A fault prediction method and device for a pressure-limiting arc-extinguishing lightning protection device

Abstract

The application provides a fault prediction method and device based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device, comprising: analyzing the uniformity of component stress distribution according to the stress distribution characteristics of each component, determining the local deformation damage area as the area whose stress distribution uniformity exceeds a preset threshold; extracting the maximum principal stress value and strain amplitude value of the local deformation damage area, calculating the fatigue damage accumulation value in combination with the stress history data of the disconnector's previous actions, and determining the fatigue damage degree of the component according to the fatigue damage accumulation value; analyzing the attenuation rate of the remaining reliable action times of the disconnector and the reduction amplitude of the success rate according to the ejection path resistance of the red indicator device in the disconnector, determining the resistance change trend according to the change rule of the resistance coefficient in the previous tests; determining the key wear position and failure mode according to the resistance change trend, and predicting the failure point of the disconnector through the degradation trajectory extrapolation method according to the determined wear rate and stress weak link.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a fault prediction method and device based on the timing of disconnection action of a voltage-limiting arc-extinguishing lightning protection device. Background Technology

[0002] Surge arresters, as key devices protecting equipment from lightning and overvoltage damage in power systems, are crucial for the stable operation of the power grid. Thermally detonating disconnectors, by promptly disconnecting the connection in the event of a surge arrester failure to prevent further damage or safety accidents, are core components ensuring system stability. However, existing disconnectors still have shortcomings in predicting reliability under complex operating conditions. Traditional methods often rely on static design parameters or single-condition testing, making it difficult to comprehensively capture the structural response characteristics of the disconnector under dynamic thermal shock, especially the rapid changes in component stress distribution and their long-term impact on operational reliability under transient shock waves. This leads to a lack of precise basis for maintenance cycle setting, increasing the risk of equipment failure or over-maintenance. Under thermal shock, the instantaneous stress distribution of each component of the disconnector becomes the primary technical factor affecting operational reliability. Due to the shock wave generated by the thermal detonation tube rupture, the disconnector is subjected to transient high stress, and uneven stress distribution among components may cause local deformation or fatigue damage. For example, the red indicator device may be obstructed by changes in resistance during its ejection path, leading to operational failure or delay. Therefore, how to analyze the instantaneous stress distribution of each component of the release device under thermal shock and its impact on the resistance of the red indicator device's ejection path, and establish a method for predicting the remaining reliable number of actions based on historical stress data, has become a key issue in formulating scientific maintenance and replacement cycles. Summary of the Invention

[0003] This invention provides a fault prediction method based on the disconnection timing of a voltage-limiting arc-extinguishing lightning protection device, mainly including:

[0004] The instantaneous stress distribution of each component of the release device under thermal explosion impact is obtained, the instantaneous stress distribution is processed by time domain filtering, and the shock wave generated by the explosion of the thermal explosion tube is obtained by collecting the pressure of the thermal explosion tube. The stress distribution characteristics of each component are obtained by combining the measured shock wave.

[0005] Based on the stress distribution characteristics of each component, the uniformity of stress distribution in the component is analyzed, and the area where the stress distribution uniformity exceeds the preset threshold is identified as the local deformation and damage area.

[0006] The maximum principal stress and strain amplitude of the local deformation and damage area are extracted, and the fatigue damage accumulation value is calculated by combining the stress history data of the release device during each operation. The degree of fatigue damage of the component is determined based on the fatigue damage accumulation value.

[0007] The remaining number of actions of the releaser mechanical structure is determined based on the degree of fatigue damage of the components, the remaining number of reliable actions of the releaser is obtained, the geometric dimensions of the red indicator device ejection channel are obtained, and the ejection path resistance of the red indicator device in the releaser is obtained by analyzing the obtained channel cross-section and surface roughness.

[0008] Based on the resistance analysis of the ejection path of the red indicator device in the release device, the attenuation rate and success rate of the remaining reliable action number of the release device are analyzed, and the trend of resistance change is determined based on the change law of the resistance coefficient in each test.

[0009] Based on the trend of resistance change, the key wear parts and failure modes are determined. Based on the determined wear rate and stress weak points, the failure point of the release device is predicted by the degradation trajectory extrapolation method.

[0010] The temperature of the conductive silicone rubber component at the fault point in the disconnector is acquired in real time to obtain the temperature rise trend of the thermal explosion tube. Based on the temperature rise trend of the thermal explosion tube, the response time delay and the degree of decrease in start-up success rate of the heating process on the reliability of the action are analyzed to determine the heating characteristics of the thermal explosion tube.

[0011] Furthermore, the instantaneous stress distribution of each component of the release device under thermal explosion impact is obtained, the instantaneous stress distribution is subjected to time-domain filtering, and the shock wave generated by the explosion of the thermal explosion tube is obtained by collecting the pressure of the thermal explosion tube. Combined with the measured shock wave analysis, the stress distribution characteristics of each component are obtained, including:

[0012] The instantaneous stress distribution is acquired, and a time-domain filter is applied to the instantaneous stress distribution to obtain the filtered stress distribution. Pressure data of the thermal explosion tube is collected to determine the shock wave morphology. The shock wave morphology and measured shock wave parameters are fused to determine the stress transmission path between components and obtain the path stress gradient. Based on the matching degree between the path stress gradient and the filtered stress distribution, local peak stress points are extracted to determine the component deformation-sensitive area. The component deformation-sensitive area and measured shock wave parameters are analyzed to obtain the stress distribution characteristics of each component.

[0013] Furthermore, the step of analyzing the uniformity of stress distribution in components based on the stress distribution characteristics of each component, and identifying areas where the uniformity of stress distribution exceeds a preset threshold as local deformation damage areas, includes:

[0014] Based on the stress distribution characteristics of each component, a local coefficient of variation is calculated, which is the standard deviation of the stress data divided by the mean. Based on the coefficient of variation, a global uniformity index is calculated. If the global uniformity index exceeds a preset threshold, the stress peak-valley difference of each component is extracted. Based on the stress peak-valley difference, a height difference region is determined, and a preset distance is extended around the height difference region to determine the local deformation damage region.

[0015] Furthermore, the extraction of the maximum principal stress and strain amplitude of the local deformation and damage area, combined with the stress history data of the release device's previous actions, yields the accumulated fatigue damage value. The degree of fatigue damage to the component is then determined based on this accumulated fatigue damage value, including:

[0016] Extract the maximum principal stress value of the local deformation and damage area to generate a principal stress vector; calculate the strain amplitude of the local deformation and damage area based on the principal stress vector to determine the strain amplitude spectrum; fuse the strain amplitude spectrum and the historical stress data to construct a historical action sequence; calculate the fatigue damage accumulation value based on the historical action sequence; map the fatigue damage accumulation value to a preset level to determine the degree of fatigue damage of the component.

[0017] Furthermore, the remaining number of actuations of the release mechanism's mechanical structure is determined based on the degree of fatigue damage to the components, thus obtaining the remaining reliable actuation number of the release mechanism. The geometric dimensions of the red indicator device ejection channel are acquired, and the obtained channel cross-section and surface roughness are analyzed to obtain the ejection path resistance of the red indicator device in the release mechanism, including:

[0018] Based on the degree of fatigue damage to the component, a damage accumulation curve is generated; based on the damage accumulation curve, a remaining action threshold is determined, curve inflection points are extracted, and the number of remaining actions is determined; the number of remaining actions and overall load data are fused to determine the number of reliable actions; based on the number of reliable actions, the cross-sectional profile of the pop-up channel of the red indicator device is measured; based on the cross-sectional profile, surface roughness values ​​are collected, and the surface roughness values ​​and the cross-sectional profile are fused to determine the pop-up path resistance.

[0019] Furthermore, after determining the degree of fatigue damage to the component based on the accumulated fatigue damage value, the process includes:

[0020] The stress concentration areas of the connectors and supports in the mechanical structure of the release device are detected to obtain the local stress field. Based on the local stress field, the stress difference between high-stress areas and adjacent components is identified to determine weak connection points and deformation-sensitive parts. The geometric deformation and surface wear state of the inner wall of the ejection channel of the red indicator device are measured to generate a deformation profile map. Based on the deformation profile map, the hindering effect of channel cross-sectional shrinkage and increased surface roughness is evaluated to identify jamming points and areas of concentrated frictional resistance. Based on the jamming points and areas of concentrated frictional resistance, the effect of channel deformation on the extension of ejection time and the attenuation of ejection force transmission is quantified to determine the risk of device jamming and ejection failure.

