A method for monitoring the installation state of a lightning arrester under voltage

Abstract

The present application relates to a kind of lightning arrester live installation state monitoring methods, in the lightning arrester ground down line arrangement broadband current measurement unit obtains current signal, voltage reference measurement unit is set at high voltage end and obtains voltage reference signal;Before installation starts, current signal is collected at first time window to obtain background noise level and set pulse event threshold, the threshold is updated with background;Current signal is continuously monitored during installation, when the pulse with amplitude exceeding threshold and duration shorter than set upper limit is detected, it is recorded as candidate contact event;Stability check is carried out to voltage reference signal in second time window before and after the candidate contact event, and it is confirmed as first electrical contact event if check is passed, and it is determined that electrical contact loop has been formed, otherwise it is rejected;If first electrical contact event is not confirmed within observation time limit, it is output that electrical contact loop has not been formed.Electrical contact state of lightning arrester can be reliably identified under live installation condition, and real-time state criterion is provided for on-site installation operation.

Description

Technical Field

[0001] This invention relates to the field of online monitoring technology for high-voltage equipment in power systems, specifically a method for monitoring the energized installation status of surge arresters. Background Technology

[0002] Current technologies, such as CN114859264A (a surge arrester live-line detection device and method), improve the convenience of signal acquisition and equipment connection during surge arrester live-line detection to some extent. However, their monitoring principle still mainly relies on measuring the overall amplitude or parameters of the surge arrester's leakage current, focusing on the detection and acquisition of the surge arrester's operating state parameters. They do not specifically identify the electrical contact formation process during surge arrester live-line installation, nor do they consider the characteristic analysis of transient current behavior during installation. Therefore, they still have significant shortcomings in actual live-line installation scenarios. This technology mainly focuses on monitoring the leakage current during the surge arrester's operating state, but does not specifically detect the first electrical contact event during surge arrester live-line installation. During live-line installation, short-duration transient current pulses are often generated when installation tools or connecting parts contact the surge arrester's conductive terminals. These pulse signals have short durations, rapid amplitude changes, and may be superimposed on the system background current. Traditional methods based on leakage current amplitude or average value monitoring are difficult to effectively identify such transient contact signals. The detection process of this scheme mainly involves sampling the leakage current at the grounding terminal of the surge arrester using a current transformer, and then monitoring or parameter detection through a detection control circuit and a monitoring host. This method is more suitable for detecting the operating status parameters of the surge arrester, such as resistive current, total current, or system voltage parameters during energization. However, it lacks a dedicated identification mechanism for transient contact behavior that occurs during installation, so it cannot determine whether the surge arrester has formed an electrical contact loop, nor can it provide real-time status feedback for the energized installation process.

[0003] This technology lacks an effective mechanism for identifying transient signals in complex electromagnetic environments. In substations or power distribution equipment environments, electromagnetic interference, system load changes, power grid disturbances, and other equipment operation factors can all generate transient current fluctuations or noise signals in the grounding loop. Traditional detection methods often only perform simple measurements or parameter calculations of the current amplitude, lacking comprehensive analysis of multi-dimensional information such as the transient characteristics, duration, and phase characteristics of the current signal. Therefore, it is easy to misjudge system disturbances as equipment abnormalities or detection events, or fail to accurately identify the true contact signal from the noise background. In addition, this technology usually does not establish a dynamic background noise model, and the detection threshold is often a fixed value or an empirically set value. When the ambient noise level changes, it can easily lead to a decrease in detection sensitivity or an increase in the false judgment rate, thus affecting the reliability of the detection results. This technology usually does not combine voltage reference signals to perform phase correlation analysis of current events. In power systems, there is a clear phase relationship between current and voltage. The transient current of the installation contact is often physically related to the voltage change. However, traditional surge arrester live detection schemes usually only collect current signals for processing, and lack collaborative analysis with grid voltage phase information. This makes it difficult for the system to use voltage phase as a criterion to filter random interference signals.

[0004] Meanwhile, during power grid operation, periodic distortions, harmonic disturbances, or short-term voltage fluctuations may occur. Without a voltage signal stability verification mechanism, it is difficult to determine whether abnormal current changes are caused by system operational disturbances or actual contact behavior, thus reducing the reliability of the detection results. This technology uses a relatively simple current signal processing method, often employing a single-channel detection structure, lacking the separation and analysis of current information across different frequency bands. During live installation, contact transient currents typically manifest as high-frequency pulses, while the leakage current of the surge arrester during normal operation mainly exhibits power frequency components. If only a single signal channel is used for detection, it is difficult to distinguish between transient pulses and power frequency current changes, and it is also impossible to effectively identify current step changes caused by changes in system operating conditions. System load changes, switching operations, or power grid disturbances may cause significant changes in the grounding loop current, but these changes do not necessarily indicate installation contact events. Without an independent detection and verification mechanism for power frequency component changes, misjudgments are easily made. Summary of the Invention

[0005] The purpose of this invention is to provide a method for monitoring the energized installation status of surge arresters, thereby addressing some of the drawbacks and shortcomings pointed out in the background art.

[0006] The present invention adopts the following technical solution to solve the above-mentioned technical problems: a method for monitoring the live installation status of a surge arrester, comprising: setting up a broadband current measurement unit on the grounding down conductor of the surge arrester to obtain a current signal, and setting up a voltage reference measurement unit at the high-voltage end to obtain a voltage reference signal;

[0007] Before installation begins, current signals are collected in the first time window to obtain the background noise level and a pulse event threshold is set. The threshold is updated with the background noise. During installation, current signals are continuously monitored. When a pulse with an amplitude exceeding the threshold and a duration shorter than the set upper limit is detected, it is recorded as a candidate contact event.

[0008] The voltage reference signal within the second time window before and after the candidate contact event is subjected to stability verification. If the verification is successful, it is confirmed as the first electrical contact event and it is determined that an electrical contact loop has been formed. Otherwise, it is discarded. If the first electrical contact event is not confirmed within the observation period, the output is that no electrical contact loop has been formed.

[0009] Furthermore, the identification of the first electrical contact event includes phase gating discrimination, which determines the power frequency phase based on the voltage reference signal. Only when the candidate pulse occurs within a preset phase interval and its polarity is consistent with the direction of voltage change within the phase interval is it confirmed as the first electrical contact event; otherwise, it is discarded.

