A non-contact detection method and system for the aging state of a surge arrester

Abstract

This invention relates to the field of fault prediction and health management, and discloses a non-contact detection method and system for the aging state of surge arresters. The method includes the following steps: acquiring the multi-node magnetic field and thermal radiation baseline of the surge arrester to complete temperature drift compensation; constructing a spatial magnetic field gradient tensor matrix by combining node weights, and decoupling the near-field resistive current by stripping away far-field uniform interference through phase-sensitive detection; extracting the magnetic field and surface thermal gradient distortion pointing vectors, calculating cosine similarity, and performing spatial cross-phase-locked verification to eliminate false anomalies; constructing a thermo-electrodynamic phase space using resistive current and temperature change rate, and extracting the topological geometric distortion parameters of the limiting hysteresis loop; using the parameters for feedback adjustment of differential calculation weights, and mapping them to the Weibull degradation equation for inverse solution of the remaining effective lifetime. This invention achieves high-purity signal reconstruction and homology-based false alarm prevention in complex environments, constructing a closed-loop system that combines adaptive detection and accurate lifetime prediction.

Description

Technical Field

[0001] This invention relates to the field of fault prediction and health management technology, specifically to a non-contact detection method and system for the aging state of surge arresters. Background Technology

[0002] Metal oxide surge arresters are devices used in power systems to limit overvoltages. During long-term operation, the insulation performance of the internal varistor of the surge arrester will age due to the influence of operating voltage, switching overvoltage, and external environmental factors. Insulation aging leads to an increase in resistive leakage current inside the surge arrester and generates Joule heating. If not detected in time, it may cause thermal breakdown of the equipment. Therefore, condition monitoring of the surge arrester's operating status is a basic requirement for equipment operation and maintenance.

[0003] Currently, surge arrester condition monitoring methods are mainly divided into two categories: contact and non-contact. Contact monitoring requires a sensor to be connected in series in the surge arrester's grounding circuit. This method requires a power outage during installation and maintenance, and the sensor, connected in series in the discharge channel, is easily damaged when the surge arrester operates. Non-contact monitoring technologies do not require changes to the equipment wiring and mainly include infrared thermography and spatial magnetic field detection.

[0004] In existing technologies, non-contact detection has certain limitations. Infrared thermometry mainly monitors the surface temperature distribution of surge arresters, but there is a physical time delay in the conduction of heat from the internal varistor to the external surface. Unilateral sunlight and wind speed differences in the external environment can cause uneven surface temperature distribution of the surge arrester, easily leading to false alarms in monitoring systems based on temperature differences. Spatial magnetic field detection technology derives leakage current by measuring the magnetic field around the equipment. Substations have high-voltage busbars and equipment, and strong electromagnetic interference from the background electric field and transmission lines. Conventional magnetic field sensors are susceptible to measurement errors due to ambient temperature drift, and conventional filtering algorithms struggle to separate the uniform interference from distant spatial interference from the weak leakage current signal in the near field of the surge arrester in complex electromagnetic environments.

[0005] Furthermore, existing monitoring systems typically rely on static thresholds of a single physical quantity for alarm judgment, failing to establish a correlation model that combines the electromagnetic and thermodynamic characteristics of the equipment. This judgment method cannot effectively distinguish between environmental coupling interference and actual internal defects, makes it difficult to assess the dynamic evolution of insulation degradation, and cannot calculate the remaining effective lifespan of the equipment. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a non-contact detection method and system for the aging state of surge arresters. This solves the problems of existing non-contact detection technologies for surge arresters, which are difficult to effectively extract weak leakage currents in complex electromagnetic environments, are prone to false alarms due to occasional external environmental factors, and lack quantitative assessment and remaining life prediction methods for the nonlinear degradation state of the equipment.

[0007] To achieve the above objectives, the present invention provides the following technical solution: The first aspect of the present invention provides a non-contact detection method for the aging state of surge arresters, comprising: The multi-node spatial magnetic induction intensity vector of multiple spatial nodes around the surge arrester is obtained, and the two-dimensional temperature distribution matrix of the surge arrester surface and the real-time thermal radiation baseline of the environmental location are obtained simultaneously. Temperature drift compensation is performed on the multi-node spatial magnetic induction intensity vector based on the real-time thermal radiation baseline. Combined with the preset node difference calculation weight, a spatial magnetic field gradient tensor matrix is ​​constructed based on the compensated multi-node spatial magnetic induction intensity vector. Based on the spatial magnetic field gradient tensor matrix, far-field uniform interference is decoupled and eliminated, and a near-field resistive current signal characterizing the near-field characteristics of the surge arrester is reconstructed. The spatial magnetic field gradient tensor matrix and the two-dimensional temperature distribution matrix are respectively decomposed and spatially mapped to extract the spatial magnetic field distortion pointing vector and the surface thermal gradient distortion pointing vector. Spatial cross-phase-locked verification is performed by calculating and comparing the degree of overlap between the two in three-dimensional space. After the spatial cross-phase-locked verification is passed, the first-order partial derivative of the two-dimensional temperature distribution matrix in the time dimension is extracted to generate the temperature change rate. The two-dimensional thermo-electrodynamic phase space trajectory is constructed with the near-field resistive current signal and the temperature change rate, and the topological geometric distortion parameters of the thermo-electrodynamic phase space trajectory are extracted. The topological geometric distortion parameters are used as feedback control parameters to dynamically reconstruct the node differential calculation weights, and the time evolution slope of the topological geometric distortion parameters is extracted. The remaining effective life of the surge arrester is calculated in reverse by combining the preset degradation equation, and a fault prediction and health management report containing pre-diagnosis results is generated.

