A method and system for predicting the deterioration trend of an arrester in extremely cold weather based on multi-physical field coupling

Abstract

The application belongs to the technical field of electrical equipment state monitoring, and specifically discloses a method and system for predicting the deterioration trend of a lightning arrester in extremely cold conditions based on multi-physical field coupling, which comprises the following steps: synchronously collecting multi-dimensional data such as bus voltage, end total current, surface temperature field and broadband mechanical vibration of the lightning arrester; based on the internal instantaneous Joule heat power, the electric parameter is deconstructed, and cross-correlation analysis is performed with the surface temperature field to calculate the effective heat conduction time delay representing the structural integrity; the newly added mechanical resonance frequency is extracted from the vibration signal, and the total current is modulated based sideband lock detection to identify the electromechanical coupling damage features causing the current modulation effect. The application combines the evolution trend of the heat conduction time delay and the detection results of the electromechanical coupling damage to generate a prediction report containing the structural thermal resistance state and the crack electrical risk level, which can distinguish the defect nature and evaluate the electrical risk, and is helpful for grading early warning.

Description

Technical Field

[0001] This invention belongs to the field of electrical equipment condition monitoring technology, and relates to a method and system for predicting the degradation trend of lightning arresters in extreme cold based on multi-physics field coupling. Background Technology

[0002] In high-voltage transmission systems, surge arresters are critical overvoltage protection devices, and their operational reliability directly impacts the safety of the entire power grid. Especially in extremely cold environments, low temperatures not only alter the physical properties of the internal metal oxide varistors and external composite insulation materials of the surge arrester, but also can cause icing and condensation on the internal or surface due to drastic changes in ambient temperature and humidity. These factors combined can induce potential defects such as material embrittlement, microcrack propagation, internal interface delamination, or moisture absorption. If these early defects are not detected in time, they can develop into serious insulation faults, leading to equipment damage or even large-scale power outages. Therefore, condition monitoring and early fault warning for surge arresters operating in such environments is a significant challenge in ensuring the stable operation of the power grid.

[0003] Currently, common methods used in the industry for online monitoring of surge arresters include monitoring the total leakage current or its resistive component, infrared thermal imaging temperature measurement, and partial discharge detection. Electrical quantity monitoring methods, such as extracting resistive current through the compensation capacitor method, can reflect changes in the electrical performance of the surge arrester varistor due to moisture or aging. Infrared thermal imaging technology captures the temperature distribution on the surface of the surge arrester in a non-contact manner to detect abnormal hot spots caused by uneven heating. Acoustic or ultra-high frequency methods are mainly used to capture partial discharge signals caused by internal insulation defects. These methods provide a basis for assessing the condition of surge arresters to a certain extent.

[0004] However, the aforementioned traditional methods have certain limitations in dealing with the degradation caused by multi-physics coupling in extremely cold environments. 1. Simple resistive current monitoring measures the macroscopic electrical response of the entire surge arrester core. When localized, non-penetrating mechanical cracks appear inside, they may not be reflected in the overall change of resistive current, resulting in insensitivity to early structural damage.

[0005] 2. Infrared thermometry measures the temperature of the outer surface of the equipment. In extremely cold environments, if the inside of the surge arrester is damp and icy or if the material delaminates to form an insulation layer, it may actually hinder the transfer of Joule heat generated inside to the outside, resulting in a normal or even low surface temperature. This could mask the true state of the internal valve plate, which may be overheated or structurally deteriorated, leading to a missed diagnosis.

[0006] 3. Vibration or acoustic emission monitoring, while capable of detecting mechanical events such as material cracking, cannot directly determine whether the mechanical crack has affected the internal conductive path. In other words, it cannot distinguish between purely structural changes and dangerous cracks with direct electrical risks, and the risk level assessment is not clear enough.

[0007] Meanwhile, the technical solutions for effectively integrating these monitoring data from different physical dimensions to reveal the inherent correlation of faults still need improvement.

[0008] Based on the above problems, the present invention aims to solve the problem that the existing technology is unable to distinguish the physical properties of internal structural defects of surge arresters under the action of multi-physics field coupling, and is unable to assess whether mechanical damage has posed a direct threat to the electrical circuit. Summary of the Invention

[0009] In order to overcome the above-mentioned defects of the prior art and to achieve the above objectives, the present invention proposes the following technical solution: a method for predicting the degradation trend of a surge arrester in extreme cold based on multi-physics field coupling, comprising: S1, acquiring multi-dimensional synchronous operation data of the target surge arrester in extreme cold environment, the multi-dimensional synchronous operation data including the instantaneous value of bus voltage, the instantaneous value of terminal full current, the surface temperature field array data distributed along the axial direction, and the broadband mechanical vibration signal at the base flange.

[0010] S2. Based on the instantaneous values ​​of the bus voltage and the total current at the end, the instantaneous Joule thermal power sequence of the internal valve plate of the surge arrester is deconstructed and used as the thermal power excitation source for thermal structure coupling analysis.

[0011] S3. Perform cross-cycle cross-correlation analysis using data from at least one temperature measurement point located in the middle region of the arrester valve plate group selected from the instantaneous Joule thermal power sequence and surface temperature field array data to calculate the effective thermal conduction delay parameter characterizing the integrity of the internal structure.

[0012] S4. Perform time-frequency domain analysis on the broadband mechanical vibration signal and compare it with the reference vibration data under healthy conditions to identify the set of new mechanical resonance characteristic frequencies caused by material deterioration or microcracks.

[0013] S5. Based on the mechanical resonance characteristic frequency set, the instantaneous value of the end current is modulated and sideband locked to identify the electromechanical coupling damage characteristics that cause the current modulation effect.

[0014] S6. Based on the evolution trend of the comprehensive effective thermal conduction delay parameters and the detection results of electromechanical coupling damage characteristics, a deterioration trend prediction report is generated, which includes the structural thermal resistance state and the electrical risk level of the crack.

[0015] The second aspect of the present invention provides a system for predicting the degradation trend of a surge arrester in extreme cold conditions based on multi-physics coupling, comprising: a multi-dimensional data synchronous acquisition module, which acquires multi-dimensional synchronous operation data of a target surge arrester in extreme cold environments, including instantaneous values ​​of bus voltage, instantaneous values ​​of terminal full current, surface temperature field array data distributed along the axial direction, and broadband mechanical vibration signals at the base flange.

[0016] The thermal circuit parameter extraction module, based on the instantaneous values ​​of the bus voltage and the total current at the end, deconstructs the instantaneous Joule thermal power sequence of the valve plates inside the surge arrester, which serves as the thermal power excitation source for thermal structure coupling analysis.

[0017] The heat conduction delay analysis module uses data from at least one temperature measurement point located in the middle region of the arrester valve plate group selected from the instantaneous Joule thermal power sequence and surface temperature field array data to perform cross-cycle cross-correlation analysis and calculate the effective heat conduction delay parameters characterizing the integrity of the internal structure.