[0010] Furthermore, the wideband current measurement unit simultaneously outputs a pulse detection channel and a power frequency component channel; when the pulse detection channel detects a candidate pulse that meets the threshold and duration conditions, and the power frequency component channel does not exhibit an excessive step change within a preset time before and after the candidate pulse, the candidate pulse is input into the first electrical contact event identification and stability verification process.

[0011] Furthermore, the preset phase interval is set based on the zero crossover point of the voltage reference signal, and is set separately for the rising half-cycle and the falling half-cycle; only when the candidate pulse falls into the phase interval of the corresponding half-cycle and its polarity is consistent with the direction of voltage change of that half-cycle is it confirmed as the first electrical contact event.

[0012] Furthermore, zero-crossover tracking and validity judgment are adopted when determining the power frequency phase; when the voltage reference signal is distorted, missing or abruptly changed, making zero-crossover unreliable, the phase gating judgment is suspended and the candidate pulses during this period are judged as invalid candidate pulses.

[0013] Furthermore, a protection discrimination is set for the candidate pulses that pass through the phase gating; if a second pulse with opposite polarity that falls into the phase interval appears within the protection time, it is judged as contact jitter, and the candidate pulse pair is not regarded as the first electrical contact event; if the second pulse does not appear within the protection time, the first electrical contact event is confirmed.

[0014] Furthermore, the step change exceeding the tolerance is determined by a dual-window comparison. A comparison time window is set before and after the candidate pulse, and the power frequency component amplitude index is calculated. When the difference between the comparison results before and after exceeds the preset tolerance, it is judged as a step change. Only when it does not exceed the tolerance is the candidate pulse input into the first electrical contact event identification and stability verification process.

[0015] Furthermore, when the power frequency component channel experiences an excessively tolerant step change within a preset time before or after a candidate pulse, the candidate pulse is removed and a shielding period is triggered; during the shielding period, new candidate pulses are prohibited from entering the first electrical contact event identification and stability verification process, and the judgment is restored after the shielding ends.

[0016] Furthermore, the comparison time windows before and after the candidate pulse are aligned with the same phase point of the voltage reference signal, so that the two comparison time windows cover the same number of power frequency cycles; when an abnormal cycle of the voltage reference signal appears in either comparison time window, the abnormal cycle is removed before calculating the power frequency component amplitude index.

[0017] Furthermore, the abnormal cycle is determined by comparing the zero-crossing interval and peak amplitude of multiple adjacent power frequency cycles. When the deviation of the zero-crossing interval or peak amplitude of any power frequency cycle from the corresponding parameters of the previous and subsequent power frequency cycles exceeds a preset consistency threshold, the power frequency cycle is determined to be an abnormal cycle and is removed.

[0018] The beneficial effects of this invention are as follows: By deploying a broadband current measurement unit on the grounding lead of the surge arrester and setting a voltage reference measurement unit at the high-voltage end, synchronous acquisition and joint analysis of current signals and voltage reference signals can be achieved during installation. By establishing a background noise level and dynamically setting a pulse event threshold before installation begins, the system can adaptively adjust the detection threshold according to the on-site electromagnetic environment, thereby accurately extracting transient current signals generated by installation contact in complex operating environments. During installation, by continuously monitoring the current signal and identifying pulse signals with amplitudes exceeding the threshold and durations meeting the conditions, candidate contact events are formed. The stability of the voltage reference signal before and after the candidate events is combined to reliably confirm the first electrical contact event, thus enabling real-time determination of whether the surge arrester has formed an electrical contact loop. This method achieves online monitoring without changing the existing equipment structure, and has the advantages of simple implementation, strong adaptability to the on-site environment, and a high degree of automation in the judgment process.

[0019] A multi-layered discrimination mechanism, incorporating phase-gated discrimination, power frequency component channel-assisted judgment, and dual-window comparative analysis, is introduced to progressively screen and verify candidate pulses. By using power frequency phase information based on the voltage reference signal to determine the consistency of pulse phase and polarity, and combining this with zero-crossover tracking effectiveness detection and contact jitter protection discrimination mechanisms, misjudgments caused by voltage anomalies or contact instability can be effectively avoided. Furthermore, by identifying step changes in the power frequency component channel, setting shielding periods, and employing a phase-aligned dual-window comparison method, and by eliminating abnormal periods in statistical calculations, the influence of system operation disturbances and power grid fluctuations on the detection results can be effectively eliminated, thereby significantly improving the accuracy and stability of initial electrical contact event identification. Attached Figure Description

[0020] Figure 1 This is a functional relationship diagram of the surge arrester live installation status monitoring system of the present invention.

[0021] Figure 2 This is a schematic diagram of the background data acquisition results and adaptive threshold setting before installation in Embodiment 1 of the present invention.

[0022] Figure 3 This is a schematic diagram of voltage reference signal phase gating and candidate pulse phase determination in Embodiment 1 of the present invention.

[0023] Figure 4 This is a schematic diagram of the reverse second pulse investigation and first electrical contact event confirmation during the protection time in Embodiment 1 of the present invention.

[0024] Figure 5 This is a schematic diagram of the dual-window comparison and over-tolerance step change elimination of candidate pulse 1 in Embodiment 2 of the present invention.

[0025] Figure 6 This is a schematic diagram of the shielding period and recovery determination after candidate pulse 1 is eliminated in Embodiment 2 of the present invention.

[0026] Figure 7 This is a schematic diagram of the abnormal period identification, elimination, and retention determination of candidate pulse 2 in Embodiment 2 of the present invention. Detailed Implementation

[0027] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0028] Combined with appendix Figure 1 This invention discloses a method for monitoring the energized installation status of a surge arrester. A broadband current measurement unit is installed outside the grounding down conductor of the surge arrester and senses and measures the current changes in the grounding down conductor through a non-contact or clamp-on structure. This measurement unit has broadband response characteristics, enabling it to simultaneously capture the power frequency component and transient pulse currents generated during installation contact, thereby achieving effective detection of minute transient current changes. When the surge arrester is energized and the installation operation gradually approaches conduction, the installation tool or connecting component makes its first electrical contact with the conductive part of the surge arrester. At this time, a short-term transient current change is generated in the grounding loop. The broadband current measurement unit can quickly respond to this transient current and output a corresponding current signal. The current signal is then sent to a monitoring device for acquisition and processing, used for subsequent pulse event identification and installation status determination.