[0008] In a specific implementation, the step of temperature drift compensation based on a real-time thermal radiation baseline works by: acquiring the physical temperature difference between the real-time thermal radiation baseline and the calibration reference temperature, and inputting it into the preset temperature drift coefficient matrix corresponding to each sensing node; and then calculating the dynamic drift compensation amount for the hardware noise floor. This compensation amount is then precisely subtracted from the original vector, eliminating baseline interference from drastic fluctuations in ambient temperature on the accuracy of weak magnetic field sensing.

[0009] In a specific implementation, the core innovative mechanism of constructing the spatial magnetic field gradient tensor matrix by combining preset node difference calculation weights lies in: calculating the spatial difference partial derivatives of the compensated magnetic induction intensity vector of each node along the orthogonal direction of the three-dimensional Cartesian coordinate system, and multiplying each spatial difference partial derivative with the corresponding node difference calculation weight to construct a third-order spatial magnetic field gradient tensor matrix. :

[0010] in, This represents the constructed third-order spatial magnetic field gradient tensor matrix; Represents the compensated magnetic field vector components Along the orthogonal direction of space Spatial difference partial derivatives, subscript and The range of values ​​for all values ​​covers the Cartesian coordinate system. , , Three orthogonal axes; , , These represent the node differential calculation weights corresponding to the three spatial radiation axes.

[0011] The decoupling mechanism of this invention utilizes the physical attenuation characteristics of spatial electromagnetic fields: background electromagnetic interference or electromagnetic interference from distant transmission lines is approximately a uniform and constant field in the near-field space of the surge arrester, with its spatial first-order partial derivative being zero. By extracting the non-zero high-frequency fluctuations and fundamental components in the tensor matrix through phase-sensitive detection, the uniform interference in the far field can be directly stripped away at the mathematical level, thereby accurately reconstructing the near-field resistive current signal characterizing the actual weak leakage inside the surge arrester in a noisy electromagnetic environment.

[0012] In a specific implementation, the step of performing spatial cross-phase-locked verification is innovatively based on the homogeneity of multimodal physical fields: when the valve plates inside the surge arrester are locally aged or damp, leakage current concentration (manifested as spatial magnetic field distortion) and local abnormal heating (manifested as surface thermal gradient distortion) will inevitably occur simultaneously at the same physical location. By solving the matrix eigenvalue equation to extract the magnetic field distortion feature vector and calculating the cosine similarity in three-dimensional space with the thermal gradient direction vector, the spatial overlap of the two physical field distortion vectors can be determined. This effectively eliminates single-source environmental coupling pseudo-anomalies caused solely by uneven sunlight or sudden changes in the external magnetic field, significantly reducing the false alarm rate.

[0013] In a specific implementation, the steps of constructing the phase space trajectory and extracting topological geometric distortion parameters aim to characterize the nonlinear aging features of the equipment. Since valve aging leads to an increase in nonlinear resistivity, a time-phase hysteresis effect is generated between the electric and thermal field responses. This invention maps the reconstructed resistive current and temperature change rate over a single power frequency cycle, forming a closed trajectory exhibiting a limiting hysteresis loop. The loop area and topological eccentricity of this hysteresis loop directly and quantitatively characterize the degree of imbalance in the electro-thermodynamic coupling relationship.

[0014] In a specific implementation, the steps of dynamically reconstructing weights and inversely calculating remaining lifetime are physically significant in forming a digitized adaptive observation beam and achieving equipment-level full lifecycle prediction. By introducing the extracted topological eccentricity as a feedback variable into the weight mapping function, the analytical resolution of the nodes facing the surge arrester is autonomously enhanced. Simultaneously, the time evolution slope is extracted by time-domain fitting of the hysteresis loop area. The time evolution slope reflects the absolute rate of aging and is mapped to the scale parameter of the Weibull degradation equation; the topological eccentricity reflects the non-uniform evolution mode of aging and is mapped to the shape parameter. Combined with the critical threshold of reliability for protection failure, inverse analysis is performed to finally obtain the remaining effective lifetime of the equipment.

[0015] A second aspect of the present invention provides a non-contact detection system for the aging state of surge arresters, comprising: The multi-physics synchronous sensing module is used to acquire the multi-node spatial magnetic induction intensity vector of multiple spatial nodes around the surge arrester, and simultaneously acquire the two-dimensional temperature distribution matrix of the surge arrester surface and the real-time thermal radiation baseline of the environmental location. The temperature drift compensation and interference decoupling module is used to compensate for the temperature drift of the multi-node spatial magnetic induction intensity vector based on the real-time thermal radiation baseline, construct the spatial magnetic field gradient tensor matrix by combining the preset node difference calculation weights, and decouple and eliminate far-field uniform interference to reconstruct the near-field resistive current signal. The spatial cross-phase-locked verification module is used to extract the spatial magnetic field distortion pointing vector and the surface thermal gradient distortion pointing vector respectively, and to perform spatial cross-phase-locked verification by calculating and comparing the degree of overlap between the two in three-dimensional space. The phase space topology analysis module is used to extract the temperature change rate after the judgment is passed, construct a two-dimensional thermo-electrodynamic phase space trajectory, and extract the topological geometric distortion parameters of the trajectory; The feedback closed-loop and prediction evaluation module is used to dynamically reconstruct the node differential calculation weights by using the topological geometric distortion parameters as feedback control parameters, and to calculate the remaining effective life of the surge arrester by combining the degradation equation with reverse solution.

[0016] In a specific implementation, the multi-physics synchronous sensing module includes a composite non-contact detection hardware architecture composed of a regular tetrahedral array built based on a tunneling magnetoresistive sensor and an infrared thermal imager.