[0018] The mechanical resonance feature extraction module performs time-frequency domain analysis on broadband mechanical vibration signals and compares them with reference vibration data under healthy conditions to identify the new set of mechanical resonance characteristic frequencies caused by material degradation or microcracks.

[0019] The electromechanical coupling damage detection module performs modulation sideband locking detection on the instantaneous value of the end current based on the mechanical resonance characteristic frequency set, and identifies the electromechanical coupling damage characteristics that cause the current modulation effect.

[0020] The comprehensive degradation trend prediction module integrates the evolution trend of effective heat conduction delay parameters with the detection results of electromechanical coupling damage characteristics to generate a degradation trend prediction report that includes the structural thermal resistance state and the electrical risk level of cracks.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention establishes a mapping relationship between instantaneous Joule thermal power and surface temperature field response using cross-period cross-correlation analysis, and calculates the effective heat conduction delay parameter. Instantaneous Joule thermal power is used as the internal heat source excitation, and surface temperature field change is used as the response. By calculating the peak delay of the cross-correlation function between the two, the heat transfer lag from the core to the surface is quantified. When there are defects such as delamination or moisture inside that increase thermal resistance, it is reflected in the increase of the delay parameter, thereby realizing the quantitative characterization of internal thermal insulation defects.

[0022] (2) This invention is based on the modulation correlation between mechanical vibration and electrical parameters, and identifies electromechanical coupling damage with electrical activity by locking specific sideband components in the total current. The resonant characteristic frequency of the mechanical vibration is extracted as a reference, and sideband components centered at the power frequency and offset from this characteristic frequency are searched in the total current spectrum. The detection of sideband components confirms that the mechanical vibration causes periodic changes in the geometry of the current path or the contact resistance, thereby determining whether the mechanical damage has caused electrical contact instability. This method eliminates the ambiguity of traditional vibration monitoring in determining the electrical hazards of cracks, and realizes a direct correlation diagnosis from mechanical damage to electrical threats.

[0023] (3) This invention integrates the evolution trend of effective heat conduction delay parameters with electromechanical coupling damage characteristics to generate a degradation trend prediction result. The rate of change of heat conduction delay parameters reflects the rate of gradual damage such as aging and penetration, while electromechanical coupling characteristics are used to determine whether there is sudden damage that causes contact instability. By integrating the evaluation information from both, the invention distinguishes between the state of "slow performance degradation" or "structural damage causing electrical anomalies", thereby achieving a multi-dimensional hierarchical evaluation of the health status of surge arresters and a degradation trend prediction. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention.

[0026] Figure 2 This is a schematic diagram of the system module connections of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Example 1

[0029] Please see Figure 1As shown, the present invention proposes a method for predicting the degradation trend of surge arresters in extreme cold environments based on multi-physics field coupling, including: S1, acquiring multi-dimensional synchronous operation data of the target surge arrester in extreme cold environments, including the instantaneous value of bus voltage, the instantaneous value of terminal full current, the surface temperature field array data distributed along the axial direction, and the broadband mechanical vibration signal at the base flange.

[0030] In a preferred embodiment, acquiring multi-dimensional synchronous operation data of the target surge arrester in extremely cold environments includes: in a low-temperature environment where the ambient temperature is below a preset temperature threshold, synchronously acquiring the instantaneous value of the bus voltage and the instantaneous value of the terminal full current through a high-voltage divider and a through-type current transformer.

[0031] Activate the fiber optic grating sensor array or non-contact infrared temperature measurement module arranged on the surface of the surge arrester jacket to collect the surface temperature field array data with location tags.

[0032] A high-frequency acoustic emission sensor coupled to the grounding flange of the surge arrester is used to continuously record waves and capture broadband mechanical vibration signals containing background noise.

[0033] Specifically, an extremely cold environment refers to working conditions where the ambient temperature remains below -40 degrees Celsius. The first implementation step of the method of this invention is executed by an edge computing terminal or remote server deployed at the monitoring site. This executing entity first sends a synchronous acquisition command to the high-voltage divider and the through-type current transformer, which are electrically connected to the target surge arrester, through its data acquisition interface. After receiving the command, these two sensor modules sample the AC waveform of the current power grid cycle with a microsecond-level time resolution, and amplify and filter the acquired analog signals through their respective signal conditioning circuits. The microsecond-level time resolution is achieved by setting the sampling rate of the data acquisition card to no less than one million sampling points per second, i.e., 1 MS / s, typically set between 2 MS / s and 10 MS / s, to ensure that rapid voltage or current transient processes caused by internal discharge or external switching operations can be captured without distortion.

[0034] Subsequently, the high-speed multi-channel analog-to-digital converter (ADC) inside the execution unit synchronously converts the two analog signals into digital time series, thereby generating the instantaneous values ​​of the bus voltage and the total current at the end, both containing precise timestamps. These two data streams are then temporarily stored in the first high-speed buffer. The instantaneous bus voltage is the real-time voltage value of the phase connection terminal of the target surge arrester to ground, measured in kilovolts (kV). The instantaneous total current at the end refers to the real-time value of the total current flowing through the surge arrester core and insulating outer sleeve and finally converging into the grounding terminal, measured in milliamperes (mA).

[0035] Simultaneously, the actuator activates the fiber optic grating sensor array arranged on the surface of the arrester's outer sleeve via an optical fiber communication interface, or initiates a non-contact infrared temperature measurement module via an industrial Ethernet protocol. The activated temperature sensing system measures the temperature at multiple preset points along the arrester's axis, binding the physical coordinates or unique identifier of each measurement point to the corresponding temperature reading, thus forming a surface temperature field array data with location tags. This data is transmitted to the actuator's second data buffer. The surface temperature field array data is a structured dataset where each data point contains location information and a temperature value. For example, for a arrester with a height of 6m, a measurement point can be set every 0.5m along its axis.

[0036] Almost simultaneously, the actuator sends a continuous recording command to the high-frequency acoustic emission sensor coupled to the grounding base flange of the surge arrester. The sensor begins to capture structural vibrations and convert them into electrical signals. After being amplified by a preamplifier, this signal is continuously acquired by another ADC channel of the actuator at a sampling rate in the megahertz range. This fully records the broadband mechanical vibration signal, including inherent noise of equipment operation and potential abnormal transient signals, and stores it in a third data buffer. The broadband mechanical vibration signal at the base flange refers to the raw voltage signal acquired by the high-frequency acoustic emission sensor. Its effective bandwidth is typically set to 20kHz to 1MHz. This frequency band is selected to cover the main energy distribution range of stress waves generated by the propagation of microcracks or the formation of ice crystals within the composite insulation material.

[0037] All data acquisition processes are synchronized by a unified clock source calibrated based on the GPS pulse-second signal (PPS) or Network Time Protocol (NTP), ensuring the timing consistency of the multi-dimensional synchronous operation data. The multi-dimensional synchronous operation data is a dataset that characterizes the arrester's operating status from three physical dimensions: electrical, thermal, and mechanical.