[0029] A voltage reference measurement unit is installed at the high-voltage end of the surge arrester to acquire a voltage reference signal that changes synchronously with the grid voltage. The voltage reference measurement unit can establish a measurement relationship with the high-voltage end potential through a voltage sampling probe or a capacitive voltage divider structure, thereby outputting a reference voltage signal consistent with system voltage changes. This voltage reference signal is mainly used to reflect the current power frequency phase change and voltage stability of the grid. By continuously acquiring the voltage reference signal, voltage zero-crossing information and voltage periodic change characteristics can be obtained, providing a basis for phase discrimination and stability verification of subsequent current pulse events. A broadband current measurement unit and the voltage reference measurement unit simultaneously acquire signals, forming a synchronous monitoring system for current and voltage reference signals. The current signal reflects the transient current change characteristics in the grounding loop, while the voltage reference signal provides reference information for system voltage phase and operational stability.

[0030] Since the surge arrester has not yet made contact during installation, the current changes in the grounding down conductor mainly originate from environmental electromagnetic interference, equipment background noise, and minor fluctuations caused by normal system operation. Therefore, by statistically analyzing the current signal collected within the first time window, the background noise level under the current field conditions can be obtained. The background noise level can be obtained by calculating the amplitude distribution characteristics or equivalent amplitude index of the current signal, and this serves as the benchmark for subsequent pulse event identification. A pulse event threshold is set based on the background noise level, ensuring that the threshold value is higher than the background noise fluctuation range, thereby avoiding misidentification of environmental noise as a contact event. As the installation process continues, environmental interference or system operating status changes. Therefore, the current signal is continuously updated and statistically analyzed during monitoring, and the pulse event threshold is dynamically adjusted according to the new background noise level to maintain a balance between detection sensitivity and anti-interference capability.

[0031] When a transient change in the current signal with an amplitude exceeding the pulse event threshold is detected, the duration of this change is determined. Since actual electrical contact behavior typically manifests as short transient pulses rather than long-term stable current changes, a duration upper limit is set for this pulse signal. When the detected current pulse duration is less than the set upper limit, it is considered a pulse signal meeting transient characteristics, and the occurrence time, amplitude characteristics, and duration parameters of the event are further recorded. Pulse signals that meet both the amplitude and duration requirements are marked as candidate contact events. Candidate contact events are not directly determined as actual contact behavior but are used as input for subsequent discrimination processes.

[0032] Using the occurrence time of the candidate contact event as a reference, second time windows are set before and after it, and the voltage reference signal output by the voltage reference measurement unit is extracted within this time range. By analyzing the continuous changes of the voltage reference signal within the second time window, it can be determined whether the system voltage is in a stable operating state during the occurrence of the candidate contact event. If the voltage reference signal maintains normal periodic variation characteristics within this time range, without obvious abrupt changes, missing values, or abnormal fluctuations, the system voltage is considered to be in a stable state, and the candidate contact event is not affected by abnormal voltage factors. At this time, the candidate contact event is confirmed as the first electrical contact event in the installation process, and it is determined that the surge arrester installation circuit has formed an electrical contact path.

[0033] If a significant anomaly is detected in the voltage reference signal within the second time window, such as periodic instability, waveform distortion, or abrupt amplitude changes, it indicates that the system voltage environment was unstable during the candidate event, and the transient current pulse was caused by grid disturbance or external electromagnetic interference. In this case, the candidate contact event is deemed invalid and discarded, and is not used as a basis for the formation of electrical contact.

[0034] Throughout the installation monitoring process, the system continuously records candidate contact events and performs the aforementioned stability checks, while setting an observation period to limit the monitoring time range. If at least one candidate contact event passes the stability check within the observation period, it is confirmed that the first electrical contact event has occurred during the installation process, and the electrical contact loop is determined to have been formed. If no candidate contact event passes the check before the end of the observation period, it is considered that a valid electrical contact loop has not yet been formed during the current installation process.

[0035] The voltage reference signal acquired by the voltage reference measurement unit is used as the phase reference for the system's operating state. The current power frequency phase information of the power grid is determined by continuously sampling and analyzing the voltage reference signal. The voltage reference signal exhibits periodic variation characteristics, and its waveform undergoes a rising process and a falling process sequentially within each cycle. Therefore, the power frequency phase position corresponding to each time point can be determined by identifying the periodic pattern of voltage changes.

[0036] When identifying candidate contact events, the occurrence time of the candidate pulse is analyzed in relation to the power frequency phase, and the event is phase-gated according to a pre-defined phase interval. The pre-defined phase interval limits the allowed phase range for valid contact pulses; this range is set based on the characteristics of the relationship between transient current and voltage changes during arrester contact. When a candidate pulse occurs within the pre-defined phase interval, the relationship between the pulse polarity and the direction of voltage change within that phase interval is further determined. If the polarity of the candidate pulse is consistent with the direction of voltage change within the phase interval, it indicates that the transient current change and the system voltage change have a consistent physical correlation, conforming to the typical characteristics of current response during electrical contact formation.

[0037] If the phase position of a candidate pulse is not within the preset phase interval, or if its polarity direction is inconsistent with the voltage change direction within the corresponding phase interval, it indicates that the pulse originates from external electromagnetic interference or other non-contact factors. Candidate pulses that do not meet the phase gating conditions are deemed invalid events and discarded, and are not included in the confirmation process for the first electrical contact event.

[0038] The voltage reference signal passes through zero-crossing points (changing from negative to positive and from positive to negative) in each power frequency cycle. Therefore, the zero-crossing points can be used to divide a power frequency cycle into a rising half-cycle and a falling half-cycle. After determining the zero-crossing point, using this zero-crossing moment as the starting reference, corresponding phase intervals are set for the rising and falling half-cycles according to a pre-set phase range, thus forming two independent phase-gated regions. When a candidate pulse occurs, its power frequency phase position is determined, and it is determined whether it belongs to the rising or falling half-cycle. Then, the phase of the pulse is compared with the preset phase interval of the corresponding half-cycle. Only when the candidate pulse falls within the preset phase interval of the corresponding half-cycle, and the polarity direction of the pulse is consistent with the voltage change direction of the half-cycle, is the pulse considered to meet the physical characteristics of transient current generated by electrical contact, and thus it is confirmed as the first electrical contact event. If a candidate pulse does not fall within the phase interval of the corresponding half-cycle, or if its polarity direction is inconsistent with the voltage change direction, then the pulse is considered not to meet the phase consistency condition and is therefore rejected and not included in the confirmation of the first electrical contact event.

