status assessment conclusions.

[0083] S5. Based on the filtered discharge detection signal or the re-acquired discharge detection signal, output the partial discharge detection result.

[0084] After completing the above signal processing for different interference modes, the partial discharge events in the final discharge detection signal, i.e. the filtered discharge detection signal or the re-acquired discharge detection signal, are identified and statistically analyzed, and the detection results are finally output. The specific process is as follows: S51, the amplitude of each sampling moment in the signal to be analyzed is compared with the preset discharge threshold, wherein the discharge threshold is 4 times the root mean square value of the output signal amplitude of the device under test in the state of no discharge.

[0085] S52. Sequentially determine whether the amplitude of the signal to be analyzed at each moment exceeds the preset discharge threshold. A time period consisting of several consecutive moments with amplitudes exceeding the threshold is considered as a candidate discharge event. The minimum number of consecutive moments is determined based on the sampling rate. For example, when the sampling rate is 1MHz, three consecutive amplitudes exceeding the threshold are considered as one candidate discharge event to eliminate single-point noise interference.

[0086] S53. Record the start time of each candidate discharge event. Compare this start time with the time boundaries of each determined interference window. If the start time falls between the start and end times of any interference window, mark the candidate discharge event as a false event and exclude it from partial discharge statistics. Otherwise, mark it as a real discharge event and include it in partial discharge statistics.

[0087] Understandably, the mechanical coupling interference generated by the impact of the counterweight ball will produce spurious pulses in the discharge detection signal that are synchronized with the peak vibration time. The start time of these spurious pulses must fall within the time boundary of the interference window. By comparing the start time of the candidate discharge event with the interference window, it is possible to distinguish between real partial discharges and mechanically coupled spurious pulses.

[0088] S54. Count the number of all events marked as actual discharge events, their occurrence time, and amplitude characteristics to form partial discharge detection results. These results are output in the form of a numerical list, waveform annotation diagram, or alarm signal for subsequent analysis or maintenance.

[0089] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A discharge detection method based on a partial discharge detector, characterized in that, The method includes: The vibration signal between the counterweight ball and the shell in the anti-collision structure corresponding to the partial discharge detector is collected in real time, and the discharge detection signal output by the discharge sensor in the discharge detector is collected simultaneously. Vibration signals are captured according to a preset time window, and the matching degree between the vibration signal and the preset mechanical impact signal in each window is calculated. Windows whose matching degree exceeds the preset matching threshold are recorded as interference windows, and the interference mode is determined based on the proportion of interference windows. If the interference mode is intermittent, the corresponding time period in the discharge detection signal is filtered based on the time boundary of the interference window. If the interference mode is continuous, the partial discharge detection is paused and the vibration signal is monitored. When the vibration signal meets the conditions for dissipation, the discharge detection is restarted and the discharge detection signal is reacquired in the first complete power frequency cycle after recovery. The partial discharge detection result is output based on the filtered discharge detection signal or the re-acquired discharge detection signal. The specific processing procedure for filtering the corresponding time period in the discharge detection signal is as follows: Obtain the start and end times of each interference window, extract the occurrence times of all peak points located between the start and end times in the vibration signal, and construct a list of interference peak times; For each moment in the list of interference peak times in the discharge detection signal, perform the following sub-steps: If this moment is the first sampling moment of the discharge detection signal, then its amplitude is replaced with the amplitude of the next adjacent moment; If this moment is the last sampling moment of the discharge detection signal, then its amplitude is replaced with the amplitude of the previous adjacent moment; Otherwise, replace the discharge detection signal amplitude at that moment with the arithmetic mean of the discharge detection signal amplitudes at the previous and next adjacent moments. The replaced discharge detection signal is used as the filtered discharge detection signal; The amplitude of the discharge detection signal at adjacent time points is taken from the value of the discharge detection signal before replacement.

2. The discharge detection method based on a partial discharge detector as described in claim 1, characterized in that: The vibration signal is collected by a vibration sensor fixedly installed on the inner wall of the housing, and is a sequence of vibration signal amplitudes.

3. The discharge detection method based on a partial discharge detector as described in claim 1, characterized in that: The calculation process for the matching degree between the vibration signal and the preset mechanical impact signal is as follows: The time and amplitude values ​​of each peak are extracted from the vibration signal to form the actual peak sequence. At the same time, the time position and amplitude values ​​of each peak point in the preset mechanical impact signal are extracted to form the reference peak sequence. The actual peak sequence is paired with the benchmark peak sequence, and the number of successful pairings is counted. If the quantity is 0, the matching degree between the output vibration signal and the preset mechanical impact signal is 0; otherwise, take the maximum value between the peak number of the actual peak sequence and the peak number of the reference peak sequence, divide the number of successfully matched peaks by the maximum value, and obtain the peak time similarity. For each successfully paired peak, if the amplitude values ​​of the two vibration signals are the same, the signal amplitude similarity is assigned a value of 1; otherwise, the smaller vibration signal amplitude value is divided by the larger vibration signal amplitude value, and the ratio is used as the signal amplitude similarity. Calculate the mean of the signal amplitude similarity corresponding to all successfully paired peaks, and use it as the final signal amplitude similarity. The product of peak time similarity and signal amplitude similarity is used as the matching degree between the vibration signal and the preset mechanical impact signal.