[0038] For example, in a specific application scenario, the monitoring system at a certain moment Data acquisition was performed on a 220kV surge arrester installed in an area with an ambient temperature of -47℃. First, the system used an ADC with a sampling rate of 2.5MS / s to acquire the instantaneous bus voltage of 181.3kV and the instantaneous terminal current of 0.85mA at that moment. These two values, along with the timestamp... Together, they were recorded. Next, the system polled the eight fiber Bragg grating temperature sensors deployed on the surge arrester, acquiring a set of surface temperature field array data with location tags: {sensor 1, -46.2℃}, {sensor 2, -46.1℃}, {sensor 3, -46.1℃}, {sensor 4, -45.8℃}, {sensor 5, -45.9℃}, {sensor 6, -46.0℃}, {sensor 7, -46.3℃}, {sensor 8, -46.5℃}. Subsequently, the system continuously acquired acoustic emission signals at the base flange at a sampling rate of 5 MS / s. Within a 1-millisecond time window, a segment of the digitized broadband mechanical vibration signal was captured, which manifested as a series of voltage values, such as [-0.11, 0.08, 0.23, -0.15, ...], in millivolts (mV). Finally, the system integrated and packaged these three different physical quantity data into a single file with a unified timestamp. The multidimensional synchronous running data frames are used for subsequent analysis and processing.

[0039] S2. Based on the instantaneous values ​​of the bus voltage and the total current at the end, the instantaneous Joule thermal power sequence of the internal valve plate of the surge arrester is deconstructed and used as the thermal power excitation source for thermal structure coupling analysis.

[0040] In a preferred embodiment, the instantaneous Joule thermal power sequence of the internal valve plate of the surge arrester is deconstructed, including: performing orthogonal decomposition or fast Fourier transform on the instantaneous values ​​of the bus voltage and the terminal full current.

[0041] The capacitive component in the instantaneous value of the total current at the end is removed, and the resistive current waveform characterizing the nonlinear conductivity of the valve plate is extracted.

[0042] The resistive current waveform is used as the basic current component for calculating the heating power.

[0043] In a further preferred embodiment, the instantaneous Joule thermal power sequence of the internal valve plate of the surge arrester is deconstructed, which further includes: performing a time-domain multiplication operation on the instantaneous value of the bus voltage and the resistive current waveform to generate an original power sequence that reflects the real-time heating state of the surge arrester core.

[0044] The original power sequence is denoised and smoothed to construct an instantaneous Joule thermal power sequence for subsequent calculation of the thermal response transfer function.

[0045] Specifically, after obtaining multi-dimensional synchronous operation data in the previous implementation step, the execution entity then processes the instantaneous values ​​of the bus voltage and the terminal total current temporarily stored in the first high-speed cache to deconstruct the instantaneous Joule thermal power sequence, which serves as the excitation source of thermal power. The instantaneous Joule thermal power sequence is a time series data, where each data point represents the total thermal power generated by all valves inside the surge arrester at the corresponding moment, serving as the internal virtual heat source for subsequent heat conduction analysis.

[0046] First, the execution unit reads the two synchronized time-series data from the buffer and performs frame processing, with each frame covering several power frequency cycles. For each data frame, the processing unit employs a Fast Fourier Transform algorithm. Specifically, it processes the instantaneous bus voltage value... and the instantaneous value of the total current at the end By performing transformations respectively, we can obtain their complex representations in the frequency domain. and This includes the amplitude and phase information of each frequency component.

[0047] Subsequently, the system performs orthogonal decomposition in the frequency domain, using the fundamental power frequency and its main odd harmonic frequencies as the analysis objects, such as 50Hz, 150Hz, and 250Hz. At each analysis frequency point... At this point, the total current phasor Decomposed into voltage phasors Parallel resistive components The capacitive component perpendicular to the direction of the current. The resistive component. The complex representation of the total current phasor is expressed by... voltage phasor The projection operation is performed. This operation utilizes the phase difference between the two phasors to separate the current component that is in phase with the voltage. After completing the calculations for all key frequency points, the system converts these resistive current components. Perform an inverse fast Fourier transform to reconstruct the resistive current waveform in the pure time domain. The resistive current waveform refers to the current component that is in phase with the instantaneous value of the bus voltage. This component is the direct cause of Joule heating in metal oxide varistors, and its amplitude and waveform distortion can reflect the nonlinear current-voltage characteristics and aging state of the varistor material. The data frame length used in Fast Fourier Transform (FFT) is typically set to cover 10 to 20 power frequency cycles. For example, in a 50Hz power grid, a data length of 200 to 400 milliseconds can be selected to obtain sufficient frequency resolution.

[0048] Next, the executing entity will convert the instantaneous value of the bus voltage in the time domain. With the extracted resistive current waveform Perform point-by-point multiplication to generate an original instantaneous power sequence. This sequence reflects the real-time heating status of the valve plates inside the surge arrester core.

[0049] Finally, to construct a stable input reference signal for subsequent calculations, the system applies digital filtering techniques to denoise and smooth the original power sequence. When smoothing the original power sequence, the window length of the moving average filter is set to one power frequency period, i.e., 20 milliseconds. This setting effectively filters out power fluctuation ripple at twice the power frequency while retaining the slow heating trend caused by valve plate degradation or changes in operating conditions. One implementation method uses a moving average filter to filter out high-frequency noise and glitches introduced by grid fluctuations by calculating the power average within a sliding time window, ultimately generating an instantaneous Joule thermal power sequence. It is then stored in the derived database for use in subsequent steps.

[0050] The formula for calculating Joule heat power is:

[0051]

[0052] in, This represents the instantaneous Joule thermal power sequence, measured in watts (W). This is the instantaneous value of the bus voltage, in volts (V). The resistive current waveform is obtained by frequency domain decomposition and time domain reconstruction of the instantaneous value of the total current at the end, and the unit is ampere (A).

[0053] For example, continuing from the previous step, the system obtains the instantaneous value of the bus voltage. and the instantaneous value of the total current at the end The waveform data. Assuming a 50Hz power frequency cycle, FFT analysis shows that the effective voltage value at the 50Hz fundamental frequency is... The phase is 0 degrees; the effective value of the total current is The phase leads the voltage by 85 degrees. Based on this, the system calculates the effective value of the resistive current under the 50Hz fundamental frequency. ,Right now The system performs the same calculations for other harmonic frequencies and synthesizes the resistive components of all frequencies into a time-domain resistive current waveform using inverse FFT. Subsequently, the system will generate the original instantaneous bus voltage value sequence. With the calculated resistive current waveform Perform point-by-point multiplication. For example, at time [time value missing]... , The calculated value is 256.4 kV. If the current is 0.105mA, then the original power at that moment is... The system performs this operation on the entire time series, obtaining a raw power series containing high-frequency fluctuations. Finally, the system applies a 20-millisecond moving average filter to smooth the raw power series. Assuming that... Within the 20-millisecond window of the given moment, the average value of the original power is 24.5W, so the value of the instantaneous Joule thermal power sequence generated by the system at that moment is 24.5W.