[0039] In determining the power frequency phase, the system continuously monitors the voltage reference signal and uses a zero-crossing tracking method to identify the zero-crossing moment in each power frequency cycle in real time. By continuously recording the time interval between adjacent zero-crossing points, stable power frequency cycle information can be obtained, thereby achieving continuous tracking of the voltage signal phase position. Simultaneously, the validity of the zero-crossing identification is judged during the process. When the voltage reference signal exhibits waveform distortion, signal loss, or abnormal amplitude abrupt changes, the zero-crossing point identification becomes unreliable, affecting the accuracy of phase judgment. When such abnormalities are detected, the system pauses the phase gating discrimination process and does not perform phase consistency judgment on candidate pulses during this period. Instead, all candidate pulses appearing within this time period are uniformly judged as invalid candidate pulses to avoid misjudgments due to unreliable phase references. Once the voltage reference signal returns to normal periodic variation and stable zero-crossing information is obtained again, the phase gating discrimination function is resumed.

[0040] After the candidate pulse is phase-gated and judged, a protection time is set, and the current signal changes are monitored within this protection time. If a new pulse signal is detected again within the protection time, and this pulse also falls within the preset phase interval but its polarity is opposite to that of the previous pulse, it indicates a contact jitter phenomenon where the contact point momentarily separates and then re-contacts. In this case, the aforementioned candidate pulse and the subsequent pulse are jointly determined as contact jitter events, and are not confirmed as the first electrical contact event, thereby avoiding misjudging unstable contact as valid contact. If no second pulse meeting the above conditions is detected within the protection time, it indicates that the contact process is stable and there is no obvious jitter. At this time, the candidate pulse can be finally confirmed as the first electrical contact event, and it can be determined that the surge arrester installation circuit has formed a stable electrical contact.

[0041] The pulse detection channel is used to extract transient components from the current signal. This channel sensitively detects short-time current pulses through wideband response or high-frequency enhancement processing, thus enabling timely capture of transient current signals generated at the moment of installation contact. The power frequency component channel is used to extract power frequency variation characteristics from the current signal. This channel suppresses high-frequency transient components through filtering or frequency separation, retaining only power frequency current variations related to grid operation, which are used to reflect the steady-state current state in the grounding loop.

[0042] Using the occurrence time of the candidate pulse as a reference, preset time ranges are set before and after it, and the current signal output by the power frequency component channel is continuously analyzed within these time ranges. If no obvious current step change is detected within these time ranges, and the power frequency component maintains a stable periodic variation characteristic, then the candidate pulse is considered not to be accompanied by a sudden change in the loop current structure, which is consistent with the typical behavior of contact transient current. In this case, the candidate pulse is input as a valid candidate event into the initial electrical contact event identification and voltage stability verification process.

[0043] If a significant step change is detected in the power frequency component channel within a preset time range before and after the candidate pulse, it indicates a substantial change in the current state in the grounding loop. This change may originate from current disturbances caused by system operation state switching or other equipment operation. In this case, the pulse signal is not directly related to the surge arrester installation contact, and therefore it is not input into the subsequent identification process to avoid the influence of system operation disturbances on the contact determination.

[0044] Using the occurrence time of the candidate pulse as a benchmark, comparison time windows are set before and after it. The power frequency component amplitude index is calculated by statistically analyzing the power frequency component signal within the two time windows. The amplitude index reflects the overall change level of the power frequency current within the corresponding time period, and stable amplitude characteristics can be obtained by periodically statistically analyzing the current signal.

[0045] When the difference in amplitude calculated within the time window before and after a candidate pulse exceeds the preset tolerance, it indicates a significant change in power frequency current near the candidate pulse. This change manifests as a sudden or step change in current amplitude, indicating a significant alteration in the grounding loop current state. In this case, the change is more likely to originate from system operation adjustments or external equipment operation, rather than transient currents generated by surge arrester installation contact behavior. Therefore, when the amplitude difference is detected to exceed the preset tolerance, this situation is judged as an out-of-tolerance step change, and the corresponding candidate pulse is discarded, no longer entering the initial electrical contact event identification and voltage stability verification process. If the difference in amplitude within the time window before and after a candidate pulse does not exceed the preset tolerance, the power frequency component is considered to remain stable, indicating that the pulse is not accompanied by a significant change in loop current, thus having a higher probability of originating from contact transient current. In this case, the candidate pulse is input into the initial electrical contact event identification and stability verification process for further judgment.

[0046] When the power frequency component channel detects an excessively volatile step change within a preset time range before and after a candidate pulse, the system not only eliminates the current candidate pulse but also triggers a shielding period to suppress consecutive false positives within a short period. The shielding period begins immediately after the step change is identified. During this period, the system continues to acquire current signals but temporarily prohibits new candidate pulses from entering the initial electrical contact event identification and stability verification process. The purpose of setting the shielding period is to prevent continuous interference pulses caused by current fluctuations or equipment recovery processes from affecting the judgment results within a short period after a sudden change in the system current state. As the shielding period ends, the system resumes the normal candidate pulse discrimination process and allows new candidate pulses to re-enter the initial electrical contact event identification and stability verification process.

[0047] A continuous voltage reference signal is acquired through a voltage reference measurement unit, and the current power frequency cycle characteristics of the power grid are determined based on this signal. On this basis, a stable phase reference point is determined by identifying the periodic variation characteristics of the voltage reference signal, and this phase reference point is used as the time alignment benchmark. Subsequently, comparison time windows are set before and after the occurrence of candidate pulses, with the starting positions of the two time windows corresponding to the same phase point of the voltage reference signal. In this way, the two comparison time windows are kept in phase, ensuring that both time windows cover the same number of power frequency cycles. Since the power frequency current signal has different instantaneous amplitude characteristics at different phase positions, if the phase positions of the two time windows are inconsistent, it will lead to deviations in the amplitude statistics results.