4. The discharge detection method based on a partial discharge detector as described in claim 3, characterized in that: The specific steps of matching the actual peak sequence with the reference peak sequence are as follows: In chronological order, each peak in the actual peak sequence is paired with a peak in the reference peak sequence, with the pairing time deviation not exceeding three sampling periods.

5. The discharge detection method based on a partial discharge detector as described in claim 1, characterized in that: The specific process for determining the interference mode includes: Using the power frequency cycle as the time length, the time axis is divided into continuous power frequency cycle intervals; Within each power frequency cycle interval, the ratio of the number of interference windows falling within that interval to the total number of windows within that interval is used as the interference percentage for that power frequency cycle. If the proportion of interference exceeds the preset proportion threshold for at least two consecutive power frequency cycles, it is determined to be a continuous interference mode. If the proportion of interference in at least one power frequency cycle exceeds the proportion threshold, but does not meet the conditions for a continuous interference mode, it is determined to be an intermittent interference mode. If the proportion of interference in all power frequency cycles is lower than the proportion threshold, it is determined to be an interference-free mode.

6. The discharge detection method based on a partial discharge detector as described in claim 3, characterized in that: The time boundary of the interference window is determined in the following way: Extract the first and last peak points of the vibration signal within each interference window, and take the moment when the first peak point appears as the starting moment of the time boundary; Starting from the moment the last peak occurs, the vibration signal amplitude at each moment is compared with the preset background noise threshold value along the time axis. If the vibration signal amplitudes at at least two adjacent moments are both below the background noise threshold, then the first moment is taken as the end moment of the time boundary; otherwise, the end moment of the interference window is taken as the end moment of the time boundary.

7. The discharge detection method based on a partial discharge detector as described in claim 1, characterized in that: The specific extinction conditions are set as follows: If the pause duration exceeds the preset timeout threshold, the pause state will be forcibly exited, the currently acquired discharge detection signal will be output, and a warning command will be initiated. Otherwise, during the period of suspended discharge detection, the vibration signal is divided into preset time windows, and the root mean square value of the vibration signal amplitude at all times in each window is calculated, as well as the matching degree between the vibration signal in that window and the preset mechanical impact signal. When several consecutive preset time windows simultaneously meet the following two conditions, the fading condition is determined to be satisfied: The root mean square value of the vibration signal amplitude at all times within the window is lower than the preset background noise threshold. The matching degree between the vibration signal within the window and the preset mechanical impact signal is lower than the preset matching threshold.

8. The discharge detection method based on a partial discharge detector as described in claim 1, characterized in that: After reacquiring the discharge detection signal, the process also includes: Before the occurrence of continuous interference, the complete power frequency cycle in which the matching degree between the last vibration signal and the preset mechanical impact signal is lower than the matching threshold in each preset time window is marked as the interference-free reference power frequency cycle. If there is no interference-free reference power frequency cycle, the newly acquired discharge detection signal is directly used as the discharge detection result for that power frequency cycle, and a prompt message is generated. If an interference-free reference power frequency cycle exists, the discharge detection signal collected within the interference-free reference power frequency cycle is used as the reference discharge detection signal. The discharge detection signal re-acquired during the first complete power frequency cycle after recovery will be compared with the reference discharge detection signal at corresponding moments on the time axis, and the time difference will be calculated step by step. The differential signal is used as the discharge detection result for that power frequency cycle, replacing the original re-acquired signal for output.

9. The discharge detection method based on a partial discharge detector as described in claim 1, characterized in that: Before outputting the partial discharge detection result, the following is also included: The filtered discharge detection signal or the re-acquired discharge detection signal is used as the signal to be analyzed. The amplitude of the signal to be analyzed at each moment is determined sequentially to see if it exceeds the preset discharge threshold. The time period consisting of several consecutive moments when the amplitude exceeds the threshold is taken as a candidate discharge event. Record the start time of each candidate discharge event. If the start time falls within the time boundary of any determined interference window, the candidate discharge event is marked as a false event; otherwise, it is marked as a real discharge event and included in the partial discharge statistics.

Images