[0054] S3. Perform cross-cycle cross-correlation analysis using data from at least one temperature measurement point located in the middle region of the arrester valve plate group selected from the instantaneous Joule thermal power sequence and surface temperature field array data to calculate the effective thermal conduction delay parameter characterizing the integrity of the internal structure.

[0055] In a preferred embodiment, calculating the effective thermal conduction delay parameter characterizing the integrity of the internal structure includes: setting the instantaneous Joule thermal power sequence as the input signal and setting the temperature change of the temperature measurement point located in the middle region of the arrester valve plate group in the surface temperature field array data as the response signal.

[0056] Calculate the cross-correlation function between the input signal and the response signal;

[0057] Search for the time lag corresponding to the peak value of the cross-correlation function, and define this time lag as the effective heat conduction delay parameter.

[0058] In a further preferred embodiment, the process of generating the degradation trend prediction report further includes: continuously monitoring and recording the periodically calculated effective heat conduction delay parameters to form its historical time series dataset;

[0059] Based on historical time series datasets, the relative deviation between the current effective heat conduction delay parameter and the preset baseline value under the health status of the equipment is quantified;

[0060] Perform trend analysis on historical time series datasets to calculate and quantify the rate of increase of the effective heat conduction delay parameter;

[0061] If either the relative deviation or the rate of increase exceeds the preset alarm threshold within M consecutive calculation cycles, it is diagnosed that there is a structural layering or cavity defect inside the surge arrester that hinders heat flow transmission, where M is an integer greater than 1.

[0062] Specifically, after generating the instantaneous Joule thermal power sequence in the preliminary steps, the execution entity retrieves this sequence from the derived database and reads the surface temperature field array data from the second data buffer to perform cross-period cross-correlation analysis. Cross-period cross-correlation analysis refers to the analysis time window. The frequency is much greater than the power grid frequency cycle, and is usually selected within the range of 1 to 24 hours. The purpose is to capture the slow thermal dynamic process caused by changes in equipment operating conditions or daily changes in ambient temperature, rather than high-frequency electrical signal fluctuations.

[0063] First, the executing entity will execute the complete instantaneous Joule thermal power sequence. The input signal is set. Next, the actuator selects a temperature change sequence during this period from the surface temperature field array data at a measurement point located in the middle region of the arrester valve assembly. This measurement point is typically selected on the surface of the outer sleeve in the middle of the arrester valve assembly because this location is most sensitive to changes in the internal heat source. The selected temperature time series is set as the response signal.

[0064] Subsequently, the executing entity performs cross-correlation function calculations on the input and response signals over a pre-defined, relatively long time window, such as several hours. This calculation is achieved by shifting one signal in the time domain, then multiplying it point-by-point with the other signal and summing or integrating the results to obtain a value describing the time lag between the two signals. Cross-correlation function of similarity After the calculation is complete, the system searches for the global maximum value, or peak point, on the result curve of the function. The horizontal axis corresponding to this peak point represents the time lag. The effective thermal conduction delay parameter is defined and recorded by the system as the effective thermal conduction delay parameter within the current time window. The effective thermal conduction delay parameter is a macroscopic physical quantity that comprehensively reflects the equivalent thermal conductivity from the valve plate to the outer sleeve surface inside the surge arrester. Any internal structural changes that lead to increased thermal resistance, such as material aging, interface delamination, or moisture and icing, will increase its value.

[0065] Finally, the executing entity stores the newly calculated effective heat conduction delay parameters into a long-term trend database. Subsequently, the system performs a two-dimensional diagnostic assessment: First, it performs a state deviation assessment. The system compares the latest effective heat conduction delay parameters with a preset benchmark value stored in the database that represents the health status of the equipment, and calculates the relative deviation between the two. The preset benchmark value is obtained by measuring and averaging the values ​​after at least one complete thermal stabilization cycle (e.g., 48 hours) when the surge arrester is confirmed to be in a healthy state during factory testing or initial commissioning.

[0066] Secondly, the system assesses the evolution of trends. It retrieves historical data points from the long-term trend database within a preset time window (e.g., the past 90 days), performs trend analysis on the time series (e.g., least squares linear regression), and calculates the rate of increase that characterizes the change.

[0067] The system comprehensively judges two indicators, relative deviation and increase rate, based on a preset diagnostic rule base. If the relative deviation exceeds its corresponding diagnostic threshold (e.g., a typical setting range of 15% to 25%), or the increase rate exceeds its corresponding diagnostic threshold (e.g., a typical setting range of 0.05 to 0.15 seconds / day), the execution entity generates a diagnostic flag, determining that there is a structural delamination or cavity defect inside the surge arrester that hinders heat flow, and records an alarm event. The diagnostic thresholds are set based on the aging model of this type of surge arrester or laboratory accelerated aging test data.

[0068] The formula for calculating the effective heat conduction delay parameter is:

[0069]

[0070] Among them, the cross-correlation function Defined as:

[0071]

[0072] In the formula, This represents the effective heat conduction delay parameter, in seconds. Representing the time lag variable, this is the independent variable used to time-shift the response signal in cross-correlation calculations; its physical meaning is the assumed heat conduction time. The search range is estimated based on the physical dimensions of the surge arrester and the thermal diffusivity of its materials, and this range must cover all possible aging states. For example, for a device with a health value of 60s, the search range can be set to [0s, 300s]. This indicates finding the variable that maximizes the subsequent function expression. The value of . It is an instantaneous Joule thermal power sequence. In cross-correlation calculations, the original response signal is represented. It was moved forward on the timeline. The signal waveform after a time unit. This is the length of the data time window used for cross-correlation analysis. To ensure the reliability of the calculation results and cover typical variations in equipment operating conditions, its value is usually set to several hours, for example, 3 hours. It should be noted that to ensure the validity of the cross-correlation calculation, the total duration of the collected instantaneous Joule thermal power sequence and surface temperature field data is crucial. Should meet , This represents the maximum search latency.

[0073] For example, continuing from the previous step, the system uses an instantaneous Joule thermal power sequence with an average value of 24.5W as the input signal. Simultaneously, the system selects temperature data from sensor 4 (located in the middle of the surge arrester) from the surface temperature field array data as the response signal. Assume that at time T0, due to increased load, the average value of the input signal jumps from 24.5W to 35.0W and remains thereafter. The system continuously monitors the response signal, observing a significant upward trend in its reading starting 0.5 hours after T0, followed by stabilization. The system extracts a data window from time T0 to 3 hours after T0 (a total of 3 hours) and performs cross-correlation calculations on the input and response signals. The calculated cross-correlation function... The peak value occurs at 75 seconds. Therefore, the system determines the effective heat conduction delay parameter for this calculation to be 75 seconds. The system queries the database and finds that the preset baseline value for the surge arrester during initial operation is 60 seconds, and the alarm threshold is set at 20%. The current parameter deviation is... Since 25% exceeds the 20% threshold, the system determines that the insulation layer inside the surge arrester may have deteriorated, and then generates a high-priority warning message: "Increased structural thermal resistance detected, suspected internal moisture and icing or material delamination."