[0048] In real-world operating environments, voltage reference signals are affected by grid fluctuations, transient interference, or measurement noise, resulting in abnormal changes in certain power frequency cycles. Directly including these abnormal cycles in statistical calculations causes the amplitude index to deviate from the true current state, thus affecting the accuracy of step change determination. Therefore, before calculating the power frequency component amplitude index, a cycle-by-cycle analysis of the voltage reference signal within the comparison time window is performed. When a significant abnormal change in the voltage reference signal is detected in a certain cycle, that cycle is marked as an abnormal cycle and removed from the statistical range. After the abnormal cycles are removed, the power frequency component amplitude index is recalculated using the remaining valid cycle data to obtain more stable and reliable current amplitude characteristics.

[0049] A continuous voltage reference signal is acquired using a voltage reference measurement unit, and the start position of each power frequency cycle is determined using a zero-crossing identification method. Within each cycle, the voltage reference signal exhibits zero-crossing points that change from negative to positive and from positive to negative. By continuously recording the time interval between adjacent zero-crossing points, the cycle length of each power frequency cycle can be determined. Simultaneously, the peak amplitude of the voltage reference signal is extracted within each cycle range to obtain the amplitude characteristics of the corresponding cycle.

[0050] The zero-crossing interval of the current cycle is compared with that of the adjacent cycle, and the peak amplitude of the current cycle is also compared with that of the adjacent cycle. If there is a significant difference between the current cycle and the adjacent cycle in terms of cycle length or peak amplitude, it indicates that the voltage signal of that cycle is affected by a disturbance or measurement anomaly. To quantify the degree of this difference, a consistency threshold is preset in the system to limit the allowable range of variation between normal cycles. When the deviation of the zero-crossing interval or peak amplitude of a cycle relative to its adjacent cycle exceeds the consistency threshold, the cycle is considered inconsistent with the characteristics of a normal power frequency cycle, and thus the cycle is judged as an abnormal cycle.

[0051] Once an abnormal period is identified, it is removed from the statistical data when calculating the power frequency component amplitude index, and only the remaining normal period data is used for amplitude index calculation. By identifying and removing abnormal periods, the influence of local voltage fluctuations, transient interference, or measurement noise on the statistical results can be avoided, thus ensuring that the calculated power frequency component amplitude index within the comparison time window can truly reflect the stable state of the system current.

[0052] Example 1:

[0053] A 220kV substation underwent live-line installation of zinc oxide surge arresters under uninterrupted power conditions. The work was scheduled from 9:12 AM to 9:18 AM, with cloudy skies, an ambient temperature of approximately 26°C, and a wind speed of approximately 2.1 m / s. Stable power frequency electromagnetic interference was generated by the busbar and nearby energized equipment. The installation personnel first hoisted and positioned the surge arrester body, then gradually approached the high-voltage connection point with the connecting tools, while ensuring reliable connection of the measurement circuit at the grounding down conductor. During this process, the monitoring system first collected background noise before installation began, and then continuously collected the grounding down conductor current signal during the installation. Combined with the high-voltage reference signal, candidate pulses were identified, phase-gated, and their validity checked to determine whether an initial electrical contact event had occurred and an electrical contact circuit had been formed.

[0054] A broadband current measurement unit is installed at the grounding lead of the surge arrester to collect the pulse current signal triggered by the moment of contact establishment during installation. The sampling rate of the broadband current measurement unit is set to 5MHz, which can distinguish the pulse duration in microseconds. A voltage reference measurement unit is set at the high-voltage end of the surge arrester to extract the power frequency reference waveform and implement zero-crossing tracking. In this embodiment, the power grid frequency is 50Hz, corresponding to a power frequency period of 20ms. The monitoring host opens the first time window for background data acquisition before the installation begins. After the installation begins, it continuously monitors the current signal. Once a pulse with an amplitude exceeding the adaptive threshold and a duration shorter than the set upper limit is detected, it is recorded as a candidate contact event. Then, the voltage reference signal is called to perform stability verification and phase gating discrimination. To avoid misjudging contact bounce as the first electrical contact, a protection time is set to perform a reverse second pulse check on the candidate pulses that have passed the phase gating.

[0055] Before installation, the monitoring system set a first time window of 120ms between 9:12:00.000s and 9:12:00.120s to continuously collect the grounding lead current signal to characterize the background noise level. The collection results showed that, under the combined effects of power frequency interference and slight wind-induced swaying, the equivalent noise mean was 3.8mA, and the standard deviation was 4.6mA. Based on the background statistical results, the following adaptive pulse event threshold expression was used to establish the candidate event triggering threshold:

[0056]

[0057] In the formula, For pulse event threshold, This represents the mean of the background noise. The standard deviation of the background noise. This is the threshold coefficient. Take... Substituting the measured data, we get:

[0058] mA

[0059] Therefore, 22.2mA is set as the amplitude discrimination threshold for candidate pulses during installation, and dynamic adjustment capability is retained along with the background update strategy. At the same time, the upper limit of the candidate pulse duration is set to 300μs to exclude non-contact disturbances with longer durations. Figure 2 The background curve in the data is generally below the threshold, indicating that the data collected in the first time window is mainly the background disturbance before installation rather than the actual contact pulse. Therefore, the data corresponding to this time window can be used as the threshold basis for subsequent candidate contact event identification.

[0060] At 9:13:24.286s during installation, the installer moved the connection tool further towards the high-voltage connection position. The gap between the tool head and the target contact point rapidly narrowed, and a significant transient pulse was detected at the grounding lead. This pulse had a peak value of 68.5mA, a duration of 96μs, and positive polarity. Its peak value was significantly higher than the 22.2mA threshold, and its duration was shorter than the 300μs upper limit, therefore it was initially recorded as a candidate contact event. The monitoring system then extracted the voltage reference signal within the second time window before and after the pulse, checking for waveform continuity, abrupt distortion, and reliable zero-crossing tracking. Because a brief voltage reference waveform distortion occurred approximately 18ms before the candidate pulse, the system first performed zero-crossing validity judgment on this distorted segment and paused phase gating during the distortion period, not considering any pulses within that period as valid candidates. After the voltage reference stabilized, the zero-crossing information corresponding to the 68.5mA candidate pulse was reconfirmed. Figure 3 The diagram shows a single-cycle waveform of the voltage reference signal used to illustrate phase gating determination, the position of the most recent effective zero crossover, the time of the candidate pulse occurrence, and the point of local waveform distortion. It can be seen intuitively that the system does not perform phase discrimination uniformly on all voltage reference segments, but first removes distorted segments, and then performs subsequent phase calculation and gating determination on the basis of the effective zero crossover after the stability is restored.