[0017] In a specific implementation, the feedback closed loop and prediction evaluation module has a built-in Weibull degradation equation engine, which is configured to nonlinearly map the time evolution slope and topological eccentricity of the extracted hysteresis loop area to the scale parameter and shape parameter of the degradation equation, respectively, so as to realize the dynamic prediction of the remaining effective life of the equipment.

[0018] This invention provides a non-contact detection method and system for the aging condition of surge arresters. It has the following advantages: 1. This invention overcomes the detection bottleneck of weak leakage current extraction under complex electromagnetic backgrounds. It utilizes the physical attenuation characteristic of far-field interference sources, where the spatial first-order partial derivative approaches zero, and constructs a three-dimensional spatial magnetic field gradient tensor matrix using multi-node data for differential decoupling. This mechanism eliminates the need for bulky and expensive physical shielding, effectively removing substation background noise at a purely mathematical analytical level. It reconstructs the near-field resistive current signal characterizing internal defects with high fidelity, significantly improving the anti-interference limit of non-contact detection.

[0019] 2. This invention fundamentally eliminates the problem of false alarms induced by occasional environmental factors. Recognizing the inherent physical similarity between surge arrester degradation and the accompanying "concentrated leakage current and localized heating," this invention introduces a spatial cross-phase-locked verification mechanism. By rigorously verifying the cosine similarity between the magnetic field distortion vector and the surface thermal gradient direction vector in three-dimensional space, it accurately filters out single-physics-field spurious anomalies caused by unilateral solar exposure or occasional external magnetic disturbances, thus giving the system extremely high diagnostic confidence.

[0020] 3. This invention achieves keen detection of nonlinear degradation characteristics and quantitative prediction of remaining life of equipment. It creatively maps near-field current and temperature change rate together to generate a phase-space limiting loop characterizing the electro-thermal hysteresis effect. By extracting its topological eccentricity and area evolution slope and directly mapping them to parameters of the Weibull degradation equation, it can not only keenly detect early minute resistance changes in valve plate aging, but also combine reliability thresholds to calculate accurate remaining effective life, truly propelling surge arresters towards "predictive health management." Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention. Figure 2 This is a schematic diagram of the non-contact detection hardware architecture and spatial topology deployment principle of the present invention; Figure 3 This is the logic diagram for temperature drift compensation calculation based on real-time thermal radiation baseline of the present invention; Figure 4 This is a schematic diagram illustrating the principle of spatial magnetic field gradient tensor construction and phase-sensitive detection decoupling in this invention. Figure 5 This is the logic diagram for preventing false alarms in the multiphysics spatial cross-phase-locked loop verification of the present invention; Figure 6 This is the geometric topology diagram of the two-dimensional thermo-electrodynamic phase space hysteresis loop of the present invention; Figure 7 This is a flow diagram of adaptive feedback and lifetime prediction based on the Weibull equation of the present invention. Detailed Implementation

[0022] The technical solutions in 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.

[0023] Please see the appendix Figure 1 -Appendix Figure 7 This invention provides a non-contact detection method and system for the aging state of surge arresters, comprising the following steps: The multi-node spatial magnetic induction intensity vector of multiple spatial nodes around the surge arrester is obtained, and the two-dimensional temperature distribution matrix of the surge arrester surface and the real-time thermal radiation baseline of the environmental location are obtained simultaneously.

[0024] To achieve accurate acquisition of the aforementioned multiphysics data, this embodiment employs a composite non-contact detection hardware architecture to perform the detection task. This hardware architecture is integrated into the near-field space of the surge arrester and consists of a tetrahedral array based on a tunneling magnetoresistive sensor and an infrared thermal imager.

[0025] The tunneling magnetoresistive sensor possesses extremely high magnetic field resolution and extremely low low-frequency noise floor, making it suitable for capturing the micro-nano Tesla-level alternating magnetic field radiated outward from the weak resistive leakage current inside the surge arrester. To construct a three-dimensional observation perspective with spatial differential capability, this embodiment arranges multiple sensing nodes in a tetrahedral topology.

[0026] Specifically, the tetrahedral array is established in an independently defined local three-dimensional Cartesian coordinate system. Its physical structure consists of a reference sensing node and three measurement sensing nodes. The reference sensing node is located at the origin of the local coordinate system, and the three measurement sensing nodes are respectively located on three mutually orthogonal radiation axes extending from the origin, with each measurement sensing node having the same physical arm length from the origin reference sensing node.

[0027] During synchronous acquisition, each node of the tetrahedral array independently outputs a three-dimensional magnetic field strength vector at its spatial location. The set of multi-node spatial magnetic field strength vectors acquired by the array is defined as follows: , any number of its internal The magnetic field vector of each spatial node is represented as:

[0028] in, The value is the index number of each sensor node in the array; , , These represent the spatial node in the local Cartesian coordinate system. axis, axis, The magnetic flux density components in the three orthogonal directions of the axis.

[0029] While performing magnetic field spatial sampling, the infrared thermal imager in the composite non-contact detection hardware architecture triggers synchronous acquisition with the same system clock stamp. The main optical axis of the infrared thermal imager is aligned with the arrester body to acquire a high-resolution infrared thermal image of the arrester's outer surface, and then converts it into a quantitative two-dimensional temperature distribution matrix through an internal analytical algorithm. Each element in this matrix represents the absolute thermodynamic temperature value of a specific spatial grid on the arrester surface.

[0030] While scanning the surge arrester, the infrared thermal imager extracts temperature data from environmental areas at the edge of the field of view where there is no significant heat source interference, thereby generating a real-time thermal radiation baseline of the environmental location. This baseline not only serves as an absolute reference characterizing the background thermal fluctuations of the external environment, but also provides the necessary physical input parameters for subsequent temperature drift compensation of the underlying hardware of the high-sensitivity magnetic sensor.