[0074] S4. Perform time-frequency domain analysis on the broadband mechanical vibration signal and compare it with the reference vibration data under healthy conditions to identify the set of new mechanical resonance characteristic frequencies caused by material deterioration or microcracks.

[0075] In a preferred embodiment, identifying the new set of mechanical resonance characteristic frequencies caused by material degradation or microcracks includes: applying wavelet packet transform or short-time Fourier transform to the broadband mechanical vibration signal to obtain a time-frequency domain distribution map.

[0076] By comparing the time-frequency domain distribution map with the reference vibration map under healthy conditions, noise bases that are absent or have lower energy than the reference vibration map, and narrowband spectral components with energy amplitudes exceeding the diagnostic threshold set based on the reference vibration map and fixed frequency positions are identified.

[0077] Extract the center frequency of the narrowband spectral component and construct a set of mechanical resonant characteristic frequencies consisting of one or more center frequencies.

[0078] Specifically, after acquiring multidimensional synchronous operation data, the execution entity processes the broadband mechanical vibration signals stored in the third data buffer to identify new mechanical resonance characteristics caused by material degradation or microcracks.

[0079] First, the system reads a representative broadband mechanical vibration signal time series from the buffer, for example, data of 1 second in length. Next, the system applies a wavelet packet transform algorithm to this signal sequence, decomposing it to a preset depth to obtain multiple sub-band signals with uniform bandwidth covering the entire signal frequency band. Wavelet packet transform is a time-frequency analysis tool that, compared to short-time Fourier transform, provides finer frequency resolution for the high-frequency components of the signal, making it particularly suitable for identifying high-frequency, narrow-band resonant signals generated by microcracks. The wavelet basis function used can be the Daubechies 'db4' wavelet, with the decomposition level set to 5 to 8 layers to obtain kHz-level frequency resolution in the high-frequency bands of interest. By calculating the energy of each sub-band signal within a continuous time window, the system constructs a time-frequency domain distribution map of the current state. This map is a two-dimensional matrix where row indices correspond to frequencies, column indices correspond to time, and matrix elements are signal energy.

[0080] Subsequently, the executing entity retrieves a pre-stored reference vibration spectrum representing the arrester's healthy state from the local model library. This reference vibration spectrum, obtained during the initial commissioning of the equipment or through extensive statistical analysis of similar equipment, represents the vibration energy distribution caused solely by background noise and normal operation when the equipment is free of obvious defects. The system performs a differential analysis between the currently calculated time-frequency domain distribution spectrum and the reference vibration spectrum to identify narrowband spectral components that are absent in the reference vibration spectrum or have energy below the noise floor, but whose energy amplitude exceeds a preset diagnostic threshold in the current time-frequency domain distribution spectrum. The specific identification logic is as follows: the system iterates through each time-frequency unit in the current time-frequency domain distribution spectrum. If the energy amplitude of a unit not only exceeds the preset diagnostic threshold corresponding to its own frequency but also has an energy amplitude at the noise floor level at the same position in its corresponding reference vibration spectrum, then that unit is marked as an anomaly. When these anomalies exhibit clustering on the frequency axis, forming a narrowband with a relatively fixed frequency position, this narrowband is ultimately identified as a narrowband spectral component caused by a potential defect.

[0081] The diagnostic threshold is a frequency-related absolute energy amplitude threshold, set based on a large amount of laboratory fault simulation test data or historical fault data of the same type of equipment. Its purpose is to distinguish between real, defect-induced narrowband resonant signals and random background noise fluctuations. Only when the energy amplitude at a certain frequency point in the current spectrum exceeds the threshold corresponding to that frequency is it preliminarily determined to be a potential anomaly.

[0082] The noise floor level is also a frequency-related energy amplitude threshold. It defines the upper limit of energy in the reference vibration spectrum that is considered pure background noise under healthy equipment conditions. Its setting method is similar to that of diagnostic thresholds, typically based on statistical analysis of a large number of reference vibration spectra under healthy conditions (e.g., calculating the mean energy at each frequency point plus three times the standard deviation). In the diagnostic logic, if the energy at a certain frequency position in the reference spectrum is lower than this noise floor level, it is considered that there is no valid signal at that position under healthy conditions.

[0083] For each identified narrowband spectral component, the system further calculates its energy-weighted center frequency. The formula for calculating the center frequency is:

[0084]

[0085] in, The energy-weighted center frequency representing the extracted narrowband spectral component is an element in the set of mechanical harmonic characteristic frequencies, and its unit is Hertz (Hz). It is the number of discrete frequency points contained in an identified narrowband spectral component. It is the first in this frequency band The frequency values ​​at discrete frequency points are expressed in Hz. This represents the index of a discrete frequency point within the narrowband spectral component, with values ​​ranging from 1 to N. It corresponds to the frequency point. The power spectral density value at that point is the energy amplitude in the time-frequency domain distribution spectrum, reflecting the energy intensity of the signal at that frequency point, and is expressed in units of (m / s²)² / Hz.

[0086] Finally, the system integrates all calculated center frequency values ​​into a set, constructing a mechanical resonant characteristic frequency set consisting of one or more frequency values, and stores this set as the "locked frequency" for subsequent steps in the intermediate results database. The mechanical resonant characteristic frequency set is a dynamically updated list that directly reflects the current state of the arrester's physical structure and serves as a crucial bridge connecting mechanical damage and electrical signal modulation.

[0087] For example, following the previous steps, the system reads a broadband mechanical vibration signal with a sampling rate of 5 MS / s from the third data buffer. The system processes the signal using a 5-layer 'db4' wavelet packet transform, obtaining 32 sub-bands, each with a bandwidth of 78.125 kHz. The system generates a time-frequency domain distribution map for the current period and retrieves the reference vibration map of the arrester under healthy conditions from the database. During the comparison, the system finds that in the 13th frequency band, from 937.5 kHz to 1015.625 kHz, the value in the energy gain matrix remains consistently around 9.5, far exceeding the set threshold of 8. The system identifies this frequency band as an abnormally narrow spectral component. Subsequently, the system performs a fine spectral analysis of the signal energy within this frequency band and finds that at a certain frequency point... Energy at the location for , Energy at the location for , Energy at the location for The energy at other frequencies is negligible. The center frequency is calculated using the formula: After scanning the entire spectrum, no other abnormal frequency bands were found. Therefore, the final set of mechanical resonant characteristic frequencies generated by the system is { The frequency value is then passed to the next step for lock detection.