[0021] Furthermore, the analysis of the attenuation rate and success rate reduction of the remaining reliable operation count of the release device based on the ejection path resistance of the red indicator device in the release device, and the determination of the resistance change trend based on the variation law of the resistance coefficient in previous tests, includes:

[0022] Based on the pop-up path resistance, path correlation data is generated, which is a correlation matrix between resistance and the number of actions. Based on the path correlation data, the decay rate of the remaining number of actions is calculated. Based on the decay rate, the magnitude of the success rate reduction is fused to determine the magnitude of the success rate reduction. The sequence values ​​of the resistance coefficient in each test are collected to construct a resistance coefficient sequence. Based on the resistance coefficient sequence, the change pattern is extracted, the change pattern is extended, and the trend of resistance change is determined.

[0023] Furthermore, the process of determining key wear locations and failure modes based on resistance change trends, and predicting the failure point of the release device using degradation trajectory extrapolation based on the determined wear rate and stress weak points, includes:

[0024] Based on the resistance change trend, a trend distribution map is generated; based on the trend distribution map, the failure mode is determined; based on the failure mode, the local value of the wear rate is extracted to determine the wear rate; based on the wear rate, the rate influence zone of the stress weak link is identified, and a degradation trajectory curve is generated; based on the degradation trajectory curve, high-risk points are located; based on the high-risk points, the local characteristics of the stress weak link are fused to determine the failure point of the release device.

[0025] Furthermore, the temperature of the conductive silicone rubber component at the fault point in the disconnector is acquired in real time to obtain the temperature rise trend of the thermal detonation tube. Based on the temperature rise trend of the thermal detonation tube, the response time delay and the degree of decrease in start-up success rate of the heating process on the reliability of the action are analyzed to determine the heating characteristics of the thermal detonation tube, including:

[0026] The temperature of the conductive silicone rubber component is acquired, and a sliding window filter is applied to obtain a filtered temperature sequence. Based on the filtered temperature sequence, a temperature rise trend of the heat-explosion tube is generated. Based on the temperature rise trend, heating stages are divided. Based on the heating stages, the response time delay is quantified, and the response time delay and the degree of success rate reduction are fused to determine the degree of success rate reduction. Based on the degree of success rate reduction, the rate component of the heating characteristic is extracted to determine the peak temperature position. Based on the peak temperature position, the heating rate and peak temperature are integrated to determine the heating characteristics of the heat-explosion tube.

[0027] This invention provides a fault prediction device based on the disconnection timing of a voltage-limiting arc-extinguishing lightning protection device, mainly comprising:

[0028] The stress distribution acquisition and analysis module is used to acquire the instantaneous stress distribution of each component of the release device under thermal explosion impact, perform time-domain filtering on the instantaneous stress distribution, obtain the shock wave generated by the thermal explosion tube by collecting the pressure of the thermal explosion tube, and analyze the stress distribution characteristics of each component by combining the measured shock wave.

[0029] The stress uniformity analysis module is used to analyze the uniformity of stress distribution in components based on the stress distribution characteristics of each component, and to identify areas where the stress distribution uniformity exceeds a preset threshold as local deformation damage areas.

[0030] The fatigue damage assessment module is used to extract the maximum principal stress value and strain amplitude of the local deformation damage area, and calculate the fatigue damage accumulation value by combining the stress history data of the release device in each operation. The degree of fatigue damage of the component is determined based on the fatigue damage accumulation value.

[0031] The remaining action and resistance analysis module is used to determine the remaining number of actions of the releaser mechanical structure based on the degree of fatigue damage of the components, obtain the remaining reliable number of actions of the releaser, obtain the geometric dimensions of the red indicator device ejection channel, and analyze the obtained channel cross-section and surface roughness to obtain the ejection path resistance of the red indicator device in the releaser.

[0032] The reliability decay and resistance trend analysis module is used to analyze the decay rate and success rate reduction of the remaining reliable action number of the release device based on the ejection path resistance analysis of the red indicator device in the release device, and to determine the resistance change trend based on the change law of the resistance coefficient in each test.

[0033] The fault prediction module is used to determine the key wear parts and failure modes based on the resistance change trend. Based on the determined wear rate and stress weak points, the failure point of the release device is predicted by the degradation trajectory extrapolation method.

[0034] The temperature monitoring and thermal characteristic analysis module is used to acquire the temperature of the conductive silicone rubber component at the fault point in the disconnector in real time, obtain the temperature rise trend of the thermal explosion tube, analyze the response time delay and the degree of decrease in start-up success rate of the heating process based on the temperature rise trend of the thermal explosion tube, and determine the heating characteristics of the thermal explosion tube.

[0035] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0036] This invention discloses a fault prediction method and device based on the timing of disengagement action of a voltage-limiting arc-extinguishing lightning protection device. It addresses the operational scenario problem of decreased operational reliability caused by uneven stress distribution, fatigue damage accumulation, and increased ejection path resistance under high-temperature and high-pressure impacts. The invention acquires the shock wave from the thermal explosion tube, combines it with time-domain filtering to process instantaneous stress distribution, analyzes the uniformity of stress distribution in each component, identifies local deformation damage areas and stress concentration weak points, extracts the maximum principal stress and strain amplitude, calculates fatigue damage accumulation using historical stress data, and predicts the remaining number of reliable actions. Simultaneously, by measuring the geometry and surface roughness of the ejection channel of the red indicator device, it analyzes the resistance change trend, identifies jamming points and areas of concentrated frictional resistance, and assesses the attenuation effect of channel deformation on ejection time and force transmission. Based on real-time temperature monitoring of conductive silicone rubber components and analysis of the thermal explosion tube's temperature rise trend, it predicts the fault point and the extent of reliability degradation. This invention significantly improves the accuracy of disengagement performance degradation prediction, optimizes operational reliability assessment, and provides a scientific basis for mechanical structure design and maintenance under complex operating conditions. Attached Figure Description

[0037] Figure 1 This is a flowchart of a fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to the present invention.

[0038] Figure 2 This is a schematic diagram of the structure of a fault prediction device based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to the present invention. Detailed Implementation

[0039] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0040] like Figure 1-2 This embodiment of a fault prediction method and device based on the timing of disconnection action of a voltage-limiting arc-extinguishing lightning protection device may specifically include:

[0041] Step S101: Obtain the instantaneous stress distribution of each component of the release device under thermal explosion impact, perform time-domain filtering on the instantaneous stress distribution, obtain the shock wave generated by the thermal explosion tube by collecting the pressure of the thermal explosion tube, and analyze the stress distribution characteristics of each component in combination with the measured shock wave.

[0042] The instantaneous stress distribution of the release device components under the thermal explosion impact is captured by a sensor array. A time-domain filter is applied to this instantaneous stress distribution to obtain a filtered stress distribution. Pressure data is collected from inside the thermal explosion tube to determine the shock wave pattern generated by the tube's rupture. Based on this shock wave pattern, the measured shock wave parameters corresponding to the pattern are fused to determine the stress transmission path between components and obtain the path stress gradient. The matching degree between the path stress gradient and the filtered stress distribution is obtained. If the matching degree exceeds a preset threshold, local peak stress points are extracted to determine the component deformation-sensitive area. The influence of the measured shock wave parameters on the component deformation-sensitive area is analyzed to obtain the stress distribution characteristics of each component.

[0043] Specifically, in one embodiment, the instantaneous stress distribution of the release device components under thermal shock is captured by a sensor array deployed at key locations on the release device.

[0044] Specifically, the sensor array includes multiple strain gauges and piezoelectric sensors, which are arranged on the housing, connectors, and supports of the release device. When the thermal detonator ruptures and generates a shock wave, the sensors record stress changes on the component surfaces in real time. The capture process involves synchronously acquiring multi-channel signals to ensure data time alignment. Time-domain filtering is applied to the captured instantaneous stress distribution. This filtering process involves using a low-pass filter to remove high-frequency noise interference, such as a Butterworth filter to smooth the signal. The filter order is set according to the signal frequency characteristics, resulting in a filtered stress distribution. This filtered stress distribution reflects the pure stress response of the release device components under impact, providing fundamental data for subsequent analysis. In power systems, this capture and filtering method is applicable to release devices of different specifications, ensuring accurate monitoring of structural response in lightning overvoltage scenarios. Based on the filtered stress distribution, pressure data is acquired from inside the thermal detonator.