[0031] To ensure the correspondence of multimodal data in physical space, an external parameter transformation matrix needs to be established between the local coordinate system of the tetrahedral array and the field-of-view coordinate system of the infrared thermal imager during the system measurement initialization phase. This transformation matrix maps the acquired multi-node spatial magnetic flux density vector and two-dimensional temperature distribution matrix to a global three-dimensional spatial coordinate system with the arrester base as the geometric center, thereby achieving spatial alignment and temporal phase locking of the multiphysics information.

[0032] In this embodiment, after obtaining the multi-node spatial magnetic induction intensity vector of multiple spatial nodes around the surge arrester, the step of performing temperature drift compensation on the multi-node spatial magnetic induction intensity vector based on the real-time thermal radiation baseline specifically includes the following operations.

[0033] When capturing weak alternating magnetic fields, the inherent thermal sensitivity of the underlying sensing element in a highly sensitive tunneling magnetoresistive sensor causes the output baseline to drift to zero as the ambient temperature fluctuates. To prevent this thermally coupled drift from overwhelming the micro-nano Tesla-level near-field signal characterizing the leakage current, the system must perform rigorous baseline dynamic calibration at the pure hardware signal level.

[0034] In practice, the system first extracts the real-time thermal radiation baseline obtained in the preceding steps and acquires the preset calibration reference temperature. This calibration reference temperature is the physical ambient temperature during the initial calibration of the sensor array in a constant-temperature laboratory environment free from magnetic field interference.

[0035] Subsequently, the difference between the real-time thermal radiation baseline and the preset calibration reference temperature is calculated. This physical difference directly reflects the absolute amount of thermodynamic deviation between the current actual substation testing environment and the reference calibration environment.

[0036] For each individual node in the tetrahedral array, the system is equipped with a dedicated preset temperature drift coefficient matrix. This matrix was obtained through multi-temperature zone cyclic calibration experiments and accurately characterizes the linear and nonlinear drift rates of the magnetic field output of the reference sensing node and each measurement sensing node on the three-dimensional orthogonal axes as a function of temperature.

[0037] The calculated temperature difference is multiplied by the preset temperature drift coefficient matrices corresponding to the reference sensing node and the measurement sensing node, respectively, to obtain the three-dimensional magnetic field compensation of each node at the current ambient temperature. The mathematical and physical model of this process is expressed as follows:

[0038] in, Indicates the first The three-dimensional magnetic field compensation of each sensing node; Indicates the first The preset temperature drift coefficient matrix corresponding to each sensing node is a third-order square matrix. This represents the real-time thermal radiation baseline extracted by an infrared thermal imager; This indicates the preset calibration reference temperature.

[0039] After obtaining the corresponding compensation values, a direct subtraction operation is performed on the underlying signals. The calculated compensation value is subtracted from the original acquired multi-node spatial magnetic field strength vector to obtain the compensated multi-node spatial magnetic field strength vector, which is expressed as follows:

[0040] in, This indicates the first time after temperature drift compensation is completed. The spatial magnetic flux density vector of each node; This represents the original multi-node spatial magnetic induction intensity vector.

[0041] Through the above closed-loop calculation, the system completely eliminates the common-mode hardware error introduced by the drastic fluctuations in the external thermal environment, eliminates the interference of temperature variables on weak magnetic sensing, and provides pure, stable and real underlying electromagnetic field source data for the subsequent construction of a high-precision spatial magnetic field gradient tensor.

[0042] In this embodiment, after completing the temperature drift compensation of the underlying signal, the system combines the preset node differential calculation weights, constructs a spatial magnetic field gradient tensor matrix based on the compensated multi-node spatial magnetic induction intensity vector, and decouples and eliminates far-field uniform interference based on the spatial magnetic field gradient tensor matrix to reconstruct the near-field resistive current signal characterizing the near-field characteristics of the surge arrester.

[0043] In practice, the system extracts compensated magnetic field data from three measurement sensing nodes in a tetrahedral array and a reference sensing node located at the origin. Using a fixed physical arm length between the measurement and reference sensing nodes, the system calculates the differential partial derivatives of each magnetic field component along the three orthogonal spatial axes of the local three-dimensional Cartesian coordinate system. To adaptively balance the detection sensitivity from different spatial perspectives, the system multiplies each calculated spatial differential partial derivative with a preset node differential calculation weight.

[0044] Using the weighted spatial difference partial derivatives, the complete third-order spatial magnetic field gradient tensor matrix is ​​constructed, and its mathematical model is expressed as follows:

[0045] in, This represents the constructed third-order spatial magnetic field gradient tensor matrix; Represents the compensated magnetic field vector components Along the orthogonal direction of space Spatial difference partial derivatives, subscript and The range of values ​​for all values ​​covers the Cartesian coordinate system. , , Three orthogonal axes; , , These represent the node differential calculation weights corresponding to the three spatial radiation axes.

[0046] The essence of constructing this matrix lies in utilizing the spatial attenuation law of physical fields in complex electromagnetic environments. In actual substation operation environments, widely distributed distant high-voltage busbars and adjacent electrical equipment generate strong background electromagnetic radiation. However, when these interference waves reach the near-field space of the target surge arrester, their wavefronts have flattened and evolved into a spatially uniform field. The core physical characteristic of this far-field uniform interference source is that its three-dimensional spatial first-order partial derivative is zero. Meanwhile, the source region magnetic field excited by the weak leakage current inside the surge arrester has high non-uniformity, exhibiting drastic gradient changes within an extremely short baseline distance.