[0088] S5. Based on the mechanical resonance characteristic frequency set, the instantaneous value of the end current is modulated and sideband locked to identify the electromechanical coupling damage characteristics that cause the current modulation effect.

[0089] In a preferred embodiment, identifying electromechanical coupling damage characteristics that induce current modulation effects includes: using the center frequency contained in the mechanical resonance characteristic frequency set as the carrier center frequency and the power frequency fundamental frequency as the frequency offset, calculating the target search frequency point;

[0090] Spectral analysis is performed on the instantaneous value of the end current to detect the presence of sideband components with amplitudes exceeding the noise floor only at the target search frequency point;

[0091] If a sideband component is detected, it confirms that the mechanical crack corresponding to the mechanical resonance characteristic frequency set has periodically modulated the internal conductive channel, generating electromechanical coupling damage characteristics.

[0092] Specifically, after constructing the mechanical resonant characteristic frequency set in the previous step, the actuator immediately performs modulation sideband lock-in detection on the instantaneous value of the final full current based on this frequency set. Modulation sideband lock-in detection is a highly efficient signal detection method that utilizes a known modulation frequency. To locate the modulated carrier signal The sidebands on the side avoid blindly searching in broadband noise, thereby improving the signal-to-noise ratio and accuracy of detection.

[0093] First, the execution unit reads the mechanical resonance characteristic frequency set from the intermediate result database and retrieves the time series of instantaneous full current values ​​at the end, which were acquired synchronously with the vibration signal, from the first high-speed buffer. The system sets the power grid's fundamental frequency. This is the frequency offset. The fundamental power frequency is the nominal frequency of the power transmission system, such as 50Hz or 60Hz. For each center frequency included in the mechanical resonance characteristic frequency set... The system uses this center frequency as the carrier center frequency to calculate a pair of target search frequency points that need to be locked, i.e., the upper sideband frequency. and lower sideband frequency .

[0094] The formula for calculating the target search frequency point is:

[0095]

[0096] in, This represents the calculated target search frequency point, measured in Hertz (Hz), and includes the upper sideband frequency. and lower sideband frequency Two values.

[0097] Subsequently, the execution unit performs spectral analysis on the data segment of the instantaneous total current at the terminal. Spectral analysis is achieved by using long-time data windows (e.g., data longer than 10 seconds) in conjunction with window functions with high sidelobe suppression (such as the Blackman-Harris window), ensuring that extremely weak sideband signals can be distinguished from the strong power frequency fundamental wave. This analysis is not a full-band scan, but rather employs algorithms such as the Chirp-Z transform (CZT) to concentrate computational resources on a very narrow frequency band around the target search frequency point, achieving microhertz-level frequency resolution. The system extracts the target search frequency point from the calculated spectrum. and The spectral amplitude at that location.

[0098] Next, the system compares the extracted amplitude with a pre-set noise floor threshold. The noise floor threshold is set based on long-term statistical analysis of the spectral amplitude of the frequency band near the target frequency point (excluding known harmonic frequencies) during healthy equipment operation. The average amplitude is then multiplied by a safety factor (e.g., 3 to 5 times) to define the boundary between normal noise fluctuations and abnormal signals. If the amplitude at any target search frequency point exceeds this threshold, the system determines that a sideband component has been detected. Once this component is detected, the system confirms that the mechanical crack corresponding to the mechanical resonance characteristic frequency set has periodically modulated the internal conductive path and immediately generates an electromechanical coupling damage feature with a Boolean value of "true" or a status code of "1" to clearly identify that the crack possesses electrical activity. The electromechanical coupling damage feature is a qualitative diagnostic indicator; its appearance is a key basis for determining whether a mechanical crack has developed to a dangerous stage, as it directly proves that an interaction channel has been established between the mechanical damage and the electrical path. If the amplitude at all target search frequency points does not exceed the threshold, the generated feature is "false" or "0".

[0099] For example, continuing from the previous step, the set of mechanical resonant characteristic frequencies received by the system is { The system is set to the fundamental frequency of the power line. The frequency is 50Hz. First, the system uses... Given the carrier center frequency, calculate the target search frequency point. The upper sideband frequency is... The lower sideband frequency is Next, the system performs a spectral analysis on the instantaneous value of the total current at the terminal, focusing the analysis on two frequency points. The analysis results show that the average noise floor amplitude near the frequencies of 990086Hz and 989986Hz is 0.002mA. The system sets the noise floor threshold to four times the average floor, i.e., 0.008mA. Detection revealed a spectral amplitude of 0.015mA at 990086Hz and 0.013mA at 989986Hz. Since both 0.015mA and 0.013mA are greater than the threshold of 0.008mA, the system determines that a sideband component has been detected. Therefore, the system ultimately generates an electromechanical coupling damage characteristic and sets its status to "confirmed," indicating that the frequency is... The mechanical resonant source (i.e., microcrack) has produced a measurable modulation effect on the internal current path of the surge arrester, exhibiting high-risk electrical activity.

[0100] S6. Based on the evolution trend of the comprehensive effective thermal conduction delay parameters and the detection results of electromechanical coupling damage characteristics, a deterioration trend prediction report is generated, which includes the structural thermal resistance state and the electrical risk level of the crack.

[0101] In a preferred embodiment, a degradation trend prediction report including structural thermal resistance status and crack electrical risk level is generated, including: based on the increase rate of effective heat conduction delay parameter, quantitatively assessing the degree of aging and delamination of the internal materials of the surge arrester or the heat insulation effect caused by moisture and icing, and generating structural thermal resistance status.

[0102] Based on the presence of electromechanical coupling damage characteristics, we can distinguish between purely mechanical structural changes and dangerous cracks that have already induced current modulation.

[0103] Based on the combined evaluation results of structural thermal resistance state and electromechanical coupling damage characteristics, a graded degradation trend prediction signal is output.

[0104] For states exhibiting electromechanical coupling damage characteristics, an insulation breakdown warning is triggered and written into the degradation trend prediction report.

[0105] Specifically, after obtaining long-term data on effective heat conduction delay parameters and the latest detection results of electromechanical coupling damage characteristics in the preceding steps, the executing entity performs a final comprehensive diagnosis and trend prediction. First, the executing entity accesses the long-term trend database and extracts all historical effective heat conduction delay parameters of the target surge arrester over a past period (e.g., the last 90 days), forming a time series. The system uses the least squares method to perform linear regression analysis on this time series, calculating the slope of the trend line as the series evolves over time. This slope represents the rate of increase of the effective heat conduction delay parameters. Based on this rate value, the system refers to a pre-set degradation assessment rule library to quantitatively assess the degree of aging and delamination of the internal materials of the surge arrester or the insulation effect caused by moisture and icing, and outputs the corresponding structural thermal resistance status, such as "stable," "slowly increasing," or "accelerated degradation."