[0045] Specifically, pressure sensors, such as dynamic pressure probes, are installed on the inner wall of the thermal explosion tube. These sensors are connected to a data acquisition system via wired or wireless means to record the internal pressure fluctuation curve at the moment of the thermal explosion tube's rupture. The acquired data includes the pressure peak value and duration. The shock wave morphology generated by the thermal explosion tube's rupture is determined using curve fitting methods. This morphology describes the propagation speed and waveform characteristics of the shock wave, such as a sine wave or pulse waveform. This acquisition method is widely used in surge arrester fault simulation testing to ensure that the data accurately reflects the rupture process. Further, for the shock wave morphology, the measured shock wave parameters corresponding to the shock wave morphology are fused. This fusion process refers to combining the waveform data of the shock wave morphology with measured parameters such as peak amplitude and attenuation rate, and calculating a comprehensive index using a weighted average method. For example, the morphological feature vector and parameter vector are concatenated and then input into a fusion function to obtain a unified description. The stress transmission path between components is determined. This determination is based on the fusion results and tracks the propagation trajectory of stress from the thermal explosion tube to the outer shell and connectors. For example, finite element simulation is used to track stress changes at path nodes and identify the main transmission chain. The path stress gradient is obtained by calculating the slope value obtained by dividing the stress difference between adjacent nodes by the distance, quantifying the attenuation or amplification trend of stress along the path. In surge arrester disconnectors, this fusion and judgment helps to reveal the interaction effects between components under thermal shock, providing a basis for structural optimization.

[0046] Preferably, the matching degree between the path stress gradient and the filtered stress distribution is obtained. This matching degree calculation process involves using the cosine similarity formula to perform a dot product operation between the path stress gradient vector and the filtered stress distribution vector, dividing by the product of their respective magnitudes, to obtain a value between 0 and 1. If the matching degree exceeds a preset threshold, for example, a threshold set to 0.8, local peak stress points are extracted. This extraction is achieved by scanning the vector using a peak detection algorithm to locate the position of maximum stress. The component deformation-sensitive area is determined; this area refers to the region surrounding the peak point, and its boundary is defined using a radius search method.

[0047] In one possible implementation, this matching and extraction is applied to the periodic maintenance testing of surge arrester disconnectors to ensure the identification of potential deformation risks. The influence of the measured shock wave parameters is analyzed through the deformation-sensitive areas of the components.

[0048] Specifically, the analysis process includes calculating the correlation between stress data in the deformation-sensitive area and measured shock wave parameters such as amplitude and frequency. For example, the Pearson correlation coefficient is used to assess the impact of parameter changes on the stress within the area; the coefficient is calculated as the product of covariance and standard deviation. This analysis reveals how the shock wave causes localized stress concentration, thereby affecting the overall stability of the release device. The stress distribution characteristics of each component are obtained; these characteristics are represented by a distribution map generated by integrating the analysis results, describing the stress patterns of components such as the red indicator and support members under impact. In another implementation, the sensor array's capture can be extended to multi-dimensional monitoring, including adding a temperature sensor to assist in stress data correction. The time-domain filtering process is further optimized to adaptive filtering, dynamically adjusting the cutoff frequency based on the signal-to-noise level to ensure the accuracy of the filtered stress distribution.

[0049] It should be noted that pressure data acquisition can be combined with a real-time monitoring system to achieve automated triggering and recording. When determining the shock wave pattern, a waveform classification method is used to distinguish different blasting types, such as rapid blasting or gradual blasting, thereby adapting to the application of surge arrester disconnectors under different operating conditions.

[0050] Specifically, the detailed process of integrating measured shock wave parameters involves parameter standardization followed by principal component analysis to extract key influencing factors, which are then integrated with the shock wave morphology. Stress propagation paths can be determined by modeling components as nodes and paths as edges using graph theory, calculating the shortest stress propagation path. Once the path stress gradient is obtained, it can be used to simulate the cumulative effects of multiple impacts. In surge arrester systems, this method improves response analysis to complex thermal explosion scenarios. Furthermore, after identifying the deformation-sensitive areas of components, the analysis of influence can include the application of finite deformation theory to assess the plastic deformation threshold of materials within these areas. When stress distribution characteristics are obtained, a three-dimensional visualization map is generated, highlighting high-risk components. This feature directly supports the design iteration of disconnectors, enabling more reliable action prediction in the field of power protection.

[0051] Step S102: Analyze the uniformity of stress distribution in each component based on the stress distribution characteristics of each component, and identify the areas where the stress distribution uniformity exceeds a preset threshold as local deformation damage areas.

[0052] Based on the stress distribution characteristics of each component, the local coefficient of variation of stress distribution is obtained. This coefficient of variation is calculated by dividing the standard deviation of the stress data by the mean, resulting in a coefficient of variation distribution. Using this coefficient of variation distribution, a global uniformity index of stress distribution is calculated. This index is a weighted average of the coefficients of variation, determining the uniformity index value. If the uniformity index value exceeds a preset threshold, stress peak-valley differences are extracted for the component regions exceeding the threshold. This difference is calculated by subtracting the valley value from the peak value, obtaining a set of peak-valley differences. The elevation difference regions within this set of peak-valley differences are then identified, and boundary expansion is applied to these regions. This expansion involves extending the region by a preset distance to determine the local deformation and damage areas.

[0053] Specifically, in one implementation, the local coefficient of variation of the stress distribution of each component is obtained based on its stress distribution characteristics. This coefficient of variation is the ratio calculated by dividing the standard deviation of the stress data within the component by the mean. The standard deviation reflects the dispersion of stress values, and the mean represents the average stress level. By dividing the component into grid regions, this calculation is performed on the stress points in each region to obtain the coefficient of variation distribution. This distribution is presented in the form of a heat map, highlighting stress variation hotspots in disconnector components such as connectors and supports. In surge arrester thermal shock tests, this method is suitable for simulating multiple fault scenarios, ensuring the capture of dynamic stress changes. The global uniformity index of the stress distribution of the component is calculated using the coefficient of variation distribution. This index is a weighted average of the coefficients of variation, where the weights are allocated according to the component area or importance; for example, supports have a higher weight than the outer shell. The uniformity index value is determined by summing the coefficients of variation, multiplying by the weights, and then dividing by the total weight. This value quantifies the overall balance of stress distribution, providing a basis for assessing structural stability in the maintenance of surge arrester disconnectors in power systems. Furthermore, if the uniformity index value exceeds a preset threshold, then for the component area exceeding the threshold, the stress peak-valley difference value is extracted. This difference value is the value obtained by subtracting the peak value and valley value identified from the stress curve. The peak value is the local maximum stress point, and the valley value is the adjacent minimum point. By scanning the stress data sequence within the region, these points are located and the difference is calculated to obtain the peak-valley difference value set. This set reflects the severity of the uneven distribution and is used to identify potential crack initiation zones in the fault analysis of surge arrester disconnectors.

[0054] Preferably, the elevation difference region in the set of peak-valley differences is obtained, and a boundary extension is applied to the elevation difference region. This extension refers to extending the elevation difference region by a predetermined distance, for example, drawing a circular boundary centered on the elevation difference point and extending it outward by a fixed width to form an extension area, thereby determining the local deformation and damage area. This area includes deformed parts that may be affected by thermal shock.

[0055] In one embodiment, the acquisition of local variation coefficients can be extended to three-dimensional stress fields, and variation can be calculated through volumetric meshes to ensure application on complex curved surface components such as thermal burst pipe interfaces.

[0056] Specifically, the calculation of the global uniformity index can incorporate dynamic weights, adjusting the weighting scheme according to the impact intensity to adapt to the evaluation of surge arresters at different voltage levels.

[0057] For example, when extracting the stress peak-valley difference, a sliding window method is used to scan the data, and the window width is set according to the component size to improve accuracy in noisy environments.

[0058] It should be noted that the boundary expansion process can be combined with image processing techniques, such as dilation operations, to expand the boundaries of elevation difference areas and achieve automated damage area delineation. Furthermore, in multiple thermal explosion tests of surge arresters, this method can achieve early warning of damaged areas by continuously monitoring uniformity changes.

[0059] Step S103: Extract the maximum principal stress value and strain amplitude of the local deformation and damage area, calculate the fatigue damage accumulation value by combining the stress history data of the release device in each operation, and determine the degree of fatigue damage of the component based on the fatigue damage accumulation value.