[0061] The zero-crossing point is used as the power frequency phase reference, and phase gating intervals are set for the rising and falling half-cycles respectively. For the rising half-cycle, the preset effective phase interval is 30° to 75°, and the polarity of the candidate pulse is required to be consistent with the direction of voltage change within this half-cycle. That is, during the voltage change from low to high, only positive polarity candidate pulses are accepted. For the falling half-cycle, a corresponding interval can be set separately, but this branch does not need to be triggered. The system uses continuous zero-crossing tracking to determine the power frequency phase. When the interval between adjacent zero-crossing points deviates from the 10ms half-cycle reference by more than 8%, or when the peak value of the reference waveform changes by more than 12% between adjacent half-cycles, the zero-crossing is deemed unreliable, phase gating is suspended, and pulses detected during this period are considered invalid candidates. A local distortion of the reference waveform occurred near 9:13:24.268s, with the interval between adjacent half-cycle zero-crossing points reaching 10.96ms, a deviation of 9.6%, exceeding the 8% criterion. Therefore, this period was suspended. After 9:13:24.280s, the reference waveform stabilized. The zero-crossing intervals for the three consecutive half-cycles were 10.03ms, 9.98ms and 10.01ms, respectively, with peak fluctuations controlled within 4%, meeting the zero-crossing validity requirement. Therefore, the phase calculation of the 68.5mA candidate pulse was performed using the recovered valid zero-crossing data. Figure 3 The explanatory information also indicates that the zero-crossing interval of the local distortion half-cycle is 10.96ms, which is about 9.6% of the deviation from the 10ms half-cycle reference, exceeding the 8% criterion. Therefore, this segment of the reference waveform is not used for candidate pulse phase confirmation. The zero-crossing intervals of the three consecutive half-cycles after stabilization are 10.03ms, 9.98ms and 10.01ms, respectively, with peak fluctuations controlled within 4%, so they can be used as effective phase references.

[0062] After confirming the validity of the zero-crossing, the system performs phase gating on the candidate pulses. Using the most recent valid zero-crossing moment as the phase starting point, the candidate pulse is calculated to be within the effective range of the rising half-cycle, and its polarity is positive, consistent with the voltage rise direction of that half-cycle; therefore, it passes phase gating. Subsequently, the system sets a protection time of 1.5 ms to determine if a second pulse with opposite polarity caused by contact bounce exists. Monitoring results show that within 1.5 ms after 9:13:24.286s, only two weak disturbances with amplitudes of 8.4mA and 11.1mA were detected, neither exceeding the 22.2mA threshold. No reverse valid pulses falling within the corresponding phase range were observed, and no second pulse with opposite polarity to the first candidate pulse appeared. Therefore, this candidate pulse is determined not to be a pulse pair caused by contact bounce and is not considered contact bounce; it should be confirmed as the first electrical contact event. This process is... Figure 4The diagram provides a direct corresponding representation, showing the protection time range, the primary candidate pulse, and the time and amplitude relationship between the two weak disturbances. With a protection time of 1.5 ms, only two weak disturbances of 8.4 mA and 11.1 mA were detected within this timeframe, both of which were below the threshold. The pulse amplitude is mA, therefore it cannot constitute a second valid pulse in reverse corresponding to the confirmed candidate pulse. This eliminates the possibility of misjudgment due to contact bounce or contact jitter, and further supports the identification of this 68.5mA pulse as the first electrical contact event.

[0063] The candidate pulse phase is calculated using the following expression:

[0064]

[0065] In the formula, The candidate pulse phase angle, The time when the candidate pulse occurs. The most recent valid zero-crossing time. This is the power frequency cycle. The power frequency is 50Hz, therefore:

[0066] ms

[0067] The recorded time difference between the candidate pulse occurrence time and the most recent valid zero-crossing time is 2.9 ms. Substituting the data, we can obtain:

[0068]

[0069] Calculation results show that the candidate pulse has a phase angle of 52.2°, falling within the preset phase range of 30° to 75° during the rising half-cycle. Simultaneously, the pulse peak value of 68.5mA is greater than the threshold of 22.2mA, the duration of 96μs is less than the upper limit of 300μs, the polarity is positive and consistent with the direction of voltage change during the rising half-cycle, and the corresponding voltage reference signal has zero-crossing continuous validity. No reverse second pulse meeting the conditions appeared within the 1.5ms protection time. Furthermore, within the 500ms observation period, the system did not detect any other valid contact events preceding this event. Therefore, this pulse can be considered the first electrical contact event during the installation process.

[0070] Example 2:

[0071] Based on Example 1, the live-line installation of the zinc oxide surge arrester continued under the same conditions. The work time was set from 10:08 AM to 10:15 AM, with an ambient temperature of 27°C, relative humidity of 61%, and wind speed of approximately 2.4 m / s. A stable power frequency electromagnetic background was generated by the substation busbar and adjacent live equipment. The installers completed the hoisting and positioning of the surge arrester and made minor adjustments to the grounding down conductor connection position. Due to slight mechanical swaying of the connecting components during the initial adjustment, a short-duration pulse appeared in the current signal at the grounding down conductor. However, the power frequency component channel experienced a synchronous abrupt change before and after the pulse, so it was not used as a basis for establishing a real contact. Subsequently, the installers recalibrated the connection position and slowly approached the connection point again, generating a second candidate pulse. After dual-channel joint discrimination, dual-window comparison, and abnormal cycle elimination, the system confirmed that the second candidate pulse could enter the subsequent first electrical contact event identification and stability verification process, and finally output that an electrical contact circuit had been formed.

[0072] A wideband current measurement unit is installed at the grounding down conductor of the surge arrester. This unit simultaneously outputs a pulse detection channel and a power frequency component channel. The pulse detection channel is used to capture short-duration transient contact pulses, with a sampling rate set to 5MHz, capable of resolving microsecond-level pulse details. The power frequency component channel is used to extract the power frequency amplitude variation characteristics in the grounding down conductor current, with an update period set to 1ms, for evaluating the power frequency stability state before and after the candidate pulse. A voltage reference measurement unit is installed at the high-voltage end of the surge arrester to provide a common phase point alignment reference for the comparison time windows before and after. Before installation, the monitoring system collects background noise and sets a candidate pulse threshold. During installation, it continuously monitors the output of the pulse detection channel. When the pulse detection channel detects a pulse with an amplitude exceeding the threshold and a duration shorter than the set upper limit, the system does not directly confirm it as a contact event. Instead, it further retrieves data from the power frequency component channel, constructs comparison time windows before and after the candidate pulse, and compares whether there is an over-tolerance step change in the power frequency component amplitude index within the two windows, thereby improving the reliability of real contact identification.