[0047] Based on this physical characteristic, the system performs anti-interference decoupling on the constructed spatial magnetic field gradient tensor matrix. Specifically, the system uses a phase-sensitive detection algorithm to synchronously demodulate each weighted difference element within the matrix. Since matrix operations essentially calculate the rate of change in space, the far-field uniform interference component, which has zero partial derivative characteristics, naturally disappears and is completely stripped away in this mathematical space.

[0048] Subsequently, phase-sensitive detection separates the non-zero high-frequency fluctuations and power frequency fundamental components retained in the spatial magnetic field gradient tensor matrix. These non-zero gradient components are purely caused by the distortion of the directional charge movement within the arrester itself. The system extracts these pure non-zero components, performs parametric mapping using an inverse electromagnetic inversion algorithm, and then reconstructs with high fidelity the near-field resistive current signal that directly characterizes the insulation degradation features inside the arrester.

[0049] In this embodiment, in order to fundamentally eliminate false alarms of single physics fields induced by accidental environmental factors, the system performs feature decomposition and spatial mapping on the aforementioned constructed spatial magnetic field gradient tensor matrix and the acquired two-dimensional temperature distribution matrix, extracts the spatial magnetic field distortion pointing vector and the surface thermal gradient distortion pointing vector, and performs spatial cross-phase-locked verification by calculating and comparing the degree of overlap between the two in three-dimensional space.

[0050] Specifically, the system first solves the eigenvalue equations of the spatial magnetic field gradient tensor matrix. The eigenvalues ​​of the matrix characterize the degree of variation of the spatial magnetic field gradient along a specific three-dimensional principal axis. The system extracts the eigenvalue with the largest absolute value using a numerical decomposition algorithm and obtains the corresponding eigenvector. This eigenvector precisely indicates the spatial orientation of the most intense local magnetic field distortion in the near-field region. The system defines and extracts this eigenvector as the spatial magnetic field distortion pointing vector, denoted as... .

[0051] Simultaneously, the system performs spatial difference operations on the two-dimensional temperature distribution matrix acquired by the infrared thermal imager to calculate the two-dimensional spatial physical gradient field on the surface of the surge arrester. Based on this data, the system traverses the gradient field and locates the spatial extremum point where the gradient magnitude reaches its maximum, extracting the two-dimensional direction vector at that point. This vector points to the geometric tangent direction of the fastest outward diffusion of local abnormal heating.

[0052] To eliminate the perspective differences between heterogeneous sensors, the system then invokes the temperature measurement extrinsic parameter matrix preset during system initialization. Using this parameter matrix, the extracted two-dimensional direction vector is subjected to coordinate algebraic transformation, projecting it losslessly from the two-dimensional plane of the infrared field of view to a global three-dimensional coordinate system sharing a common reference with the tetrahedral array. After this round of spatial dimension upscaling and alignment, the system generates a surface thermal gradient distortion pointing vector, denoted as... .

[0053] After extracting the two key vectors characterizing the distortion of the multimodal physical field, the system enters the core false alarm prevention logic. The physical basis of this mechanism is that when the internal varistor of the surge arrester becomes damp or deteriorates, it will inevitably cause a high concentration of leakage current and local abnormal heating at the same weak insulation location. Therefore, the real internal defect source must have a spatial homology between the electric field and the thermal field.

[0054] Based on this shared origin attribute, the system calculates the cosine similarity between the pointing vector of the spatial magnetic field distortion and the pointing vector of the surface thermal gradient distortion in three-dimensional space. The mathematical model is as follows:

[0055] in, Represents cosine similarity in three-dimensional space; This indicates the vector pointing towards the extracted spatial magnetic field distortion. This represents the vector pointing to the surface thermal gradient distortion generated by the mapping; This represents the dot product operation of two spatial vectors; This indicates the calculation of the magnitude of the corresponding vector.

[0056] The system will calculate the cosine similarity parameters. The system performs a logical comparison with a preset spatial overlap verification threshold. When the cosine similarity is greater than or equal to the preset spatial overlap verification threshold, the system determines that the currently detected electromagnetic anomaly source and thermal anomaly source are physically matched, that is, the multi-physics field characteristics of the internal real degradation defects mutually confirm each other. Based on this, the system determines that the spatial cross-phase-locked verification is passed and triggers the subsequent aging depth assessment procedure.

[0057] Conversely, when the calculated cosine similarity is less than the preset threshold, it indicates that the current physical field fluctuations exhibit a non-homogeneous dissipation state. Based on this, the system determines that the current signal fluctuation is an environmental coupling pseudo-anomaly induced by uneven unilateral solar radiation, external airflow heat dissipation disturbance, or sudden magnetic field distortion from afar. The system will directly intercept this alarm event and discard the abnormal data as invalid background noise. This spatial cross-validation mechanism completely breaks the traditional crude logic chain of alarming based on a single physical quantity exceeding its limit.

[0058] In this embodiment, after the spatial cross-phase-locked loop verification is passed, the system enters the stage of fine-grained characterization of the aging depth of the surge arrester. The system first extracts the two-dimensional temperature distribution matrix obtained from the continuous time series, calculates the first-order partial derivative of the matrix in the time dimension, and thus generates the temperature change rate that dynamically characterizes the surface heat dissipation rate.

[0059] During long-term operation and aging, the grain boundary barrier structure of the zinc oxide varistor inside the surge arrester undergoes irreversible degradation, leading to distortion of the material's nonlinear current-voltage characteristics. This microscopic material degradation manifests at the macroscopic physical detection level as a significant time phase hysteresis effect between the resistive leakage current under alternating electric field excitation and the varistor's heat dissipation process.