[0106] The formula for calculating the rate of increase of the effective heat conduction delay parameter is:

[0107]

[0108] in, The rate of increase of the effective heat conduction delay parameter is expressed in seconds per day (s / day). It is the number of historical data points used for trend analysis. It is the first The time coordinates of each data point are usually in days. It is in time The measured effective heat conduction delay parameter is expressed in seconds (s).

[0109] The structural thermal resistance state is classified into at least three levels. For example, when the rate of increase is less than 0.05 s / day, it is defined as "stable"; when the rate of increase is between 0.05 s / day and 0.15 s / day, it is defined as "slow growth"; and when the rate of increase is greater than 0.15 s / day, it is defined as "accelerated degradation". These thresholds were determined based on extensive accelerated aging tests and statistical analysis of long-term operating data from similar surge arresters.

[0110] Simultaneously, the executing entity reads the Boolean state value of the electromechanical coupling damage characteristic generated in the previous step. If the value is true, the system determines that the existing mechanical crack has affected the internal conductive channel and is a dangerous crack that has triggered current modulation; if it is false, the system determines that the current structural state change is purely mechanical and the electrical risk is low.

[0111] Finally, the executing entity combines the existence of structural thermal resistance status and electromechanical coupling damage characteristics based on a two-dimensional decision matrix to generate a final degradation trend prediction report (kV). The degradation trend prediction report is generated in a standard data format (such as JSON or XML) and includes fields such as device ID, timestamp, structural thermal resistance status level, crack electrical risk level, and key raw data (such as growth rate and sideband amplitude), facilitating archiving and access by upper-level applications.

[0112] The electrical risk level of the crack is divided into two levels: "No electrical risk" corresponds to a false electromechanical coupling damage characteristic, and "High electrical risk" corresponds to a true electromechanical coupling damage characteristic. Specifically, for any state where an electromechanical coupling damage characteristic is detected, the system will automatically trigger a high-level insulation breakdown early warning signal, which will be sent to the monitoring master station via an industrial communication protocol. The high-level insulation breakdown early warning is a message with a specific code sent via standard protocols such as Modbus TCP or IEC61850, used to trigger audible and visual alarms and emergency work order processes in the monitoring system.

[0113] For example, following the example results from all previous steps, the system has confirmed that the electromechanical coupling damage characteristic is "confirmed" (Boolean value is true). Now, the system retrieves historical data of the effective heat conduction delay parameters of the surge arrester for the past 90 days from the long-term trend database, including 60.1s on day 1, 65.2s on day 30, 70.5s on day 60, and 75s measured on day 90 (currently). The system applies least squares linear regression analysis to these 90 data points to calculate the slope of the trend line. It is 0.16s / day. According to the preset rules, because... If the thermal resistance value exceeds a threshold of 0.15 s / day, the system assesses the structural thermal resistance state as "accelerated degradation." Next, the system combines the two assessment results: the structural thermal resistance state is "accelerated degradation," and the electromechanical coupling damage characteristic is true. According to the two-dimensional decision matrix, this combination corresponds to the highest risk level. Therefore, the system generates a degradation trend prediction report in JSON format, with the following content: {"DeviceID":"MOA-220kV-01","Timestamp":"2023-11-20T14:30:00Z","ThermalState":"Accelerated Degradation","ThermalTrendSlope":0.16,"CrackRiskLevel":"High Electrical Risk","Recommendation":"Immediate Power Outage and Repair"}. Simultaneously, due to the crack electrical risk level being "High Electrical Risk," the system immediately triggers a high-priority insulation breakdown warning, sending an emergency alarm message to the dispatch center's SCADA system.

[0114] Example 2

[0115] Please see Figure 2 As shown, based on Embodiment 1, the second aspect of the present invention provides a system for predicting the degradation trend of an extreme cold surge arrester based on multi-physics field coupling, comprising: a multi-dimensional data synchronous acquisition module, a thermal path parameter extraction module, a heat conduction time delay analysis module, a mechanical resonance feature extraction module, an electromechanical coupling damage detection module, and a comprehensive degradation trend prediction module.

[0116] The multidimensional data synchronization acquisition module acquires multidimensional synchronous operation data of the target surge arrester in extremely cold environments. The multidimensional synchronous operation data includes the instantaneous value of the bus voltage, the instantaneous value of the terminal full current, the surface temperature field array data distributed along the axial direction, and the broadband mechanical vibration signal at the base flange.

[0117] The thermal circuit parameter extraction module, based on the instantaneous values ​​of the bus voltage and the terminal full current, deconstructs the instantaneous Joule thermal power sequence of the internal valve plate of the surge arrester, which serves as the thermal power excitation source for thermal structure coupling analysis.

[0118] The heat conduction delay analysis module uses data from at least one temperature measurement point located in the middle region of the arrester valve plate group selected from the instantaneous Joule thermal power sequence and surface temperature field array data to perform cross-cycle cross-correlation analysis and calculate the effective heat conduction delay parameter characterizing the integrity of the internal structure.

[0119] The mechanical resonance feature extraction module performs time-frequency domain analysis on broadband mechanical vibration signals and compares them with reference vibration data under healthy conditions to identify the new set of mechanical resonance feature frequencies caused by material deterioration or microcracks.

[0120] The electromechanical coupling damage detection module performs modulation sideband locking detection on the instantaneous value of the end current based on the mechanical resonance characteristic frequency set, and identifies the electromechanical coupling damage characteristics that cause the current modulation effect.

[0121] The degradation trend prediction module integrates the evolution trend of effective heat conduction delay parameters with the detection results of electromechanical coupling damage characteristics to generate a degradation trend prediction report that includes the structural thermal resistance state and the electrical risk level of cracks.

[0122] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for predicting the degradation trend of surge arresters in extreme cold weather based on multi-physics coupling, characterized in that, include: S1. Acquire multi-dimensional synchronous operation data of the target surge arrester under extremely cold environment. The multi-dimensional synchronous operation data includes the instantaneous value of bus voltage, the instantaneous value of terminal total current, the surface temperature field array data distributed along the axial direction, and the broadband mechanical vibration signal at the base flange. S2. Based on the instantaneous values ​​of the bus voltage and the total current at the end, the instantaneous Joule thermal power sequence of the internal valve plate of the surge arrester is deconstructed and used as the thermal power excitation source for thermal structure coupling analysis. S3. Perform cross-period cross-correlation analysis using data from at least one temperature measurement point located in the middle region of the arrester valve plate group selected from the instantaneous Joule thermal power sequence and surface temperature field array data to calculate the effective heat conduction delay parameter characterizing the integrity of the internal structure. S4. Perform time-frequency domain analysis on broadband mechanical vibration signals and compare them with reference vibration data under healthy conditions to identify the set of new mechanical resonance characteristic frequencies caused by material deterioration or microcracks. S5. Based on the mechanical resonance characteristic frequency set, the instantaneous value of the end current is modulated and sideband locked to identify the electromechanical coupling damage characteristics that cause the current modulation effect. S6. Based on the evolution trend of the comprehensive effective thermal conduction delay parameters and the detection results of electromechanical coupling damage characteristics, a deterioration trend prediction report is generated, which includes the structural thermal resistance state and the electrical risk level of the crack.

2. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 1, characterized in that, Acquire multi-dimensional synchronous operation data of target surge arresters in extremely cold environments, including: In low-temperature environments where the ambient temperature is below a preset temperature threshold, the instantaneous values ​​of the bus voltage and the instantaneous values ​​of the terminal current are synchronously collected through a high-voltage divider and a through-type current transformer. Activate the fiber optic grating sensor array or non-contact infrared temperature measurement module arranged on the surface of the surge arrester jacket to collect the surface temperature field array data with location tags. A high-frequency acoustic emission sensor coupled to the grounding flange of the surge arrester is used to continuously record waves and capture broadband mechanical vibration signals containing background noise.

3. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 1, characterized in that, The instantaneous Joule thermal power sequence of the internal varistor of the surge arrester was deconstructed, including: Perform orthogonal decomposition or fast Fourier transform on the instantaneous values ​​of bus voltage and terminal total current; The capacitive component in the instantaneous value of the total current at the end is removed, and the resistive current waveform characterizing the nonlinear conductivity of the valve plate is extracted. The resistive current waveform is used as the basic current component for calculating the heating power.

4. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 3, characterized in that, The instantaneous Joule thermal power sequence of the internal varistor of the surge arrester was deconstructed, and it also included: The instantaneous value of the bus voltage and the resistive current waveform are multiplied in the time domain to generate the original power sequence reflecting the real-time heating state of the surge arrester core. The original power sequence is denoised and smoothed to construct an instantaneous Joule thermal power sequence for subsequent calculation of the thermal response transfer function.

5. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 1, characterized in that, Calculate the effective thermal conduction delay parameters characterizing the integrity of the internal structure, including: The instantaneous Joule thermal power sequence is set as the input signal, and the temperature change of the temperature measurement point located in the middle region of the arrester valve group in the surface temperature field array data is set as the response signal. Calculate the cross-correlation function between the input signal and the response signal; Search for the time lag corresponding to the peak value of the cross-correlation function, and define this time lag as the effective heat conduction delay parameter.

6. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 1, characterized in that, The process of generating a degradation trend prediction report also includes: Continuously monitor and record the effective heat conduction delay parameters obtained from periodic calculations to form its historical time series dataset; Based on historical time series datasets, the relative deviation between the current effective heat conduction delay parameter and the preset baseline value under the health status of the equipment is quantified; Perform trend analysis on historical time series datasets to calculate and quantify the rate of increase of the effective heat conduction delay parameter; If either the relative deviation or the rate of increase exceeds the preset alarm threshold within M consecutive calculation cycles, it is diagnosed that there is a structural layering or cavity defect inside the surge arrester that hinders heat flow transmission, where M is an integer greater than 1.

7. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 1, characterized in that, Identify the new set of mechanical resonance characteristic frequencies caused by material degradation or microcracks. include: Wavelet packet transform or short-time Fourier transform are applied to broadband mechanical vibration signals to obtain time-frequency domain distribution maps; By comparing the time-frequency domain distribution map with the reference vibration map under healthy conditions, noise bases that are absent or have lower energy than the reference vibration map, and narrowband spectral components with energy amplitudes exceeding the diagnostic threshold set based on the reference vibration map and fixed frequency positions are identified. Extract the center frequency of the narrowband spectral component and construct a set of mechanical resonant characteristic frequencies consisting of one or more center frequencies.

8. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 7, characterized in that, Identify electromechanical coupling damage characteristics that induce current modulation effects, including: The target search frequency point is calculated using the center frequency contained in the mechanical resonance characteristic frequency set as the carrier center frequency and the power frequency fundamental frequency as the frequency offset. Spectral analysis is performed on the instantaneous value of the end current to detect the presence of sideband components with amplitudes exceeding the noise floor only at the target search frequency point; If a sideband component is detected, it confirms that the mechanical crack corresponding to the mechanical resonance characteristic frequency set has periodically modulated the internal conductive channel, generating electromechanical coupling damage characteristics.

9. The method for predicting the degradation trend of extreme cold surge arresters based on multi-physics coupling according to claim 6, characterized in that, Generate a degradation trend prediction report that includes the structural thermal resistance state and the electrical risk level of cracks, including: Based on the rate of increase of the effective heat conduction delay parameter, the degree of aging and delamination of the internal materials of the surge arrester or the heat insulation effect caused by moisture and icing is quantitatively evaluated to generate the structural thermal resistance state. Based on the presence of electromechanical coupling damage characteristics, we can distinguish between purely mechanical structural changes and dangerous cracks that have already induced current modulation. Based on the combined evaluation results of structural thermal resistance state and electromechanical coupling damage characteristics, a graded degradation trend prediction signal is output. For states exhibiting electromechanical coupling damage characteristics, an insulation breakdown warning is triggered and written into the degradation trend prediction report.

10. A system for executing the method for predicting the degradation trend of extreme cold surge arresters based on multiphysics coupling as described in any one of claims 1-9, characterized in that, include: The multi-dimensional data synchronization acquisition module acquires multi-dimensional synchronous operation data of the target surge arrester in extremely cold environments. The multi-dimensional synchronous operation data includes the instantaneous value of the bus voltage, the instantaneous value of the terminal full current, the surface temperature field array data distributed along the axial direction, and the broadband mechanical vibration signal at the base flange. The thermal circuit parameter extraction module, based on the instantaneous values ​​of the bus voltage and the terminal full current, deconstructs the instantaneous Joule thermal power sequence of the internal valve plate of the surge arrester, which serves as the thermal power excitation source for thermal structure coupling analysis. The heat conduction delay analysis module uses data from at least one temperature measurement point located in the middle region of the arrester valve plate group selected from the instantaneous Joule thermal power sequence and surface temperature field array data to perform cross-cycle cross-correlation analysis and calculate the effective heat conduction delay parameters characterizing the integrity of the internal structure. The mechanical resonance feature extraction module performs time-frequency domain analysis on broadband mechanical vibration signals and compares them with reference vibration data under healthy conditions to identify the new set of mechanical resonance characteristic frequencies caused by material deterioration or microcracks. The electromechanical coupling damage detection module performs modulation sideband locking detection on the instantaneous value of the end current based on the mechanical resonance characteristic frequency set, and identifies the electromechanical coupling damage characteristics that cause the current modulation effect. The comprehensive degradation trend prediction module integrates the evolution trend of effective heat conduction delay parameters with the detection results of electromechanical coupling damage characteristics to generate a degradation trend prediction report that includes the structural thermal resistance state and the electrical risk level of cracks.

Images