[0060] The maximum principal stress value is extracted from the locally deformed and damaged area. This principal stress value refers to the maximum eigenvalue of the principal stress tensor within the area. A principal stress vector is generated from this maximum principal stress value by projecting the maximum value along the coordinate axis. Based on the principal stress vector, the strain amplitude of the locally deformed and damaged area is obtained. This amplitude refers to the difference between the maximum and minimum strain values. The strain amplitude fluctuation range is determined, and the range refers to the standard deviation of the amplitude, thus determining the strain amplitude spectrum. For the strain amplitude spectrum, historical stress data from each action of the release device is fused. This fusion refers to the multiplication of the spectrum with the historical data matrix. A historical action sequence is constructed using the historical stress data. This sequence refers to historical stress points ordered by time, resulting in a sequence fusion result. Based on the sequence fusion result, a fatigue damage accumulation value is calculated. This value is the sequence integral. If the fatigue damage accumulation value exceeds a preset threshold, the accumulation value weight is adjusted by multiplying by a preset coefficient to obtain a weighted accumulation value. Based on the weighted accumulation value, the degree of fatigue damage to the component is evaluated. This evaluation refers to mapping the accumulation value to a preset level to determine the damage level.

[0061] Specifically, in one implementation, the maximum principal stress value is extracted from the locally deformed and damaged region. This principal stress value refers to the maximum eigenvalue of the principal stress tensor within the region. The principal stress tensor is a second-order tensor describing the three-dimensional stress state within the material, and its eigenvalue is obtained by solving the characteristic equation. The maximum eigenvalue represents the peak intensity along the principal stress direction. In the thermal explosion impact scenario of a surge arrester disconnector, this extraction process involves scanning the stress grid points of the damaged region and selecting the highest eigenvalue. A principal stress vector is generated from the maximum principal stress value. This vector represents the maximum value projected along the coordinate axes, i.e., decomposed into x, y, and z components to form a vector representation. This generation supports the quantification of stress orientation in the damaged region and is suitable for analyzing the directional deformation of connectors and supports in power system surge arrester fault simulation. Based on the principal stress vector, the strain amplitude of the locally deformed and damaged region is obtained. This amplitude refers to the difference between the maximum and minimum strain. The corresponding strain field is calculated through the vector components; the maximum strain is the vector magnitude multiplied by the material modulus, and the minimum strain is the negative projection. The difference reflects the deformation amplitude. The strain amplitude fluctuation range is determined, which refers to the amplitude standard deviation, i.e., taking the square root of the variance over multiple amplitude samples to quantify the fluctuation instability. A strain amplitude spectrum is then determined, which is a curve plotting the amplitude distribution according to frequency. In multiple operation tests of the surge arrester disconnector, this acquisition and determination reveals the strain dynamics under thermal shock. Further, for the strain amplitude spectrum, historical stress data from previous disconnector operations are fused. This fusion involves matrix multiplication of the spectrum and historical data; the spectrum is converted into a row vector, and the historical data into a column matrix. Element-wise multiplication and summation are performed according to matrix multiplication rules to generate a fusion matrix, integrating the current spectrum with past stress patterns. A historical action sequence is constructed using the historical stress data; this sequence refers to historical stress points ordered chronologically, with the stress values ​​of each point arranged in the order of action occurrence. The sequence fusion result is obtained by flattening the fusion matrix into a vector and concatenating it with the sequence.

[0062] It should be noted that this fusion process, in the reliability assessment of surge arrester disconnectors, processes data from previous thermal explosion events to ensure the capture of cumulative effects.

[0063] In one possible implementation, fusion can be extended to weighted matrix multiplication, assigning weights based on the intensity of the action to improve sensitivity to high-impact events.

[0064] Preferably, the fatigue damage accumulation value is calculated based on the sequence fusion result. This value refers to the sequence integral, that is, the cumulative sum along the time axis of the fusion result vector, with each point added to the previous point to form the final value of the accumulation curve. If the fatigue damage accumulation value exceeds a preset threshold, the accumulation value weight is adjusted. This weight refers to multiplying by a preset coefficient, for example, a coefficient based on the material fatigue limit, to obtain a weighted accumulation value. This adjustment quantifies the over-threshold damage amplification, providing correction for cumulative deviations in surge arrester disconnector maintenance.

[0065] In one embodiment, the degree of fatigue damage to a component is assessed based on the weighted accumulated values. This assessment involves mapping the accumulated values ​​to preset levels, such as segmenting the accumulated values ​​into minor, moderate, and severe levels, and determining the specific level through comparisons within threshold ranges. The damage level is then determined, represented by numbers or labels, reflecting the fatigue state of the component, such as a red indicator device. In the field of surge arrester disconnectors in power systems, this assessment supports the development of replacement cycles.

[0066] For example, the extraction of principal stress values ​​can be combined with finite element mesh refinement to improve accuracy in areas with complex geometric damage, such as the interface of a thermally exploded pipe.

[0067] Specifically, the determination of the strain amplitude spectrum can involve Fourier transform to convert the amplitude to the frequency domain, highlighting periodic fluctuations, which can be applied in repetitive thermal shock scenarios. Furthermore, the construction of historical action sequences can include data cleaning, removing outliers, and sorting the data to ensure the reliability of the fusion results. In long-term monitoring of surge arrester disconnectors, this sequence processing reveals damage accumulation patterns.

[0068] Understandably, the calculation of fatigue damage accumulation can be performed using the trapezoidal integral method, which sums the sequence in a trapezoidal shape to accurately capture nonlinear accumulation.

[0069] In one embodiment, the determination of the damage level can be extended to probability mapping, using accumulated values ​​to calculate the failure probability, thereby enhancing the statistical depth of the assessment.

[0070] Step S104: Determine the remaining number of actions of the releaser mechanical structure based on the degree of fatigue damage of the components, obtain the remaining reliable number of actions of the releaser, acquire the geometric dimensions of the red indicator device ejection channel, and analyze the acquired channel cross-section and surface roughness to obtain the ejection path resistance of the red indicator device in the releaser.

[0071] Based on the fatigue damage degree of the component, a damage accumulation curve of the mechanical structure is obtained. This curve is obtained by integrating the damage degree along the number of actions. Using the damage accumulation curve, the remaining action threshold of the mechanical structure is determined. This threshold is the point where the curve slope changes. If the remaining action threshold is lower than a preset level, the curve inflection point is extracted. This point is obtained through inflection point detection to determine the remaining number of actions. For the remaining number of actions, the overall load data of the release device is fused. This fusion refers to weighting and superimposing the load data onto the number of actions to obtain a reliable number of actions. Based on the reliable number of actions, the geometric dimensions of the pop-up channel of the red indicator device are obtained. The cross-sectional profile of the channel is measured using these geometric dimensions. This measurement refers to obtaining the profile using laser scanning to obtain a cross-sectional profile diagram. For the cross-sectional profile diagram, the surface roughness value of the channel is collected. This value is the reading obtained by a roughness meter. By fusing the surface roughness value with the cross-sectional profile diagram, which is obtained by multiplying the roughness by the integral of the profile area, the pop-up path resistance is determined.

[0072] Specifically, in one implementation, a damage accumulation curve of the mechanical structure is obtained based on the degree of fatigue damage of the component. This curve is obtained by integrating the damage degree along the number of actions. The integration process involves using the damage degree value as a function and performing numerical integration along the number of actions axis, for example, using the trapezoidal rule to sum the area accumulation between adjacent points, forming an accumulation path from the initial to the current action. In the thermal explosion impact scenario of the surge arrester disconnector, this acquisition is applicable to damage tracking of support components and connectors, obtaining a damage accumulation curve. This curve depicts the accumulation trend, supporting the quantification of the durability decay of the mechanical structure. The remaining action threshold of the mechanical structure is determined through the damage accumulation curve. This threshold refers to the point where the curve slope changes. The slope is calculated by dividing the difference between adjacent points by the action interval. The location of the slope change is detected by scanning the curve and marked as the threshold point. If the remaining action threshold is lower than a preset level, the curve inflection point is extracted. This point refers to the point obtained by inflection point detection. The inflection point detection process includes calculating the second derivative and locating the position where the derivative crosses zero. The remaining number of actions is determined, which is the action increment projected from the inflection point to the end of the curve. This judgment and extraction reveals the critical point at which the structure approaches failure, providing a basis for remaining life assessment. Further, for the remaining number of actions, the overall load data of the disconnector is fused. This fusion refers to weighting and superimposing the load data (including thermal shock intensity and environmental factors) onto the number of actions, multiplying them by the remaining number of actions according to their weights, with weights allocated according to load type (e.g., shock intensity has a higher weight). A reliable number of actions is obtained, reflecting a conservative prediction after fusion. Under arrester disconnector failure simulation testing, this fusion process integrates the influence of multiple variables to ensure the reliability of the number of actions. Based on the reliable number of actions, the geometry of the red indicator device ejection channel is obtained. The channel cross-sectional profile is measured using these geometry. This measurement refers to obtaining the profile using laser scanning. The laser scanning process involves emitting a laser beam onto the channel surface, reconstructing a three-dimensional profile point cloud from the reflected signal, and then fitting it to a cross-sectional curve. A cross-sectional profile map is obtained, describing the elliptical or rectangular cross-sectional shape of the channel. In arrester disconnector design, this acquisition and measurement supports the evaluation of the geometric stability of the ejection mechanism, especially in scenarios where the channel deforms after multiple actions. For the cross-sectional profile map, the surface roughness value of the channel is collected. This value refers to the reading obtained by a roughness tester. The roughness tester process includes sliding a probe along the channel surface, recording microscopic height deviations, and calculating the average roughness index. The surface roughness value is fused with the cross-sectional profile map. This fusion involves multiplying the roughness by the profile area integral, first calculating the cross-sectional area of ​​the profile map, then multiplying it by the roughness value and integrating along the path length to generate a resistance coefficient. The ejection path resistance is determined, and this resistance quantification indicates the frictional resistance encountered by the ejection device.