[0060] To explicitly extract this nonlinear decay feature hidden in the time-domain signal, the system constructs a two-dimensional thermo-electrodynamic phase space. In specific operation, the system uses the reconstructed near-field resistive current signal as the horizontal axis parameter and the generated temperature change rate as the vertical axis parameter, and performs synchronous parametric mapping on these two cross-physical domain data in the time integration domain of a single power frequency cycle.

[0061] During this mapping process, the trajectory state points in phase space can be defined using the following parametric equation model:

[0062] in, Representing time variables The trajectory state point in the two-dimensional thermo-electrodynamic phase space at any given moment; This represents the near-field resistive current signal reconstructed by decoupling from the spatial magnetic field gradient tensor; Indicates time The rate of temperature change generated by the first-order partial derivative at any given time.

[0063] Due to the aforementioned electro-thermodynamic hysteresis effect, the state point of the parametric mapping cannot completely coincide with the initial state at the end of a single complete power frequency cycle, thus evolving into a closed state trajectory in two-dimensional phase space, exhibiting a limiting hysteresis loop. The geometry of this trajectory directly maps the imbalance state of thermo-electric energy conversion inside the surge arrester.

[0064] Subsequently, the system extracts the geometric features of the generated limiting hysteresis loop. The system calculates the loop area enclosed by the hysteresis loop using a surface integral algorithm. The physical essence of this area value represents the additional irreversible heat loss caused by insulation degradation within a single power frequency cycle. Simultaneously, the system calculates and extracts the topological eccentricity of the hysteresis loop to quantitatively characterize the degree of geometric deformation caused by the asymmetric degradation of the valve plate's forward and reverse nonlinear impedances.

[0065] The system ultimately extracts the hysteresis loop area and topology eccentricity together as topology geometric distortion parameters. This parameter set breaks away from the limitations of traditional amplitude monitoring, accurately capturing early signs of equipment degradation from a high-dimensional topology space, and providing high-fidelity data parameter input for subsequent adaptive adjustment of control weights and lifetime prediction.

[0066] In this embodiment, the system uses the extracted topological geometric distortion parameters as feedback control parameters, dynamically reconstructs the aforementioned node differential calculation weights, extracts the time evolution slope of the topological geometric distortion parameters, and combines the preset degradation equation to reverse calculate the remaining effective life of the surge arrester, ultimately generating a fault prediction and health management report containing pre-diagnosis results.

[0067] Specifically, to form a highly digitized adaptive detection beam, the system extracts the topological eccentricity from the topological geometric distortion parameters. This topological eccentricity is then used as a feedback variable in a preset weighted mapping control function. As the eccentricity increases due to worsening local degradation of the valve plate, the control function synchronously calculates and amplifies the node differential calculation weights of the corresponding measurement sensor nodes facing the side of the surge arrester where degradation is more severe. This feedback control mechanism autonomously enhances the analytical resolution in a specific spatial dimension without changing the physical hardware position, ensuring that the sensitivity of the detection beam is always focused on the deteriorated area.

[0068] In a specific embodiment, to achieve precise dynamic adjustment of the weights, the preset weight mapping control function specifically adopts a linear saturation mapping model, the mathematical expression of which is as follows:

[0069] In the formula, This represents the adjusted node differential calculation weights for the nodes facing the surge arrester. This represents the baseline weights during system initialization; This represents the topological eccentricity of the currently extracted limiting hysteresis loop; This indicates the preset threshold for determining eccentricity. This represents the system's preset gain coefficient.

[0070] Furthermore, after obtaining the weight increment facing the surge arrester using the above mapping function, the system will perform normalization processing on the node differential calculation weights of all nodes to ensure the stability of the total calculated weights.

[0071] While implementing adaptive spatial feedback, the system quantitatively assesses the degradation over time. The system retrieves continuously stored hysteresis loop area data from historical operating cycles and performs least-squares fitting on this data set, using the time series as the independent variable. Through fitting, the system extracts the slope of the fitted line reflecting the loop area's growth over time, defining and extracting it as the time evolution slope. This slope data rigorously characterizes the absolute degradation rate of the overall insulation aging of the surge arrester in a physical sense.

[0072] Subsequently, the system invokes the built-in Weibull degradation equation engine to perform lifetime inverse calculation. The Weibull degradation equation is a classic reliability mathematical model characterizing fatigue failure of nonlinear devices. At this stage, the system performs nonlinear mapping between the extracted feature parameters and the key parameters of the degradation equation. Specifically, the system maps the time evolution slope to the scale parameter of the degradation equation and the topological eccentricity to the shape parameter of the degradation equation.

[0073] The mathematical model of the Weibull degeneracy equation after mapping is expressed as follows:

[0074] in, Indicates the device's future operating time The probability of operational reliability at that time; This represents the scale parameter generated by the time evolution slope mapping, and the lifetime time span of the dominant equation; The shape parameters generated by the topological eccentricity mapping represent the degradation failure modes and non-uniform evolution curve characteristics of the dominant equation. This is an operation of the natural exponential function.

[0075] Based on the constructed degradation equation, the system obtains the critical threshold for reliability under protected conditions as preset in the operation and maintenance specifications. The system uses this critical threshold as the target reliability. Substituting into the Weibull degradation equation above, the corresponding critical failure time point is calculated through analytical inversion. Finally, the system subtracts the current service time of the surge arrester from this critical failure time point to accurately obtain the remaining effective life of the surge arrester.