[0073] Before installation, the system completed background data acquisition within a 100ms time window, setting the candidate pulse threshold to 24.0mA and the maximum candidate pulse duration to 300μs. At 10:10:12.436s, the installer fine-tuned the angle of the grounding lead connection clamp, releasing the mechanical stress between the clamp and the metal connector. The pulse detection channel detected the first candidate pulse, with a peak value of 95.0mA and a duration of 140μs, meeting the amplitude and duration requirements, and was therefore recorded as candidate pulse 1. Subsequently, the system extracted comparison time windows of 60ms each before and after the occurrence of this candidate pulse, with each time window covering three complete power frequency cycles. To evaluate the stability of the power frequency components before and after the pulse, the system used the power frequency component amplitude index.

[0074]

[0075] In the formula, To compare the amplitude of the power frequency component within the time window, The power frequency component channel within this time window Each sample value, This represents the number of samples. The power frequency component channel update period is 1ms, therefore each 60ms time window corresponds to... For candidate pulse 1, the system statistically analyzed the data within the window and found that the amplitude index of the front window was 18.2mA and the amplitude index of the rear window was 26.7mA, indicating that the adjustment action by the installer caused a significant increase in the power frequency component. This change is not typical of the stable state before and after contact establishment, but is more consistent with the characteristics of loop current distribution changes caused by mechanical disturbance. Figure 5 The results shown are consistent with this. The figure directly reflects that the amplitude index of the two windows before and after the pulse increased from 18.2mA to 26.7mA. The peak value of candidate pulse 1 itself is 95.0mA and the duration is 140μs. This indicates that although the pulse meets the pulse detection conditions, the state of its power frequency components before and after the pulse is not stable.

[0076] To further determine whether candidate pulse 1 should proceed to the next step, the system calculates the difference between the two windows before and after:

[0077] Δ

[0078] In the formula, Δ This represents the difference in the amplitude of the power frequency component between the preceding and following time windows. For the front window amplitude index, This is the amplitude index for the back window. Substituting the measured data of candidate pulse 1, we get:

[0079] Δ mA

[0080] The over-tolerance threshold for the power frequency component is set to 5.0mA. Since 8.5mA is significantly greater than 5.0mA, it can be determined that there is an over-tolerance step change before and after candidate pulse 1. Therefore, this candidate pulse is not entered into the first electrical contact event identification and stability verification process, but is directly rejected. This demonstrates the necessity of the dual-channel approach: the pulse detection channel is responsible for detecting transient events, while the power frequency component channel is responsible for determining whether the system state is stable before and after the event, avoiding misjudging mechanical disturbances as actual contact. Figure 5 The corresponding calculation results are as follows: the update period for the power frequency component channel is 1ms, and the comparison time window is 60ms for each time window. Therefore, the number of samples corresponding to each time window is... And by mA, mA yields Δ =8.5mA. This result indicates that there is a significant step change before and after candidate pulse 1 occurs. Therefore, the reason for its rejection is not due to insufficient pulse amplitude or duration, but rather because the power frequency component channel detected a sudden change in the loop state caused by mechanical disturbance.

[0081] After candidate pulse 1 was rejected due to an excessively large step change, the system immediately triggered a shielding period to prevent new disturbance pulses from entering the subsequent judgment while the installed components were still in the rebound or displacement recovery phase. The shielding period was set to 60ms, from 10:10:12.436s to 10:10:12.496s. During this shielding period, although the pulse detection channel captured two short pulses with amplitudes of 31.4mA and 28.7mA respectively, the system did not send them to the initial electrical contact event identification and stability verification process according to the shielding rules. After the shielding ended, the power frequency component channel monitoring results showed that the amplitude fluctuations converged back to the stable range, the installers completed the recalibration of the connection positions, and the system resumed normal judgment of new candidate pulses. Figure 6 Correspondingly, it shows that a shielding period is triggered immediately after candidate pulse 1 is rejected. Even though two short pulses of 31.4mA and 28.7mA appear during this 60ms shielding period, the system does not send them into the subsequent process. The shielding period is used to isolate subsequent disturbances generated during the springback or displacement recovery phase of the installed component, preventing pulses on the same mechanical disturbance chain corresponding to candidate pulse 1 from being mistakenly sent into the subsequent process. It also shows the process by which the system resumes normal judgment of candidate pulse 2 after the shielding ends.

[0082] At 10:10:12.584s, the installers moved the connection points closer together again at a slower speed. The pulse detection channel detected a second candidate pulse with a peak value of 88.0mA and a duration of 110μs. The amplitude was above the 24.0mA threshold, and the duration was below the 300μs upper limit; therefore, it was recorded as candidate pulse 2. Unlike the first candidate pulse, the power frequency component channel did not show a significant overall rise or fall, indicating that the field circuit state tended to stabilize after the connection position was recalibrated. To avoid comparison deviations caused by different power frequency phases, the system set the comparison time windows before and after candidate pulse 2 according to the same phase point alignment of the voltage reference signal, ensuring that both windows covered the same number of power frequency cycles. The comparison windows before and after were still 60ms each, and were extracted from the same phase starting point of the voltage reference signal, each covering 3 power frequency cycles, thus ensuring the comparability of the amplitude indicators before and after. Figure 7 The comparison between the first, second, and third cycles within the candidate pulse 2 pre-window is shown accordingly. The second cycle deviates significantly from the normal reference cycle in both the zero crossover interval and the peak amplitude, and is therefore identified as an abnormal cycle.