[0073] Preferably, the acquisition of the damage accumulation curve can be extended to multi-dimensional integration, incorporating the temperature variable as an additional axis to improve accuracy under varying operating conditions.

[0074] Specifically, the determination of the remaining action threshold can be combined with machine vision-assisted curve analysis to automatically detect slope changes and adapt to large-scale testing. The fusion of overall load data can introduce a time decay factor, weighting the load by action interval to enhance the focus on recent events.

[0075] In one possible implementation, optical scanning can be used instead of lasers to measure the channel cross-sectional profile, improving applicability in dusty environments.

[0076] It should be noted that the fusion of surface roughness and profile map can be further subdivided into piecewise integration, calculating the resistance separately for different parts of the channel, highlighting local high-resistance areas. Furthermore, after obtaining the number of reliable actions, it can be linked to the dynamic monitoring of the channel geometry; changes in the number of actions trigger remeasurement, ensuring the timeliness of resistance determination.

[0077] In one embodiment, the determination of ejection path resistance can incorporate fluid dynamics corrections, integrate airflow effects, and be extended to high-speed ejection scenarios.

[0078] The system detects stress concentration areas in the connectors and supports of the release device's mechanical structure, identifies stress differences between high-stress areas and adjacent components, determines weak connection points and deformation-sensitive parts in the stress transmission path, measures the geometric deformation and surface wear of the inner wall of the ejection channel of the red indicator device, analyzes the hindering effect of channel cross-sectional shrinkage and increased surface roughness on the device's ejection, identifies jamming points and areas of concentrated frictional resistance in the ejection path, assesses the impact of channel deformation on the extension of ejection time and the attenuation of ejection force transmission, and determines the risk of device jamming due to local deformation and the risk of ejection failure due to increased resistance.

[0079] The stress concentration areas of the connectors and supports in the mechanical structure of the release device are detected by a sensor array, and the local stress field of the stress concentration area is obtained. This field refers to the grid formed by the sensor readings. Based on the local stress field, the stress difference between the high-stress area and adjacent components is identified. The stress difference is mapped and transmitted through a chain, which refers to the chain formed by connecting the difference gradients, to determine the weak connection points and deformation-sensitive parts in the stress transmission path. For the weak connection points and deformation-sensitive parts, the geometric deformation and surface wear state of the inner wall of the ejection channel of the red indicator device are measured. If the geometric deformation exceeds a preset threshold, a deformation profile map is extracted. This map refers to the map obtained by profile fitting. Based on the deformation profile map, the hindering effect of the channel cross-section shrinkage and the increase in surface roughness on the ejection of the device is evaluated. The distribution characteristics of the hindering effect are fused. This fusion refers to the superposition of the effect values ​​to identify the jamming points and friction resistance concentration areas in the ejection path. Based on the jamming point and the area of ​​concentrated frictional resistance, the effect of channel deformation on the extension of device ejection time and the attenuation of ejection force transmission is quantified. This quantification refers to calculating the time difference and force attenuation value to determine the risk of device jamming caused by local deformation and the risk of ejection failure caused by increased resistance.

[0080] Specifically, in one embodiment, stress concentration areas in the connectors and supports of the release mechanism's mechanical structure are detected using a sensor array. This sensor array includes strain sensors arranged at connector interfaces and support nodes, which capture localized high-stress points when thermal shock occurs. The local stress field of the stress concentration area is obtained; this field refers to a grid formed by sensor readings. The readings are extended into a continuous field using interpolation methods, such as Kriging interpolation connecting adjacent readings to form a stress intensity distribution grid. The local stress field quantifies the stress peak value in the concentration area. Based on the local stress field, the stress difference between the high-stress area and adjacent components is identified. This difference is calculated by subtracting the stress values ​​of the high-stress point from those of neighboring points, highlighting the non-uniform distribution. A stress difference mapping is used to transfer stress through a chain, which is a chain formed by connecting differential gradients. The gradient is calculated using the derivative of the stress field, and connecting the endpoints of the gradient vectors generates a path chain. Weak connection points and deformation-sensitive locations in the stress transfer path are identified; these points and locations refer to the locations of maximum gradient inflection points on the chain. Furthermore, for the weak connection points and deformation-sensitive areas, the geometric deformation and surface wear state of the inner wall of the red indicator device ejection channel are measured. This measurement uses an optical scanner to record the coordinate deviation and wear depth of the inner wall. Geometric deformation is calculated by comparing the offset with the original design coordinates, and surface wear is assessed by texture scanning to evaluate the degree of pitting. If the geometric deformation exceeds a preset threshold, a deformation profile map is extracted. This map is obtained through profile fitting, where the fitting process involves approximating the offset points using the least squares method to form a closed profile curve. The resulting deformation profile map depicts the deformation morphology of the inner wall of the channel. In the scenario of a surge arrester disconnector thermal explosion test, this measurement and extraction quantifies the channel changes after multiple actions, providing deformation evolution data.

[0081] Preferably, the deformation profile is used to evaluate the hindering effect of the channel cross-section shrinkage and the increase in surface roughness on the device ejection. This evaluation calculates the shrinkage rate as the proportion of the reduction in profile area and the increase in roughness as the increment of the average height deviation. The effects of both are integrated using the friction coefficient formula, and the hindering effect manifests as a slowdown in ejection speed. The distribution characteristics of the hindering effect are then fused; this fusion refers to the superposition of the effect values, adding the shrinkage and roughness values ​​point-by-point on the profile to form a comprehensive hindering field. Jamming points and areas of concentrated frictional resistance in the ejection path are identified; these points and areas refer to the peak positions and clusters of the hindering field.

[0082] In one embodiment, the effect of channel deformation on the extension of device ejection time and the attenuation of ejection force transmission is quantified based on the jamming point and the area of ​​concentrated frictional resistance. This quantification involves calculating the time difference and force attenuation value; the time difference is obtained by integrating the stagnation field through a simulated ejection trajectory, and the force attenuation is calculated by the cumulative reduction in resistance area. The risk of device jamming due to local deformation and the risk of ejection failure due to increased resistance are determined, and the probability of these risks is quantified by threshold comparison. The acquisition of the local stress field can be combined with vibration sensors to expand the mesh and capture dynamic stress fluctuations, which can be applied in high-frequency impact scenarios.

[0083] Specifically, threshold filtering can be introduced to identify stress differences, eliminate low-difference noise, and improve the accuracy of chain mapping.

[0084] It should be noted that Bezier curve fitting can be used to extract the deformation profile, improving the description of irregular deformations, particularly at the curved parts of the channel. Furthermore, the fusion of the hindrance effect can be layered and superimposed, processing shrinkage first and then adding roughness to ensure the accuracy of the distribution characteristics.

[0085] In one embodiment, the quantification of risk determination can be incorporated into Monte Carlo simulations, with repeated calculations of variation to obtain a statistical risk value.

[0086] Understandably, the identification of stuck points can be linked to 3D modeling, visualizing resistance areas, and assisting in positioning during maintenance planning.

[0087] Step S105: Analyze the attenuation rate and success rate reduction of the remaining reliable action count of the release device based on the ejection path resistance analysis of the red indicator device in the release device, and determine the resistance change trend based on the change law of the resistance coefficient in previous tests.