[0076] In the specific calculation, the mathematical expression for obtaining the remaining effective lifetime through analytical inversion is as follows: First, the corresponding critical failure time point is calculated using logarithmic operations. :

[0077] Subsequently, the remaining effective life of the surge arrester was calculated. :

[0078] In the above formula This represents the preset critical threshold for reliability in case of defense failure. Represents the natural logarithm operation; and These are the scale parameters and shape parameters determined by the mapping in the preceding steps, respectively; This indicates the absolute failure time expected when the defense failure reliability critical threshold is reached. This indicates the current service evaluation time of the surge arrester.

[0079] The system summarizes the calculated remaining effective lifespan, dynamically updated phase space distortion trajectory, and electromagnetic field compensation data for each spatial node, automatically generating an equipment-level fault prediction and health management report. This report directly provides the substation operation and maintenance center with clear maintenance timeline suggestions, completely realizing the leap from simple condition monitoring to predictive management of the entire equipment lifecycle in non-contact detection.

[0080] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention. Various changes, modifications, substitutions, and variations made to these embodiments by those skilled in the art without departing from the principles and spirit of the present invention shall fall within the scope of protection of the claims of the present invention.

[0081] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A non-contact detection method for the aging condition of surge arresters, characterized in that, Includes the following steps: The multi-node spatial magnetic induction intensity vector of multiple spatial nodes around the surge arrester is obtained, and the two-dimensional temperature distribution matrix of the surge arrester surface and the real-time thermal radiation baseline of the environmental location are obtained simultaneously. Temperature drift compensation is performed on the multi-node spatial magnetic induction intensity vector based on the real-time thermal radiation baseline. Combined with the preset node difference calculation weight, a spatial magnetic field gradient tensor matrix is ​​constructed based on the compensated multi-node spatial magnetic induction intensity vector. Based on the spatial magnetic field gradient tensor matrix, far-field uniform interference is decoupled and eliminated, and a near-field resistive current signal characterizing the near-field characteristics of the surge arrester is reconstructed. The spatial magnetic field gradient tensor matrix and the two-dimensional temperature distribution matrix are respectively decomposed and spatially mapped to extract the spatial magnetic field distortion pointing vector and the surface thermal gradient distortion pointing vector. Spatial cross-phase-locked verification is performed by calculating and comparing the degree of overlap between the two in three-dimensional space. After the spatial cross-phase-locked verification is passed, the first-order partial derivative of the two-dimensional temperature distribution matrix in the time dimension is extracted to generate the temperature change rate. The two-dimensional thermo-electrodynamic phase space trajectory is constructed with the near-field resistive current signal and the temperature change rate, and the topological geometric distortion parameters of the thermo-electrodynamic phase space trajectory are extracted. The topological geometric distortion parameters are used as feedback control parameters to dynamically reconstruct the node differential calculation weights, and the time evolution slope of the topological geometric distortion parameters is extracted. The remaining effective life of the surge arrester is calculated in reverse by combining the preset degradation equation, and a fault prediction and health management report containing pre-diagnosis results is generated.

2. The non-contact detection method for the aging state of a surge arrester according to claim 1, characterized in that, The steps of acquiring the multi-node spatial magnetic flux density vector of multiple spatial nodes around the surge arrester, and compensating for temperature drift of the multi-node spatial magnetic flux density vector based on the real-time thermal radiation baseline, specifically include: The multi-node spatial magnetic induction intensity vector is obtained by arranging a regular tetrahedral array in the near field space of the surge arrester. The regular tetrahedral array consists of a reference sensing node located at the origin of the local coordinate system and three measurement sensing nodes on three mutually orthogonal radiation axes. The difference between the real-time thermal radiation baseline and the preset calibration reference temperature is obtained. The difference is multiplied by the preset temperature drift coefficient matrix corresponding to the reference sensing node and the measurement sensing node to obtain the compensation amount. The compensation amount is subtracted from the multi-node spatial magnetic induction intensity vector to complete the temperature drift compensation.

3. The non-contact detection method for the aging state of a surge arrester according to claim 1, characterized in that, The steps of combining preset node differential calculation weights, constructing a spatial magnetic field gradient tensor matrix based on the compensated multi-node spatial magnetic induction intensity vector, and decoupling and eliminating far-field uniform interference based on the spatial magnetic field gradient tensor matrix specifically include: Calculate the spatial difference partial derivatives of each component of the compensated multi-node spatial magnetic induction intensity vector along the orthogonal direction of the three-dimensional Cartesian coordinate system, and multiply the spatial difference partial derivatives with the node difference calculation weights to form a third-order spatial magnetic field gradient tensor matrix from the weighted spatial difference partial derivatives; By utilizing the physical attenuation characteristic that the first-order partial derivative of the spatial magnetic field gradient tensor matrix is ​​zero, the non-zero high-frequency fluctuations and fundamental components in the spatial magnetic field gradient tensor matrix are separated by phase-sensitive detection, thus stripping away the far-field uniform interference and reconstructing the near-field resistive current signal.

4. The non-contact detection method for the aging state of a surge arrester according to claim 1, characterized in that, The step of extracting the spatial magnetic field distortion pointing vector and the surface thermal gradient distortion pointing vector, and performing spatial cross-phase-locked verification by calculating and comparing their overlap in three-dimensional space, specifically includes: Solve the eigenvalue equation of the spatial magnetic field gradient tensor matrix, and extract the eigenvector corresponding to the eigenvalue with the largest absolute value as the spatial magnetic field distortion pointing vector; The two-dimensional spatial physical gradient of the two-dimensional temperature distribution matrix is ​​calculated, the direction vector at the maximum gradient magnitude is extracted, and it is mapped to the global three-dimensional coordinate system through a preset temperature measurement extrinsic parameter matrix to generate the surface thermal gradient distortion pointing vector. Calculate the cosine similarity between the spatial magnetic field distortion pointing vector and the surface thermal gradient distortion pointing vector in three-dimensional space. When the cosine similarity is greater than or equal to the preset spatial overlap verification threshold, it is determined that the electromagnetic anomaly source and the thermal anomaly source correspond in physical location, and the spatial cross-phase locking verification is passed; otherwise, it is determined to be an environmental coupling pseudo-anomaly and is removed.