[0083] During the initial statistical analysis of candidate pulse 2 across two windows, the system detected an anomaly in the second power frequency cycle of the preceding window. The zero-crossing interval of this cycle was 21.3 ms, while the reference value for the adjacent normal cycle was 20.0 ms; the peak amplitude of this cycle was 1.16 kV, while the reference peak amplitude for the adjacent normal cycle was 1.00 kV. To determine whether this cycle should be removed, the system used the abnormal cycle consistency deviation index:

[0084]

[0085] In the formula, For the first Consistency deviation index for each power frequency cycle For this period, there is a zero crossover interval. The reference zero crossover interval is for adjacent normal cycles. This represents the peak amplitude of the cycle. This represents the reference peak amplitude for adjacent normal cycles. Substituting the data, we get:

[0086]

[0087]

[0088]

[0089] Set a consistency threshold The value is 0.10. Since 0.16 is greater than 0.10, this period is determined to be an abnormal period and is removed from the calculation of the front window corresponding to candidate pulse 2. This processing can eliminate the amplification effect of local reference waveform fluctuations on the power frequency component amplitude index, making the comparison before and after more realistically reflect the power frequency state before and after the occurrence of the candidate pulse, thereby improving the reliability of the judgment. Figure 7 The corresponding calculation results also show that the zero-crossing interval of the second power frequency cycle in the candidate pulse 2 front window is 21.3 ms, the reference value is 20.0 ms, the peak amplitude is 1.16 kV, and the reference value is 1.00 kV. Based on the above consistency deviation indices, the time deviation term is 0.065, and the amplitude deviation term is 0.16. Therefore... greater than the threshold Therefore, the second power frequency cycle of the front window can be determined as an abnormal cycle and removed.

[0090] Before eliminating abnormal cycles, the initial calculated amplitude index for candidate pulse 2 in the front window was 22.4 mA, and the initial calculated amplitude index in the back window was 21.6 mA. Although the difference is not significant, the system does not directly adopt this result considering the presence of abnormal cycles in the front window. After eliminating abnormal cycles, the system recalculates the power frequency component amplitude index for the remaining normal cycles to obtain the front window amplitude index. 19.1mA, rear window The value is 21.6 mA. Based on the aforementioned double-window difference criterion, the recalculated value is:

[0091] Δ mA

[0092] Since 2.5mA is less than the preset tolerance of 5.0mA, it can be determined that no excessive step change occurred before or after candidate pulse 2. Therefore, candidate pulse 2 passed the dual-channel screening and was sent to the first electrical contact event identification and stability verification process. The system continued to track its subsequent state within a total observation period of 500ms and found no abnormal changes sufficient to overturn the result. The final output showed that an electrical contact loop had been formed.

Claims

1. A method for monitoring the live installation status of a surge arrester, characterized in that... include: A broadband current measurement unit is installed on the grounding lead of the surge arrester to obtain the current signal, and a voltage reference measurement unit is set up at the high-voltage end to obtain the voltage reference signal; Before installation begins, current signals are collected in the first time window to obtain the background noise level and a pulse event threshold is set. The threshold is updated with the background noise. During installation, current signals are continuously monitored. When a pulse with an amplitude exceeding the threshold and a duration shorter than the set upper limit is detected, it is recorded as a candidate contact event. The voltage reference signal in the second time window before and after the candidate contact event is checked for stability. If the check passes, it is confirmed as the first electrical contact event and it is determined that an electrical contact circuit has been formed. Otherwise, it is discarded. If the first electrical contact event is not confirmed within the observation period, then no electrical contact loop has been formed at the output. The first electrical contact event identification includes phase gating discrimination, which determines the power frequency phase based on the voltage reference signal. Only when a candidate pulse occurs within a preset phase interval and its polarity is consistent with the direction of voltage change within the phase interval is it confirmed as a first electrical contact event; otherwise, it is discarded. The wideband current measurement unit simultaneously outputs a pulse detection channel and a power frequency component channel. When the pulse detection channel detects a candidate pulse that meets the threshold and duration conditions, if no excessive step change occurs in the power frequency component channel within a preset time before and after the candidate pulse, the candidate pulse is input into the first electrical contact event identification and stability verification process. The super-tolerance step change is determined by dual-window comparison. A comparison time window is set before and after the candidate pulse and the power frequency component amplitude index is calculated. If the difference between the previous and subsequent comparison results exceeds the preset tolerance, it is judged as a step change; if it does not exceed the tolerance, the candidate pulse is input into the first electrical contact event identification and stability verification process.

2. The method for monitoring the energized installation status of a surge arrester according to claim 1, characterized in that, The preset phase interval is set based on the zero crossover point of the voltage reference signal, and is set separately for the rising half-cycle and the falling half-cycle; only when the candidate pulse falls into the phase interval of the corresponding half-cycle and its polarity is consistent with the direction of voltage change of that half-cycle is it confirmed as the first electrical contact event.

3. The method for monitoring the energized installation status of a surge arrester according to claim 1, characterized in that, When determining the power frequency phase, zero-crossover tracking is used and validity is judged; when the voltage reference signal is distorted, missing or abruptly changes, making zero-crossover unreliable, phase gating judgment is suspended and candidate pulses during this period are judged as invalid candidate pulses.

4. The method for monitoring the energized installation status of a surge arrester according to claim 1, characterized in that, A protection discrimination is set for candidate pulses that pass through the phase gate; If a second pulse with opposite polarity falls within the phase range during the protection time, it is judged as contact jitter, and the candidate pulse pair is not considered as the first electrical contact event. If the second pulse does not occur within the protection time, the first electrical contact event is confirmed.

5. The method for monitoring the energized installation status of a surge arrester according to claim 1, characterized in that, When the power frequency component channel experiences an over-tolerance step change within a preset time before or after a candidate pulse, the candidate pulse is removed and a shielding period is triggered. During the shielding period, new candidate pulses are prohibited from entering the initial electrical contact event identification and stability verification process. The judgment is restored after the shielding ends.

6. The method for monitoring the energized installation status of a surge arrester according to claim 1, characterized in that, The comparison time windows before and after the candidate pulse are set to be aligned with the same phase point of the voltage reference signal, so that the two comparison time windows cover the same number of power frequency cycles; when an abnormal cycle of the voltage reference signal appears in either comparison time window, the abnormal cycle is removed before calculating the power frequency component amplitude index.

7. The method for monitoring the energized installation status of a surge arrester according to claim 6, characterized in that, The abnormal cycle is determined by comparing the zero crossover interval and peak amplitude of multiple adjacent power frequency cycles. When the deviation of any cycle relative to the adjacent cycles exceeds a preset consistency threshold, the cycle is determined to be an abnormal cycle and is removed.

Images