[0088] Based on the ejection path resistance of the red indicator device in the disconnector of the voltage-limiting arc-extinguishing lightning protection device, path correlation data of the reliable action attenuation is obtained. This data refers to the correlation matrix between resistance and the number of actions. The attenuation rate of the reliable action attenuation is evaluated using this path correlation data. This attenuation rate refers to the slope of the correlation data. If the attenuation rate exceeds a preset threshold, the magnitude of the success rate reduction is integrated. This index is the product of the attenuation rate and the success probability to determine the magnitude of the success rate reduction. For the magnitude of the success rate reduction, the sequence values ​​of the resistance coefficient from each test are collected. A resistance coefficient sequence is constructed using these sequence values, where the sequence values ​​are sorted by time. Based on the resistance coefficient sequence, the variation pattern of the resistance coefficient is extracted. This pattern refers to the sequence fitting curve. Future values ​​are extrapolated using this variation pattern, where the value is the curve extension prediction, to determine the trend of resistance change.

[0089] Specifically, in one implementation, path correlation data of reliable action attenuation is obtained based on the ejection path resistance of the red indicator device in the surge arrester. This data refers to the correlation matrix between resistance and the number of actions. By using resistance values ​​as rows and the number of actions as columns, the covariance is calculated and the matrix elements are filled to form a correlation structure. The resulting path correlation data captures the impact of resistance on action reliability. In the surge arrester disconnector thermal explosion test scenario, this acquisition is suitable for analyzing the attenuation of the indicator device ejection due to the accumulation of channel resistance. The attenuation rate of the reliable action attenuation is evaluated using the path correlation data. This rate refers to the slope of the correlation data. The slope is calculated by fitting the matrix rows using the least squares method to quantify the attenuation rate. If the attenuation rate exceeds a preset threshold, the magnitude index of the success rate reduction is integrated. This index is the product of the rate and the success probability. The slope is multiplied by the probability value to generate a magnitude value. The magnitude of the success rate reduction is determined, which reflects the quantified weakening of the impact of resistance growth on action success. Further, for the magnitude of the success rate reduction, the sequence values ​​of the resistance coefficient in each test are collected. This sequence value is extracted from the resistance readings of each action from the test log. A resistance coefficient sequence is constructed using the sequence values, where the sequence values ​​are ordered chronologically by test date to form a time-series chain. This resistance coefficient sequence depicts the evolution of resistance over time. Based on this sequence, the variation pattern of the resistance coefficients is extracted. This pattern refers to a sequence fitting curve, generated by fitting the sequence points using polynomial regression. Future values ​​are extrapolated using this variation pattern; these values ​​represent curve extension predictions. The equation is then input into future test points to calculate the expected resistance. The resistance change trend is determined, describing the expected upward or stable pattern. This extraction and extrapolation quantifies the long-term resistance evolution.

[0090] Preferably, the acquisition of path association data can be extended to a multivariate matrix, incorporating temperature factors to improve the accuracy of association under varying operating conditions.

[0091] Step S106: Determine the key wear parts and failure modes based on the resistance change trend, and predict the failure point of the release device by using the degradation trajectory extrapolation method based on the determined wear rate and stress weak points.

[0092] Based on the resistance change trend, a trend distribution map of the key wear parts is obtained. This map refers to the spatial mapping of trend data. Using this trend distribution map, the failure mode of the key wear parts is determined. This mode refers to pattern matching of the distribution map. If the failure mode matches a preset type, the local value of the wear rate is extracted. This value refers to the peak value of the rate, thus determining the wear rate. For the wear rate, the rate influence zone of the stress-weak link is identified. This zone refers to the rate gradient region. A degradation trajectory curve is constructed using the rate influence zone. This curve is obtained by integrating the influence zone. Based on the degradation trajectory curve, an extension segment of the degradation trajectory curve is extrapolated. This segment refers to the linear extension of the curve. High-risk points are located using the extension segment. These points are the peak values ​​of the extension segment, thus obtaining high-risk points. For these high-risk points, local features of the stress-weak link are integrated. These features refer to the superposition of stress values ​​in the weak link, thus determining the failure point of the release device.

[0093] Specifically, one-dimensional time-series resistance data is converted into a multi-dimensional spatial distribution form to achieve a visual representation of the deterioration state of key wear parts.

[0094] In one embodiment, the disconnector mechanism of the voltage-limiting arc-extinguishing lightning protection device includes several key components. Miniature pressure sensors are deployed at key locations such as the contact surface, spring contact surface, and guide groove to collect resistance change data at different locations. These sensor data are rearranged according to spatial coordinates to form a three-dimensional thermogram reflecting the resistance distribution within the entire disconnector.

[0095] For example, after 500 cycles, the resistance in the contact area increases from 2.1N to 3.8N, while the resistance in the guide groove increases from 0.8N to 2.2N. By using spatial mapping technology to distribute these values ​​according to their actual physical locations, the differences in wear levels across different parts can be visually reflected. Failure mode identification relies on a pre-defined pattern matching algorithm library, which compares the actual distribution map with standard failure modes for identification.

[0096] It should be noted that the typical failure modes of the disconnector in a voltage-limiting arc-extinguishing lightning protection device mainly include four basic types: contact adhesion, spring fatigue, guide wear, and comprehensive deterioration. When the trend distribution chart shows a sharp increase in resistance in the contact area while the surrounding area shows little change, the contact adhesion failure mode is automatically matched.

[0097] Specifically, if the peak resistance in the contact area exceeds three times the average resistance of the surrounding area and exhibits a clear localized concentration, it is identified as a contact adhesion pattern. The advantage of this pattern recognition technique is its ability to quickly pinpoint the root cause of the fault, avoiding a comprehensive overhaul of the entire system and significantly improving maintenance efficiency and cost control. The extraction of local wear rate values ​​focuses on identifying rate peaks in the distribution map; these peaks represent the most severely worn critical areas.

[0098] In one possible implementation, gradient analysis is performed on the trend distribution map to calculate the rate of change of resistance between adjacent regions, and the region with the largest rate of change is identified as the rate peak point.

[0099] For example, at the spring clamping part of the release mechanism, when the resistance increases from 1.5N to 2.8N over 10 consecutive cycles, the wear rate reaches a peak level of 13% per cycle; this value becomes the characteristic wear rate of that part. By accurately extracting these local peak values, the degree of deterioration in the most dangerous areas can be quantified. The identification of the rate-affected zone of stress-weak points is based on the principle of rate gradient analysis, determining the boundary of the affected area by calculating the spatial distribution characteristics of rate changes.

[0100] For example, when the wear rate in a certain area inside the ejector changes significantly, this change spreads to the surrounding area, forming an influence area with gradient characteristics.

[0101] For example, increased resistance due to wear of the main contact affects adjacent auxiliary contacts and the spring system. By analyzing the rate gradient changes in these related areas, a circular influence zone with a radius of approximately 8 mm, centered on the main contact, can be identified. Integrating these influence zones transforms the spatially distributed effects into a time-series degradation trajectory curve. This curve not only reflects the deterioration trend at a single point but also embodies the systemic degradation evolution process. The extrapolation of the degradation trajectory curve employs a linear trend analysis method, predicting future time based on existing trajectory data.

[0102] Understandably, when the degradation trajectory curve shows that the overall performance of the disconnector decreases by 12% every 100 actions, a linear extrapolation algorithm can predict the performance status after the next 1000 actions. Peak points in the extension segment typically correspond to critical moments when the system may experience sudden changes or failures; accurate identification of these high-risk points is crucial for developing preventative maintenance strategies. This extrapolation prediction method can provide early warning information before actual failures occur, avoiding system downtime and safety risks caused by sudden failures. The fusion analysis of high-risk points and local characteristics of stress-weak links is achieved through multi-dimensional information overlay technology, comprehensively evaluating the time-domain prediction results with the spatial stress distribution.

[0103] For example, when extrapolation analysis predicts that the 800th action will be a high-risk period, the stress distribution of each weak link in the release mechanism at that moment is analyzed simultaneously. Stress concentration coefficients at the main contact and spring contact surface are calculated by superimposing stress values, revealing that they reach 2.3 and 1.8 respectively. The combined risk coefficient is 4.1, exceeding the set safety threshold of 3.5. Based on this fusion analysis, the fault point of the release mechanism is determined to be located at the interface between the main contact and the spring contact surface.

[0104] Step S107: The temperature of the conductive silicone rubber component at the fault point in the disconnector is acquired in real time to obtain the temperature rise trend of the thermal explosion tube. Based on the temperature rise trend of the thermal explosion tube, the response time delay and the degree of decrease in start-up success rate of the heating process on the reliability of the action are analyzed to determine the heating characteristics of the thermal explosion tube.