5. The non-contact detection method for the aging state of a surge arrester according to claim 1, characterized in that, The step of constructing a two-dimensional thermo-electrodynamic phase space trajectory using the near-field resistive current signal and the temperature change rate, and extracting the topological geometric distortion parameters of the thermo-electrodynamic phase space trajectory, specifically includes: Using the reconstructed near-field resistive current signal as the horizontal axis parameter and the generated temperature change rate as the vertical axis parameter, parameter mapping is performed in the time integral domain of a single power frequency cycle to generate the thermo-electrodynamic phase space trajectory that exhibits a limiting hysteresis loop. Geometric features are extracted from the limiting hysteresis loop to obtain the loop area and topological eccentricity. The loop area and topological eccentricity are used together as the topological geometric distortion parameter.

6. The non-contact detection method for the aging state of a surge arrester according to claim 5, characterized in that, The step of dynamically reconstructing the node difference calculation weights by using the topological geometric distortion parameters as feedback control parameters specifically includes: The extracted topological eccentricity is introduced as a feedback control variable into a preset weight mapping control function; When the topological eccentricity falls into the preset minimum value judgment interval, the preset weight mapping control function outputs an adjustment command to increase the node differential calculation weight of the node facing the surge arrester among the multiple spatial nodes, and simultaneously decrease the node differential calculation weight of the node away from the surge arrester among the multiple spatial nodes, forming a directional observation gain beam that enhances the spatial analytical resolution of leakage current.

7. The non-contact detection method for the aging state of a surge arrester according to claim 5, characterized in that, The preset degradation equation is the Weibull degradation equation. The step of extracting the time evolution slope of the topological geometric distortion parameters and inversely calculating the remaining effective lifetime of the surge arrester by combining the preset degradation equation specifically includes: Least square fitting is performed on the area of ​​multiple hysteresis loops within a preset time window to extract the temporal evolution slope of the hysteresis loop area. The extracted time evolution slope is used as the input parameter of the preset mapping function to determine the scale parameter of the Weibull degeneracy equation, and the extracted topological eccentricity is used as the input parameter of the preset mapping function to determine the shape parameter of the Weibull degeneracy equation. Using the Weibull degradation equation with determined parameters, the critical threshold for reliability in the defense failure is substituted into the equation for reverse calculation to obtain the expected failure time. The current assessment time is then deducted to obtain the remaining effective lifetime.

8. A non-contact detection system for the aging state of a surge arrester according to claim 1, characterized in that, A non-contact detection system for the aging condition of surge arresters, characterized in that it comprises: The multi-physics synchronous sensing module is used to acquire the multi-node spatial magnetic induction intensity vector of multiple spatial nodes around the surge arrester, and simultaneously acquire the two-dimensional temperature distribution matrix of the surge arrester surface and the real-time thermal radiation baseline of the environmental location. The temperature drift compensation and interference decoupling module is used to compensate for the temperature drift of the multi-node spatial magnetic induction intensity vector based on the real-time thermal radiation baseline, combine the preset node difference calculation weights, construct the spatial magnetic field gradient tensor matrix based on the compensated multi-node spatial magnetic induction intensity vector, and decouple and eliminate far-field uniform interference based on the spatial magnetic field gradient tensor matrix to reconstruct the near-field resistive current signal characterizing the near-field characteristics of the surge arrester. The spatial cross-phase-locked verification module is used to perform feature decomposition and spatial mapping on the spatial magnetic field gradient tensor matrix and the two-dimensional temperature distribution matrix respectively, extract the spatial magnetic field distortion pointing vector and the surface thermal gradient distortion pointing vector, and perform spatial cross-phase-locked verification by calculating and comparing the overlap of the two in three-dimensional space. The phase space topology analysis module is used to extract the first-order partial derivative of the two-dimensional temperature distribution matrix in the time dimension to generate the temperature change rate after the spatial cross-locking verification is passed, construct a two-dimensional thermo-electrodynamic phase space trajectory with the near-field resistive current signal and the temperature change rate, and extract the topological geometric distortion parameters of the thermo-electrodynamic phase space trajectory. The feedback closed-loop and prediction evaluation module is used to use the topological geometric distortion parameters as feedback control parameters, dynamically reconstruct the node differential calculation weights, extract the time evolution slope of the topological geometric distortion parameters, and combine the preset degradation equation to reverse calculate the remaining effective life of the surge arrester, generating a fault prediction and health management report containing pre-diagnosis results.

9. A non-contact detection system for the aging state of a surge arrester according to claim 8, characterized in that, The multi-physics synchronous sensing module includes a composite non-contact detection hardware architecture, which is composed of a regular tetrahedral array based on a tunneling magnetoresistive sensor and an infrared thermal imager. Each node of the regular tetrahedral array is used to synchronously acquire and output the multi-node spatial magnetic induction intensity vector.

10. A non-contact detection system for the aging state of a surge arrester according to claim 8, characterized in that, The topological geometric distortion parameters include the hysteresis loop area and the topological eccentricity; the feedback loop and prediction evaluation module has a built-in Weibull degradation equation engine, which is configured to map the extracted hysteresis loop area's temporal evolution slope and the topological eccentricity to the scale parameters and shape parameters of the preset degradation equation, respectively, to achieve dynamic prediction of the remaining effective lifetime of the device at the whole life cycle.

Images