[0105] The temperature of the conductive silicone rubber component at the fault point in the decoupler is acquired in real time using a temperature sensor. A sliding window filter is applied to the temperature sequence, which is the average temperature sequence within a window, resulting in a filtered temperature sequence. Based on the filtered temperature sequence, a temperature rise trend of the thermal explosion tube is constructed, which is a sequence difference curve. The temperature rise stages are divided based on this trend, which is the segmentation of the trend inflection points. For each temperature rise stage, the response time delay of the temperature rise process to the reliability of the action is quantified. This delay is the integral of the stage duration. The degree of decrease in the startup success rate is determined by fusing the response time delay, which is the decrease in success rate multiplied by the probability. Based on the degree of decrease in success rate, the rate component of the thermal explosion tube's temperature rise characteristic is extracted, which is the slope of the decrease. The location of the peak temperature is determined by the correlation between the rate component and the peak temperature, which is the temperature corresponding to the maximum point of the component. For the peak temperature position, the duration of the heat preservation time is measured. This duration refers to the flat section after the peak. The heating rate and the peak temperature are integrated through the duration. This integration means that the average rate within the interval is spliced ​​with the peak temperature to determine the heating characteristics of the heat explosion tube.

[0106] Specifically, in one embodiment, the temperature of the conductive silicone rubber component at the fault point in the deactivator is acquired in real time using a temperature sensor. The sensor is fixed to the silicone rubber surface and samples temperature data at intervals to form a continuous sequence. A sliding window filter is applied to the temperature sequence. This filter refers to averaging the temperature sequence within a window. By selecting the window size, the arithmetic mean of the temperature within the window is calculated for each position in the sequence, and the original value is replaced to suppress sudden noise. A filtered temperature sequence is obtained, providing a smooth temperature trajectory. This acquisition and filtering method is suitable for tracking the thermal response of the fault point. Based on the filtered temperature sequence, a temperature rise trend of the thermal burst tube is constructed. This trend refers to a sequence difference curve. By calculating the temperature difference between points before and after, a difference sequence representing the instantaneous rate of increase is generated. The temperature rise stage is divided based on the rise trend. This division refers to trend inflection point segmentation. By detecting the zero-crossing point of the derivative of the difference curve, the sequence is divided into preheating, acceleration, and saturation stages. The temperature rise stage segmentation is obtained, which identifies key turning points in the thermal process. Furthermore, for the temperature rise stage segmentation, the response time delay of the temperature rise process to the reliability of the operation is quantified. The delay refers to the integral of the stage duration. For each stage's cumulative time span, the total delay value is obtained by summing the stage durations, quantifying the delayed temperature rise trigger. The degree of decrease in the startup success rate is fused through the response time delay. This degree refers to the delay multiplied by the probability reduction. The delay value is multiplied by a preset probability coefficient to generate a decrease index. The degree of decrease in the success rate is determined; this index assesses the negative impact of thermal effects on startup. Based on the degree of decrease in the success rate, the rate component of the thermal explosion tube's temperature rise characteristics is extracted. This component refers to the slope of the decrease; a decrease curve is fitted using linear regression, and the coefficient is calculated as the rate value. The correlation between the rate component and the peak temperature is established; this correlation refers to the temperature corresponding to the component's maximum point. The peak temperature position is scanned and mapped to the temperature sequence. The peak temperature position is determined, marking the temperature extreme. For the peak temperature position, the duration of the holding time interval is measured. This interval refers to the flat section after the peak; the segment duration is calculated by checking the continuous length of the temperature fluctuation range after the peak that is less than a threshold. By integrating the heating rate and peak temperature over a sustained interval, this integration involves concatenating the average rate within the interval with the peak temperature, first calculating the interval average slope, and then combining it with the peak temperature to form a characteristic set. This determines the heating characteristics of the thermal burst tube, which encapsulates the rate, peak, and time attributes. In surge arrester disconnector design verification, this measurement and integration quantifies thermal durability, supporting the optimization of insulation materials.

[0107] Preferably, the temperature sequence filtering can dynamically adjust the window size according to the sequence variance to improve adaptability.

[0108] Specifically, the construction of the temperature rise trend can introduce higher-order differences to capture nonlinear increases and apply it under complex heat sources.

[0109] For example, the quantification of response time delay can be phased weighted integrals, emphasizing the acceleration phase, and reflected in the reliability model.

[0110] In one possible implementation, the determination of peak temperature can verify multiple peaks, handle fluctuating sequences, and be used in real-time systems.

[0111] It should be noted that the measurement of the insulation range can be adaptively set to the threshold, and the fluctuation limit can be adjusted according to the peak level, making it applicable to tests under varying environmental conditions.

[0112] This invention provides a fault prediction device based on the disconnection timing of a voltage-limiting arc-extinguishing lightning protection device, mainly comprising:

[0113] The stress distribution acquisition and analysis module is used to acquire the instantaneous stress distribution of each component of the release device under thermal explosion impact, perform time-domain filtering on the instantaneous stress distribution, obtain the shock wave generated by the thermal explosion tube by collecting the pressure of the thermal explosion tube, and analyze the stress distribution characteristics of each component by combining the measured shock wave.

[0114] The stress uniformity analysis module is used to analyze the uniformity of stress distribution in components based on the stress distribution characteristics of each component, and to identify areas where the stress distribution uniformity exceeds a preset threshold as local deformation damage areas.

[0115] The fatigue damage assessment module is used to extract the maximum principal stress value and strain amplitude of the local deformation damage area, and calculate the fatigue damage accumulation value by combining the stress history data of the release device in each operation. The degree of fatigue damage of the component is determined based on the fatigue damage accumulation value.

[0116] The remaining action and resistance analysis module is used to determine the remaining number of actions of the releaser mechanical structure based on the degree of fatigue damage of the components, obtain the remaining reliable number of actions of the releaser, obtain the geometric dimensions of the red indicator device ejection channel, and analyze the obtained channel cross-section and surface roughness to obtain the ejection path resistance of the red indicator device in the releaser.

[0117] The reliability decay and resistance trend analysis module is used to analyze the decay rate and success rate reduction of the remaining reliable action number of the release device based on the ejection path resistance analysis of the red indicator device in the release device, and to determine the resistance change trend based on the change law of the resistance coefficient in each test.

[0118] The fault prediction module is used to determine the key wear parts and failure modes based on the resistance change trend. Based on the determined wear rate and stress weak points, the failure point of the release device is predicted by the degradation trajectory extrapolation method.

[0119] The temperature monitoring and thermal characteristic analysis module is used to acquire the temperature of the conductive silicone rubber component at the fault point in the disconnector in real time, obtain the temperature rise trend of the thermal explosion tube, analyze the response time delay and the degree of decrease in start-up success rate of the heating process based on the temperature rise trend of the thermal explosion tube, and determine the heating characteristics of the thermal explosion tube.

[0120] The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. The present invention has been described in detail with reference to preferred embodiments. Those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications and substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device, characterized in that, The method includes: The instantaneous stress distribution of each component of the release device under thermal explosion impact is obtained, the instantaneous stress distribution is processed by time domain filtering, and the shock wave generated by the explosion of the thermal explosion tube is obtained by collecting the pressure of the thermal explosion tube. The stress distribution characteristics of each component are obtained by combining the measured shock wave. Based on the stress distribution characteristics of each component, the uniformity of stress distribution in the component is analyzed, and the area where the stress distribution uniformity exceeds the preset threshold is identified as the local deformation and damage area. The maximum principal stress and strain amplitude of the local deformation and damage area are extracted, and the fatigue damage accumulation value is calculated by combining the stress history data of the release device during each operation. The degree of fatigue damage of the component is determined based on the fatigue damage accumulation value. The remaining number of actions of the releaser mechanical structure is determined based on the degree of fatigue damage of the components, the remaining number of reliable actions of the releaser is obtained, the geometric dimensions of the red indicator device ejection channel are obtained, and the ejection path resistance of the red indicator device in the releaser is obtained by analyzing the obtained channel cross-section and surface roughness. Based on the resistance analysis of the ejection path of the red indicator device in the release device, the attenuation rate and success rate of the remaining reliable action number of the release device are analyzed, and the trend of resistance change is determined based on the change law of the resistance coefficient in each test. Based on the trend of resistance change, the key wear parts and failure modes are determined. Based on the determined wear rate and stress weak points, the failure point of the release device is predicted by the degradation trajectory extrapolation method. The temperature of the conductive silicone rubber component at the fault point in the disconnector is acquired in real time to obtain the temperature rise trend of the thermal explosion tube. Based on the temperature rise trend of the thermal explosion tube, the response time delay and the degree of decrease in start-up success rate of the heating process on the reliability of the action are analyzed to determine the heating characteristics of the thermal explosion tube.

2. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, The instantaneous stress distribution of each component of the release device under thermal explosion impact is obtained, the instantaneous stress distribution is processed by time-domain filtering, and the shock wave generated by the explosion of the thermal explosion tube is obtained by collecting the pressure of the thermal explosion tube. The stress distribution characteristics of each component are obtained by combining the measured shock wave with the analysis, including: The instantaneous stress distribution is acquired, and a time-domain filter is applied to the instantaneous stress distribution to obtain the filtered stress distribution. Pressure data of the thermal explosion tube is collected to determine the shock wave morphology. The shock wave morphology and measured shock wave parameters are fused to determine the stress transmission path between components and obtain the path stress gradient. Based on the matching degree between the path stress gradient and the filtered stress distribution, local peak stress points are extracted to determine the component deformation-sensitive area. The component deformation-sensitive area and measured shock wave parameters are analyzed to obtain the stress distribution characteristics of each component.

3. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, The step of analyzing the uniformity of stress distribution in components based on the stress distribution characteristics of each component, and identifying areas where the uniformity of stress distribution exceeds a preset threshold as local deformation and damage areas, includes: Based on the stress distribution characteristics of each component, a local coefficient of variation is calculated, which is the standard deviation of the stress data divided by the mean. Based on the coefficient of variation, a global uniformity index is calculated. If the global uniformity index exceeds a preset threshold, the stress peak-valley difference of each component is extracted. Based on the stress peak-valley difference, a height difference region is determined, and a preset distance is extended around the height difference region to determine the local deformation damage region.

4. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, The maximum principal stress and strain amplitude of the extracted local deformation and damage area are combined with the stress history data of the release device's previous operations to calculate the cumulative fatigue damage value. The degree of fatigue damage to the component is determined based on this cumulative fatigue damage value, including: Extract the maximum principal stress value of the local deformation and damage area to generate a principal stress vector; calculate the strain amplitude of the local deformation and damage area based on the principal stress vector to determine the strain amplitude spectrum; fuse the strain amplitude spectrum and the historical stress data to construct a historical action sequence; calculate the fatigue damage accumulation value based on the historical action sequence; map the fatigue damage accumulation value to a preset level to determine the degree of fatigue damage of the component.

5. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, The remaining number of actuations of the release mechanism's mechanical structure is determined based on the degree of fatigue damage to the components, thus obtaining the remaining reliable actuation number of the release mechanism. The geometric dimensions of the red indicator device's ejection channel are obtained, and the obtained channel cross-section and surface roughness are analyzed to obtain the ejection path resistance of the red indicator device in the release mechanism, including: Based on the degree of fatigue damage to the component, a damage accumulation curve is generated; based on the damage accumulation curve, a remaining action threshold is determined, curve inflection points are extracted, and the number of remaining actions is determined; the number of remaining actions and overall load data are fused to determine the number of reliable actions; based on the number of reliable actions, the cross-sectional profile of the pop-up channel of the red indicator device is measured; based on the cross-sectional profile, surface roughness values ​​are collected, and the surface roughness values ​​and the cross-sectional profile are fused to determine the pop-up path resistance.

6. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, After determining the degree of fatigue damage to the component based on the accumulated fatigue damage value, the process includes: The stress concentration areas of the connectors and supports in the mechanical structure of the release device are detected to obtain the local stress field. Based on the local stress field, the stress difference between high-stress areas and adjacent components is identified to determine weak connection points and deformation-sensitive parts. The geometric deformation and surface wear state of the inner wall of the ejection channel of the red indicator device are measured to generate a deformation profile map. Based on the deformation profile map, the hindering effect of channel cross-sectional shrinkage and increased surface roughness is evaluated to identify jamming points and areas of concentrated frictional resistance. Based on the jamming points and areas of concentrated frictional resistance, the effect of channel deformation on the extension of ejection time and the attenuation of ejection force transmission is quantified to determine the risk of device jamming and ejection failure.

7. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, The analysis of the ejection path resistance of the red indicator device in the release device determines the attenuation rate and success rate reduction of the remaining reliable operation count of the release device, and the trend of resistance change is determined based on the variation law of the resistance coefficient in previous tests, including: Based on the pop-up path resistance, path correlation data is generated, which is a correlation matrix between resistance and the number of actions. Based on the path correlation data, the decay rate of the remaining number of actions is calculated. Based on the decay rate, the magnitude of the success rate reduction is fused to determine the magnitude of the success rate reduction. The sequence values ​​of the resistance coefficient in each test are collected to construct a resistance coefficient sequence. Based on the resistance coefficient sequence, the change pattern is extracted, the change pattern is extended, and the trend of resistance change is determined.

8. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, The process involves determining key wear areas and failure modes based on resistance change trends, and predicting the failure point of the release mechanism using degradation trajectory extrapolation based on the determined wear rate and stress weak points. This includes: Based on the resistance change trend, a trend distribution map is generated; based on the trend distribution map, the failure mode is determined; based on the failure mode, the local value of the wear rate is extracted to determine the wear rate; based on the wear rate, the rate influence zone of the stress weak link is identified, and a degradation trajectory curve is generated; based on the degradation trajectory curve, high-risk points are located; based on the high-risk points, the local characteristics of the stress weak link are fused to determine the failure point of the release device.

9. The fault prediction method based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device according to claim 1, characterized in that, The temperature of the conductive silicone rubber component at the fault point in the disconnector is acquired in real time to obtain the temperature rise trend of the thermal explosion tube. Based on the temperature rise trend of the thermal explosion tube, the response time delay and the degree of decrease in start-up success rate of the heating process on the reliability of the action are analyzed to determine the heating characteristics of the thermal explosion tube, including: The temperature of the conductive silicone rubber component is acquired, and a sliding window filter is applied to obtain a filtered temperature sequence. Based on the filtered temperature sequence, a temperature rise trend of the heat-explosion tube is generated. Based on the temperature rise trend, heating stages are divided. Based on the heating stages, the response time delay is quantified, and the response time delay and the degree of success rate reduction are fused to determine the degree of success rate reduction. Based on the degree of success rate reduction, the rate component of the heating characteristic is extracted to determine the peak temperature position. Based on the peak temperature position, the heating rate and peak temperature are integrated to determine the heating characteristics of the heat-explosion tube.

10. A fault prediction device based on the disconnection action timing of a voltage-limiting arc-extinguishing lightning protection device, characterized in that, The device includes: The stress distribution acquisition and analysis module is used to acquire the instantaneous stress distribution of each component of the release device under thermal explosion impact, perform time-domain filtering on the instantaneous stress distribution, obtain the shock wave generated by the thermal explosion tube by collecting the pressure of the thermal explosion tube, and analyze the stress distribution characteristics of each component by combining the measured shock wave. The stress uniformity analysis module is used to analyze the uniformity of stress distribution in components based on the stress distribution characteristics of each component, and to identify areas where the stress distribution uniformity exceeds a preset threshold as local deformation damage areas. The fatigue damage assessment module is used to extract the maximum principal stress value and strain amplitude of the local deformation damage area, and calculate the fatigue damage accumulation value by combining the stress history data of the release device in each operation. The degree of fatigue damage of the component is determined based on the fatigue damage accumulation value. The remaining action and resistance analysis module is used to determine the remaining number of actions of the releaser mechanical structure based on the degree of fatigue damage of the components, obtain the remaining reliable number of actions of the releaser, obtain the geometric dimensions of the red indicator device ejection channel, and analyze the obtained channel cross-section and surface roughness to obtain the ejection path resistance of the red indicator device in the releaser. The reliability decay and resistance trend analysis module is used to analyze the decay rate and success rate reduction of the remaining reliable action number of the release device based on the ejection path resistance analysis of the red indicator device in the release device, and to determine the resistance change trend based on the change law of the resistance coefficient in each test. The fault prediction module is used to determine the key wear parts and failure modes based on the resistance change trend. Based on the determined wear rate and stress weak points, the failure point of the release device is predicted by the degradation trajectory extrapolation method. The temperature monitoring and thermal characteristic analysis module is used to acquire the temperature of the conductive silicone rubber component at the fault point in the disconnector in real time, obtain the temperature rise trend of the thermal explosion tube, analyze the response time delay and the degree of decrease in start-up success rate of the heating process based on the temperature rise trend of the thermal explosion tube, and determine the heating characteristics of the thermal explosion tube.

Images